PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of aemPermissionsJournals.ASM.orgJournalAEM ArticleJournal InfoAuthorsReviewers
 
Appl Environ Microbiol. 2010 April; 76(7): 2075–2085.
Published online 2010 January 29. doi:  10.1128/AEM.02395-09
PMCID: PMC2849255

Monitoring the Effects of Chiral Pharmaceuticals on Aquatic Microorganisms by Metabolic Fingerprinting[down-pointing small open triangle]

Abstract

The effects of the chiral pharmaceuticals atenolol and propranolol on Pseudomonas putida, Pseudomonas aeruginosa, Micrococcus luteus, and Blastomonas natatoria were investigated. The growth dynamics of exposed cultures were monitored using a Bioscreen instrument. In addition, Fourier-transform infrared (FT-IR) spectroscopy with appropriate chemometrics and high-performance liquid chromatography (HPLC) were employed in order to investigate the phenotypic changes and possible degradation of the drugs in exposed cultures. For the majority of the bacteria studied there was not a statistically significant difference in the organism's phenotype when it was exposed to the different enantiomers or mixtures of enantiomers. In contrast, the pseudomonads appeared to respond differently to propranolol, and the two enantiomers had different effects on the cellular phenotype. This implies that there were different metabolic responses in the organisms when they were exposed to the different enantiomers. We suggest that our findings may indicate that there are widespread effects on aquatic communities in which active pharmaceutical ingredients are present.

Active pharmaceutical ingredients (APIs) and their metabolites are ubiquitous in the environment (12), and the occurrence of APIs in the aquatic environment is a growing concern (13). There are a number of routes by which APIs and their metabolites and degradation products may enter these ecosystems, and a common avenue is through excretion of the APIs and their metabolites in urine and feces. It is known that APIs have different rates of metabolism in humans. For example, the β-blocker propranolol is almost completely metabolized in the liver, and only 1 to 4% of an oral dose is excreted as the unchanged API and its metabolites. In contrast, 40 to 50% of an oral dose of atenolol (also a β-blocker) is excreted as the API or its metabolites (2, 6, 7). Subsequent degradation of the APIs and their metabolites may also occur at sewage treatment plants (STPs); this degradation is usually substrate specific and varies greatly for different APIs. The rates of adsorption to activated sewage sludge during treatment differ for different APIs and are dependent on the hydrophobic and electrostatic interactions of the APIs with the particulates and microorganisms in the activated sewage sludge (13). Any remaining APIs and relevant metabolites are diluted in the surface water when the effluent is released from the STP. Hence, many APIs are present at low concentrations (ng liter−1 to μg liter−1) in aquatic environments, such as rivers, streams, and estuaries (2, 6, 12). The majority of APIs are neither persistent nor highly bioaccumulative; however, the continuous release of APIs into the aquatic environment poses a potential risk to aquatic organisms even though the concentrations of APIs in the receiving waters are quite low (12).

Despite the fact that little is known about the effects of APIs in the environment, the fact that they are designed to have a specific mode of action in humans must be taken into account (12). Adverse side effects may occur in humans at higher doses of the APIs, and it can be expected that any beneficial or adverse effect may also be observed in aquatic organisms with similar biological functions or receptors. It must also be noted that similar targets may control different metabolic processes in different species (43), and therefore APIs and their metabolites may have additional modes of action in aquatic organisms. The effects of the APIs may be subtle due to the very low concentrations observed in the aquatic environment, and as a result these effects may go unnoticed (12). It is also likely that the effect of an API has an impact on the local population dynamics in the whole ecosystem, from bacteria to higher organisms. To explore the effects of APIs on biological systems, a wide range of concentrations should be employed along with appropriate analytical platforms to profile the complement of biochemical components in cells. Indeed, it is known that APIs can become concentrated in the benthic environment of river beds, and as bacteria inhabit this niche, the bacterial community may be exposed to higher-than-expected levels of these compounds (16, 37, 49).

While the effect of APIs in the environment is currently a growing area of research, there is very little understanding of the environmental effects of chiral pharmaceuticals (5, 14). A chiral molecule is a molecule that lacks an internal plane of symmetry. The nonsuperimposable mirror images are termed enantiomers and are labeled (R) or (S) according to a priority system (Cahn Ingold Prelog priority rules) based on the atomic number of the molecule's substituents. Approximately 56% of the APIs currently in use are chiral compounds, and 88% of these chiral APIs are administered therapeutically as the racemate [i.e., an equal mixture of the two enantiomers, indicated by (±)]. The chirality of environmental contaminants, such as APIs, must be taken into consideration in order to fully understand the environmental fate and effects of these compounds. The enantiomers of a chiral API are able to interact differently with other chiral compounds, such as enzymes, and therefore potentially have different effects when they are released into the environment (5, 14, 33). It is widely known that the enantiomers of a chiral API may have different toxicological and biological effects than each other and than the racemate (an equal mixture of the two enantiomers) (25, 54). It has been shown that the (S) enantiomers of the β-blocking agents atenolol and propranolol are more potent in humans than the corresponding antipodes (3, 11, 35, 45) and that a number of the biotransformation pathways for β-blockers are stereoselective in humans (30). The mode of action of the drugs and their enantiomers in prokaryotic systems is not known. It is therefore necessary to increase our understanding of the fate and biological effects of chiral pharmaceuticals on typical microflora in the aquatic environment in order to fully appreciate the risks (19). Of particular interest is the group of APIs termed β-blockers as they all contain at least one chiral center and are generally administered therapeutically as the racemate (30). In addition, they are widely used, and for example, approximately 29 and 12 tonnes of atenolol and propranolol, respectively, are consumed each year in the United Kingdom (2, 6, 7).

In order to explore the effects of the APIs on biological systems, we employed Fourier-transform infrared (FT-IR) spectroscopy; this is a phenotypic typing technique which has previously been used to generate metabolic fingerprints of bacteria (22, 53). Previous studies have successfully discriminated bacteria to the subspecies level (31, 50, 53) through detection of subtle changes in the biochemical phenotypes of the bacteria. We recently demonstrated use of FT-IR spectroscopy coupled with suitable chemometrics to physiologically assess bioprocesses (unpublished data). In addition, a combination of FT-IR spectroscopy and trajectory analysis has been used for identification of metabolic changes in fermentations (22). FT-IR spectroscopy is an automated high-throughput technique (10 to 60 s per sample is typical) that requires minimal sample preparation, and this makes it relatively inexpensive. It is therefore an ideal screening method to explore the effects of APIs on a number of bacterial systems.

In this study the chirality-specific metabolism of the β1-selective adrenergic blocking agent atenolol and the nonselective β-adrenergic blocking agent propranolol by a range of environmental microorganisms was investigated (14, 41, 48). FT-IR spectroscopy was employed to monitor biochemical changes in the spectral fingerprints of whole bacterial cells during growth of the microorganisms in the presence of the selected APIs. In addition, we monitored the fate of the APIs with chiral high-performance liquid chromatography (HPLC), which allowed quantification of the enantiomers.

MATERIALS AND METHODS

Cultivation of bacteria.

In order to monitor the effects of the APIs used in the aquatic environment, a variety of microorganisms were selected for this investigation. All of the microorganisms employed in this study have been reported to be common in the aquatic environment and are amenable to growth in the laboratory. The following four bacteria were selected for this study: Pseudomonas putida KT2440, which is known to inhabit freshwater streams and activated sewage sludge (21, 29); Pseudomonas aeruginosa PA14, which is commonly isolated from freshwater streams (47); and Micrococcus luteus 2.13 (40) and Blastomonas natatoria 2.1 (40, 44), which have been isolated from freshwater biofilms. The bacteria were cultured in R2A medium (38) at 15°C for 24 h at 200 rpm in a Multitron (INFORS HT, Switzerland) orbital shaker unless otherwise stated. The pure enantiomers [(R) and (S) enantiomers] of both atenolol and propranolol (as a hydrochloride) were purchased from Sigma-Aldrich Company Limited (Poole, Dorset, United Kingdom).

Screening of microorganisms for growth in the presence of APIs.

The growth of each bacterium was monitored (using optical density at 600 nm determined with a Bioscreen spectrophotometer [Labsystems, Basingstoke, United Kingdom]) at a range of concentrations (10 to 130 μg ml−1) of each enantiomer of each API and the racemate. The data collected in these investigations were used to calculate the specific growth rate and death rate (the death rate is the rate when the rate of cell death or lysis exceeds the rate of growth so that there is a decrease in the turbidity of the culture [39]) for the exponential phases using the following equation: μ = 2.303(log10OD2 − log10OD1)/t2t1), where μ is the specific growth rate or death rate, log10OD1 is the log10 optical density at time point 1, log10OD2 is the log10 optical density at time point 2, t1 is time point 1, and t2 is time point 2. The growth rate data (see below) were used to select a reduced range of concentrations for further investigation; the concentrations selected were based on identifiable differences in the growth rate which did not result in cell death.

Batch growth.

Bacterial cultures were exposed in triplicate to a range of concentrations of the chosen API; both the pure enantiomers and a range of enantiomeric mixtures were used (Table (Table1).1). An aliquot (1 ml) of sterile water was added to an additional set of bacterial samples as a control. Samples were maintained at 15°C and 200 rpm in a Multitron orbital shaker for 24 h. Aliquots (2 ml) were taken in triplicate from each flask and centrifuged (5 min, 0°C, 16,089 × g) to harvest the biomass. The supernatant and pelleted biomass were stored at −80°C before further analysis.

TABLE 1.
Microorganisms and conditions used for batch growth

Quantitative analysis of API concentration by HPLC.

Concentrations of atenolol and propranolol were determined by HPLC (Agilent 1100 series). The supernatant samples were allowed to thaw at room temperature and were filtered (0.22 μm; Millipore) in order to remove any microbial cells remaining in the medium. Aliquots (25 μl) were injected onto the HPLC column in a random order. Each sample was injected three times during the analysis, resulting in three analytical replicates for each biological sample. The HPLC system was equipped with a Chirobiotic V2 column (250 mm by 4.6 mm [inside diameter]; particle size, 5 μm; ASTEC, Whippany, NY) and a UV detector operating at a wavelength of 230 nm. The column was eluted with an isocratic mixture of methanol and water (90:10, vol/vol) and 1.0% triethylamine acetate (TEAA) buffer (pH 5.0). The pH of the buffer was adjusted with acetic acid prior to the addition of methanol. The measurements were obtained at 25 ± 1°C at a flow rate of 1 ml min−1 (4).

Analysis of microbial cells by FT-IR spectroscopy.

Ninety-six-well zinc selenide plates were cleaned by rinsing them with 2-propanol and deionized water (three times) and were allowed to dry at room temperature (18, 53). The cell pellets stored at −80°C were allowed to thaw at room temperature and washed in order to remove any traces of residual API. Ice-cold sterile water (2 ml) was added to each sample and gently vortexed. The samples were centrifuged for 10 min (0°C, 16,089 × g), and the supernatants were discarded; this cycle was repeated three times. A final 100-μl aliquot of sterile water was added to each sample, and the solution was vortexed. Aliquots (20 μl) of each resuspended sample were applied to ZnSe plates and oven dried at 50°C for 10 min. Drying was used to minimize any signal arising from the absorption of water in the mid-IR region, which would mask the biologically important chemical information in the spectra. Three replicates of each of the samples were randomly applied to the ZnSe plates, and triplicate spectra were obtained using different positions in each well; a total of nine spectra (called technical replicates) per sample were collected. Each plate was loaded onto an HTS-XT motorized microplate module under computer control by the OPUS software (version 4) (53). Spectra were collected with an Equinox 55 FT-IR spectrometer (Bruker Optics Ltd.) in transmission mode using a deuterated triglycine sulfate detector over a wavelength range from 4,000 to 600 cm−1 and with a resolution of 4 cm−1. Sixty-four spectra were co-added to improve the signal-to-noise ratio. The spectra were displayed in terms of absorbance (Fig. (Fig.11 shows typical example spectra).

FIG. 1.
Typical processed FT-IR spectra for P. aeruginosa PA14 exposed to 80 μg ml−1 (R)-propranolol, (S)-propranolol, and (±)-propranolol. Control samples (Con) which were not exposed to propranolol were included. The spectra are offset ...

Analysis of FT-IR spectroscopy data. (i) Spectral preprocessing.

The ASCII data were imported into Matlab version 7.1 (The MathWorks, Inc., Natick, MA), and in the initial step spectral regions which were dominated by CO2 vibrations arising from the atmosphere (2,403 to 2,272 cm−1 and 683 to 656 cm−1) were removed and filled with a linear trend. The spectra were corrected using extended multiplicative scatter correction (EMSC), which normalizes and smoothes spectra by application of a polynomial smoothing function (28). The resulting preprocessed spectra were used for subsequent multivariate analyses.

(ii) Multivariate analysis.

The protocol used for multivariate analysis was a protocol developed previously (1, 15). Principal component analysis (PCA) is an unsupervised method for reducing the dimensionality of multivariate data while preserving the variance. This transformation was performed prior to canonical variate analysis (CVA). CVA is a supervised learning method that seeks to minimize within-group variance while it maximizes between-group variance, and it can be used in conjunction with PCA to discriminate between groups on the basis of retained principal components (PCs), given a priori knowledge of group membership of spectral replicates (26, 52). In this study, PC-CVA models were constructed with a priori knowledge of the biological replicates. In order to make sure that these models were not over- or undertrained, validation was performed using the full cross-validation method, where two of the biological replicates were used for model training and the third replicate was projected into the model for cluster validation (20). Finally, CVA also allowed statistical significance to be displayed on the score plots, and circles were used to indicate the 95% χ2 confidence region constructed around each group mean based on the χ2 distribution with 2 degrees of freedom (24).

Partial least-squares (PLS) (27) regression is a multivariate linear regression method which allows the quantitative relationship between different variables (e.g., API concentration and FT-IR spectra) to be modeled, and it can deal efficiently with data sets that are highly correlated. In this study, PLS regression was employed to predict the API concentration values from the FT-IR spectroscopy data. Like the PC-CVA, the regression models were calibrated with two of the three biological replicates, and the third replicate was used as an independent test set to validate the model and establish whether the models could generalize.

RESULTS AND DISCUSSION

Effects of the chiral APIs on the bacterial growth rates.

A number of aquatic microorganisms were exposed to the chiral APIs atenolol and propranolol, and growth rates, death rates, and maximum optical densities were determined to monitor the effects of the APIs on culture progress; Fig. Fig.22 shows data for 0, 10, 50, 90, and 130 μg ml−1 of the (R) and (S) enantiomers and the racemic mixtures. Slight variations in the specific growth rates were observed for the Pseudomonas species exposed to different concentrations (10 to 130 μg ml−1) of (R)-, (S)-, and (±)-propranolol. There was a considerable difference in the growth rates of species, and a marked effect was observed for P. aeruginosa PA14 exposed to propranolol. In contrast, minimal changes were detected in the growth rates, death rates, and maximum amounts of biomass of both Pseudomonas species exposed to 10 to 130 μg ml−1 of (R)-, (S)- and (±)-atenolol.

FIG. 2.
Specific growth rate data for P. putida KT2440 and P. aeruginosa PA14 exposed to 0 to 130 μg ml−1 of propranolol or atenolol. The maximum optical densities (OD) at 600 nm and specific death rates are also shown. The data are averages from ...

An interesting effect was observed for P. aeruginosa PA14 exposed to both of the propranolol enantiomers and the racemate. At concentrations of 50 to 70 μg ml−1 there appeared to be no death of the microbial cells. In contrast, for cells exposed to 10 to 40 μg ml−1 and to 80 to 130 μg ml−1 the death rate was equivalent to that of the control cells. This was probably because the lower concentrations (<40 μg ml−1) of propranolol had very little effect on metabolism so cells quickly reached the stationary and death phases and because the higher concentrations (>80 μg ml−1) had a negative impact on metabolism and killing cells (as also indicated by the fact that the final turbidity measurements were significantly lower than the turbidity measurements for the control cells), while the intermediate concentrations (50 to 70 μg ml−1) slowed growth but the cells did not enter the death phase. Inspection of the growth curves indicated that there was a second phase of growth several hours into the stationary phase. While a second phase may indicate that there is utilization of a secondary carbon source, this was not observed in the control cultures and thus is not the likely explanation for this observation. The biomass of the culture decreased as the concentration of the API increased, and thus the original carbon source was potentially not depleted at the onset of stationary phase. While at the higher concentrations of propranolol a slight increase in the amount of biomass was immediately followed by a noticeable decrease in the optical density of the culture (death phase), the maximum amount of biomass was severely inhibited by the presence of the API. Our observations suggest that the API has different effects depending on the concentration applied. At lower concentrations the growth is not affected by the API, and at high concentrations death occurs during the growth period. However, at intermediate concentrations production of biomass occurs throughout the growth period (onset of the death phase may have been observed if the growth had been monitored for extended periods). The APIs were not metabolized during growth (Table (Table2),2), and we hypothesize that the intermediate concentrations of propranolol affected either the transport of nutrients into the cell or the rate of metabolism.

TABLE 2.
Quantification of propranolol from HPLC data for bacterial cells exposed to different ratios of (R)-propranolol to (S)-propranolol at a concentration of 50 μg ml−1

The growth rate data for the cells exposed to propranolol clearly showed that the specific growth rate decreases as the concentration of the API increases. This trend was also observed in the maximum optical density data and the death rate data for both pseudomonads. Our findings indicate that propranolol has considerably different effects on the two Pseudomonas species. These findings are rather surprising as these species are genetically closely related. Estimates have shown that there is greater similarity (60% of the predicted coding sequences) between these two pseudomonads than between any other complete microbial genomes obtained to date (32). In addition, comparative genome analysis has shown that 85% of the genes in the P. putida KT2440 genome have homologues in the P. aeruginosa PAO1 genome (46).

The toxic effects of the APIs observed here for aquatic organisms have been previously reported. Toxicity studies carried out by Choi and coworkers with the crustacean Thamnocephalus platyurus and a fish species (Oryzias latipes) showed that propranolol caused acute toxicity in T. platyurus at a concentration of 10.61 μg ml−1 and in O. latipes at a concentration of 11.40 μg ml−1. In contrast to the results presented here, these workers found that atenolol did not have toxic effects in the aquatic organisms at the concentrations that they used (<100 μg ml−1) (9). In addition, toxicity studies have been carried out with a range of APIs (including propranolol) using the Japanese medaka fish (O. latipes), an amphipod (Hyalella azteca), and two crustaceans (Ceriodaphnia dubia and Daphnia magna). It was found that propranolol had the greatest effect on the organisms studied. The crustacean C. dubia displayed responses to toxicity at a concentration of 0.25 μg ml−1. Propranolol was the only API investigated which was found to have acute toxic effects in the Japanese medaka fish. These effects were observed at a concentration of 0.5 μg ml−1 (19).

Previous studies have suggested that β-blockers do not affect microbes due to the absence of the API receptors in these organisms (10, 23). However, in another study conducted by our group we reported that (±)-propranolol significantly reduced the amount of lipid storage components of the alga Micrasterias hardyi 649/15 and caused a marked reduction in the cellular protein content (34). In addition, the findings obtained by metabolic fingerprinting suggested that the phenotype was altered during exposure to this API (34). To our knowledge, no further studies on the metabolic effects of propranolol in aquatic microorganisms have been carried out.

The effects on the growth dynamics of the bacteria are likely to reflect changes in the metabolic potential of the cells, and this quantitative drug effect was explored further using FT-IR spectroscopy and P. putida KT2440 exposed to (±)-propranolol.

Quantitative effects of APIs on bacteria measured using FT-IR spectroscopy.

In order to assess possible quantitative effects of propranolol on the phenotype of P. putida KT2440, we employed partial least-squares (PLS) regression analysis to investigate whether the effect on the phenotype as measured using FT-IR spectroscopy was directly proportional to the concentration of API applied (Fig. (Fig.3).3). A clear linear relationship was observed between the concentration of (±)-propranolol to which the P. putida KT2440 cells were exposed and the metabolic fingerprint. In addition, we were able to predict the concentration of propranolol to which the bacterial cells were exposed with an accuracy of 95.45%. This is perhaps not surprising as the inhibitory effect of the propranolol on the cells was proportional to the concentration of API. This was a clear phenotypic effect as we were unable to collect a propranolol spectrum at these concentrations when using FT-IR spectroscopy. We also performed PLS regression analysis with the profiles of P. aeruginosa PA14 exposed to the intermediate concentrations of propranolol to determine if the secondary growth effect was proportional to the concentration of API applied (data not shown). Under these conditions it was not possible to obtain a correlation between the drug concentration and the FT-IR spectroscopy data. Therefore, the presence of the drug may have led to more complex biochemical perturbations in the organisms that we were unable to model using PLS regression.

FIG. 3.
Partial least-squares regression model for P. putida KT2440 exposed to various concentrations (0 to 100 μg ml−1 in steps of 10 μg ml−1) of (±)-propranolol. The model was trained with FT-IR spectroscopy data using ...

Quantitative analysis of API concentration with HPLC.

Chiral HPLC analysis was performed to quantify the amounts of the enantiomers remaining at the end of the growth period. This analysis was used to explore the effects shown by the growth rate data. In addition, control experiments were performed to determine the effect of the experiment on the drug concentration (i.e., the loss of API during incubation in growth medium). To determine the effects of the enantiomers on the growth of the pseudomonads, a range of ratios of propranolol enantiomers were employed using a concentration of 50 μg ml−1. The findings of the HPLC analysis (Table (Table2)2) demonstrated that neither of the enantiomers was degraded during batch growth. In addition, the pseudomonads were exposed to a range of concentrations of the API atenolol (Table (Table3);3); the concentrations selected were chosen based on the growth rate data. There was no notable indication of API degradation during growth of the bacteria.

TABLE 3.
Quantification of atenolol from HPLC data for bacterial cells exposed to different ratios of (R)-atenolol to (S)-atenolol at various concentrations

Effects of chiral APIs on FT-IR spectroscopy metabolic fingerprints.

In order to investigate whether there were any chirality-specific phenotypic changes in the various aquatic bacteria, a single drug concentration with which there was no observable difference in the growth rate between the enantiomers was chosen for investigation.

During the investigation of the chiral API-specific effects on microorganisms, the four bacteria were exposed to a number of drug concentrations (Table (Table1).1). A summary of the statistically significant differences in the data for the drug enantiomers, enantiomer ratios, and racemate is shown in Table Table44.

TABLE 4.
Concentrations of propranolol and atenolol at which a significant effect on the bacterial phenotype was observed using FT-IR spectroscopy

PC-CVA was carried out in order to investigate any chirality-specific effects on the microorganisms as determined by FT-IR spectroscopy. The distance between samples plotted on a PC-CVA score plot represents the degree of similarity or dissimilarity between the samples. A smaller distance indicates greater similarity, and a larger distance indicates that there are greater differences between samples. Loading plots provide an indication of which regions of the spectrum are used to define the patterns of separation, which allows meaningful biochemical interpretation of the results. FT-IR spectroscopy analysis demonstrated that each of the propranolol enantiomers had a metabolic effect on the cells of P. aeruginosa PA14 at a concentration of 80 μg ml−1 (Fig. (Fig.4a)4a) compared to the control. The PC-CVA score plot clearly shows that the control samples separate across PC-canonical variate 1 (CV1), which accounts for the greatest variance in the data according to the putative class assignment; as discussed above, this finding was perhaps not surprising given the effect of the APIs on the bacterial growth dynamics (Fig. (Fig.2).2). In addition, the samples exposed to (±)-propranolol are clearly separate from the samples exposed to each of the enantiomers [(R) and (S)], which in this analysis showed no separation across the first two PC-CV scores. As described above, two of the three biological replicates were used for calibration (indicated by black type in Fig. Fig.4),4), and the third biological replicate was projected into the model (indicated by gray type). The majority of the projected data are grouped with the appropriate calibration samples, indicating that the separation shown in the model was valid. Moreover, the 95% confidence intervals for the groups are also indicated in Fig. Fig.44 and show that for three of the groups there is a distinct separation in CVA score space. This analysis demonstrated that the microbial cells exposed to (R)- and the microbial cells exposed to (S)-propranolol clustered together, indicating that there were no metabolic differences in the microbial cells exposed to the two pure enantiomers. It was very surprising that the cultures exposed to the racemate formed a distinct cluster separate from the clusters containing cultures exposed to the pure enantiomers or control samples in the CVA space. The loading plot (Fig. (Fig.4b)4b) indicates that very specific changes in the metabolic fingerprints of the microbial cells account for the patterns of separation observed for the control and drug-exposed samples in the scores plot. The major chemical changes occur in the protein (1,681 to 1,629 cm−1) and carbohydrate (1,155 to 999 cm−1) regions of the FT-IR spectrum; there was a less pronounced contribution from lipid species (2,951 to 2,845 cm−1), which changed in the same direction as the carbohydrates. To investigate the effects of the racemate and enantiomers on the biochemical components of the cells further, we calculated difference spectra [for example, the metabolic fingerprint of the racemate was subtracted from that of the (R) enantiomer] for each combination of interest. The resulting data were then used to determine the relative changes in the lipid and amide components of the cells when they were exposed to drugs. Inspection of the difference spectra revealed that the bacterial cells exposed to the racemate contained lower levels of amides and higher levels of lipids than the cells exposed to either of the enantiomers. This suggests that the racemate has less metabolic effect on the bacterial cells, and this suggestion is supported by the PC-CVA scores plot, in which the cells exposed to racemate are between the control cells and the enantiomer-exposed cells across PC-CV1. As discussed above, HPLC analysis suggested that degradation or significant uptake of the APIs did not occur in the microbial cells. Therefore, it is unlikely that the increase in the levels of proteins shown by the FT-IR spectra of propranolol-exposed cells was due to expression of enzymes in order to metabolize this API. It is more probable that this effect was due to expression of an efflux system to remove the API from the bacterial cells. In addition, propranolol is a lipophilic API which is known to interact with cell membranes of mammalian cells, and the observed reduction in the level of lipids in exposed cells was likely due to interactions of the API with the bacterial cell wall. Propranolol is routinely administered to humans as the racemate. The (S) enantiomer accounts for the majority of the β-blocking effect, while the (R) enantiomer has a predominantly membrane-stabilizing effect (3, 17, 36, 51). We hypothesized that the results for the racemate [(±)-propranolol] were different from the results for either of the enantiomers because of the difference in the physical properties between the racemate and the enantiomers (8, 42).

FIG. 4.
PC-CVA score (left side) and loading (right side) plots for FT-IR spectroscopy data for P. aeruginosa PA14 exposed to (R)-propranolol, (S)-propranolol, and (±)-propranolol at a concentration of 80 μg ml−1 (a and b), for P. putida ...

To explore the chirality-specific effect observed in the experiments described above further, the pseudomonads were exposed to various ratios of (R)-propranolol to (S)-propranolol at a concentration of 50 μg ml−1. The results of the chemometric analysis of the FT-IR spectra also showed that there was a metabolic difference between the microbial cells exposed to different ratios of (R)-propranolol to (S)-propranolol and the control cells (data not shown), and the data for the controls were removed prior to PC-CVA so that only chirality-specific changes were observed. The results of this PC-CVA are shown in Fig. Fig.4c,4c, which shows the results for P. putida KT2440 exposed to 50 μg ml−1 (±)-propranolol (indicated by R:S) on the left side and the results for the 25:75 mixture of (R)-propranolol and (S)-propranolol (indicated by r:S) and for the pure enantiomers on the right side. The 75:25 mixture of (R)-propranolol and (S)-propranolol (indicated by R:s) falls between these two groups. The clear separation of P. putida KT2440 exposed to the racemate supports the observations on the effect of propranolol on the metabolic fingerprints of P. aeruginosa PA14 described above.

In contrast, no phenotypic variation was observed in the metabolic fingerprints of the samples exposed to (R)-, (S)-, and (±)-atenolol (Fig. 4e and f) as the 95% χ2 confidence regions overlap. This analysis clearly showed that there was no difference between the three treatments. This result differs from the results for P. aeruginosa PA14 exposed to propranolol (Fig. (Fig.4a)4a) and indicates that atenolol does not have a chirality-specific metabolic effect on P. aeruginosa PA14.

Due to the chirality-specific effects on the pseudomonads observed when they were exposed to propranolol, two additional bacteria were used to investigate the effects of propranolol. Propranolol had a very noticeable metabolic effect on B. natatoria 2.1 at concentrations of 40 and 50 μg ml−1 (Fig. 5a and b) and on M. luteus 2.13 at a concentration of 50 μg ml−1 (Fig. 5c and d). The greatest difference observed in these analyses was the difference between the control and API-exposed samples. To investigate the more subtle differences between the cultures exposed to the different enantiomers and the racemate, the control samples were removed from the analysis. The B. natatoria 2.1 samples in the PC-CVA score plot are separated across the first CV with respect to the enantiomers. The (S) and (R) enantiomers are clearly separated in the CVA space, and the racemate is located between them. A concentration effect was also observed in the metabolic fingerprints across CV2. This is in contrast to the chirality-specific effects of this API on the two pseudomonads, in which the greatest variation was found between the cells exposed to the racemate and the cells exposed to the enantiomers. The loading data for B. natatoria 2.1 indicate that the major chemical changes occur in the lipid (2,936 to 2,851 cm−1) region of the FT-IR spectrum and at 1,748 to 1,654 cm−1. Vibrations in this region may be attributed to the C=O stretching of esters and carboxylic acids; however, this region is dominated by amide I. The FT-IR spectra demonstrate that the cells exposed to (S)-propranolol contained lower levels of lipids but higher levels of amide and carbohydrate than the cells exposed to the (R) enantiomer. This suggests that (S)-propranolol has a greater biological effect on the bacterial cells. The effect of this API on M. luteus 2.13 and B. natatoria 2.1 is perhaps more predictable as the different metabolic effects of the enantiomers are linearly additive. The difference between the phenotypic effects on these organisms following exposure to propranolol and the phenotypic effects on pseudomonads is probably a consequence of the metabolic differences between the bacteria.

FIG. 5.
PC-CVA score (left side) and loading (right side) plots for FT-IR spectroscopy data for B. natatoria 2.1 exposed to (R)-propranolol, (S)-propranolol, and (±)-propranolol at concentrations of 40 and 50 μg ml−1 (a and b) and for ...

To our knowledge, the β-blockers atenolol and propranolol have not previously been studied to examine chirality-specific effects in microbial systems. Nevertheless, the effects of APIs in these systems are highly relevant, as microorganisms populate the lower trophic levels in food webs. Therefore, differences in the population dynamics could represent significant effects on the whole freshwater community (23).

Conclusion.

The growth data clearly showed that propranolol had a biological effect on all of the microorganisms studied. At the higher concentrations tested growth was retarded, and in most cases the death rate increased; associated changes were observed in the metabolic fingerprints. The loading plots from the PC-CVA of API-exposed and unexposed P. aeruginosa PA14 cells (Fig. (Fig.4b)4b) indicate that propranolol has a widespread effect on bacterial cells, and this effect was also observed in the other bacteria studied. The results of the HPLC analysis showed that this API was not degraded during the growth period, and this suggests that the observed changes in the multivariate analysis of the metabolic fingerprints were not due to degradation of the API but more likely were a secondary effect of the drug. Despite the genetic similarity of the two pseudomonads studied, our findings show that propranolol had different effects in the two species. In contrast, the growth data show that no effects were observed in the atenolol-exposed cultures, and this finding was reflected in the results of the multivariate analyses of the bacterial fingerprints (Fig. (Fig.4e4e).

All four aquatic bacteria were exposed to the enantiomers of propranolol at concentrations at which no difference in the growth rates was observed. The FT-IR spectroscopy analysis revealed that propranolol affected both the lipid and protein contents of the bacterial cells. We hypothesize that this was likely due to the interaction of the APIs with the microbial cell walls. A more predictable effect on the metabolic fingerprints was noted during the analysis of exposure of B. natatoria 2.1 and M. luteus 2.13 to propranolol, in which the racemate fell between the (R) and (S) enantiomers in the PC-CVA. Rather surprisingly, the most significant effect on the two pseudomonads was the effect of the racemate, while the enantiomers had identical effects on the phenotypes of the cells. It is possible that the physical properties of the racemate were significantly different from those of the (R) and (S) enantiomers and that this was reflected in how the cells responded to exposure to this API. This possibility will be investigated in the future.

In conclusion, we have shown that chirality-specific effects do occur in bacteria, which may have implications for environmental ecosystems as APIs are regularly found in the aquatic environment. We believe that FT-IR spectroscopy with appropriate chemometrics is a very powerful method for investigating the phenotypes and metabolic differences between APIs when they interact with bacterial cells.

Acknowledgments

E.S.W. thanks AstraZeneca and UK BBSRC for financial support. C.L.W., R.M.J., and R.G. thank UK BBSRC for funding.

Footnotes

[down-pointing small open triangle]Published ahead of print on 29 January 2010.

REFERENCES

1. Alsberg, B. K., W. G. Wade, and R. Goodacre. 1998. Chemometric analysis of diffuse reflectance-absorbance Fourier transform infrared spectra using rule induction methods: application to the classification of Eubacterium species. Appl. Spectrosc. 52:823-832.
2. Ashton, D., M. Hilton, and K. V. Thomas. 2004. Investigating the environmental transport of human pharmaceuticals to streams in the United Kingdom. Sci. Total Environ. 333:167-184. [PubMed]
3. Barrett, A. M., and V. A. Cullum. 1968. The biological properties of the optical isomers of propranolol and their effects on cardiac arrhythmias. Br. J. Pharmacol. 34:43-55. [PMC free article] [PubMed]
4. Bosakova, Z., E. Curinova, and E. Tesarova. 2005. Comparison of vancomycin-based stationary phases with different chiral selector coverage for enantioselective separation of selected drugs in high-performance liquid chromatography. J. Chromatogr. A 1088:94-103. [PubMed]
5. Buser, H.-R., T. Poiger, and M. D. Muller. 1999. Occurrence and environmental behavior of the chiral pharmaceutical drug Ibuprofen in surface waters and in wastewater. Environ. Sci. Technol. 33:2529-2535.
6. Carlsson, C., A.-K. Johansson, G. Alvan, K. Bergman, and T. Kuhler. 2006. Are pharmaceuticals potent environmental pollutants? Part I. Environmental risk assessments of selected active pharmaceutical ingredients. Sci. Total Environ. 364:67-87. [PubMed]
7. Carlsson, C., A.-K. Johansson, G. Alvan, K. Bergman, and T. Kühler. 2006. Are pharmaceuticals potent environmental pollutants? Part II. Environmental risk assessments of selected pharmaceutical excipients. Sci. Total Environ. 364:88-95. [PubMed]
8. Chickos, J. S., D. L. Garin, M. Hitt, and G. Schilling. 1981. Some solid state properties of enantiomers and their racemates. Tetrahedron 37:2255-2259.
9. Choi, H.-J., B.-H. Kim, J.-D. Kim, and M.-S. Han. 2005. Streptomyces neyagawaensis as a control for the hazardous biomass of Microcystis aeruginosa (Cyanobacteria) in eutrophic freshwaters. Biol. Control 33:335-343.
10. CSTEE. 2001. Opinion on draft discussion paper on environmental risk assessment of medicinal products for human use. C2/JCD/csteeop/CPMPpaperRAssessHumPharm12062001/D(01). European Commission, Brussels, Belgium.
11. Davies, C. L. 1990. Chromatography of β-adrenergic blocking agents. J. Chromatogr. B Biomed. Sci. Appl. 531:131-180. [PubMed]
12. Escher, B. I., N. Bramaz, R. I. L. Eggen, and M. Richter. 2005. In vitro assessment of modes of toxic action of pharmaceuticals in aquatic life. Environ. Sci. Technol. 39:3090-3100. [PubMed]
13. Fent, K., A. A. Weston, and D. Caminada. 2006. Ecotoxicology of human pharmaceuticals. Aquat. Toxicol. 76:122-159. [PubMed]
14. Fono, L. J., and D. L. Sedlak. 2005. Use of the chiral pharmaceutical propranolol to identify sewage discharges into surface waters. Environ. Sci. Technol. 39:9244-9252. [PubMed]
15. Goodacre, R., E. M. Timmins, R. Burton, N. Kaderbhai, A. M. Woodward, D. B. Kell, and P. J. Rooney. 1998. Rapid identification of urinary tract infection bacteria using hyperspectral whole-organism fingerprinting and artificial neural networks. Microbiology 144:1157-1170. [PubMed]
16. Halling-Sørensen, B., S. Nors Nielsen, P. F. Lanzky, F. Ingerslev, H. C. Holten Lützhøft, and S. E. Jørgensen. 1998. Occurrence, fate and effects of pharmaceutical substances in the environment—a review. Chemosphere 36:357-393. [PubMed]
17. Hanna, G. M., and F. E. Evans. 2000. Optimization of enantiomeric separation for quantitative determination of the chiral drug propranolol by 1H-NMR spectroscopy utilizing a chiral solvating agent. J. Pharm. Biomed. Anal. 24:189-196. [PubMed]
18. Harrigan, G. G., R. H. LaPlante, G. N. Cosma, G. Cockerell, R. Goodacre, J. F. Maddox, J. P. Luyendyk, P. E. Ganey, and R. A. Roth. 2004. Application of high-throughput Fourier-transform infrared spectroscopy in toxicology studies: contribution to a study on the development of an animal model for idiosyncratic toxicity. Toxicol. Lett. 146:197-205. [PubMed]
19. Huggett, D. B., B. W. Brooks, B. Peterson, C. M. Foran, and D. Schlenk. 2002. Toxicity of select beta adrenergic receptor-blocking pharmaceuticals (B-blockers) on aquatic organisms. Arch. Environ. Contam. Toxicol. 43:229-235. [PubMed]
20. Jarvis, R. M., and R. Goodacre. 2004. Discrimination of bacteria using surface-enhanced Raman spectroscopy. Anal. Chem. 76:40-47. [PubMed]
21. John, D. M., and G. F. White. 1998. Mechanism for biotransformation of nonylphenol polyethoxylates to xenoestrogens in Pseudomonas putida. J. Bacteriol. 180:4332-4338. [PMC free article] [PubMed]
22. Johnson, H. E., D. Broadhurst, D. B. Kell, M. K. Theodorou, R. J. Merry, and G. W. Griffith. 2004. High-throughput metabolic fingerprinting of legume silage fermentations via Fourier transform infrared spectroscopy and chemometrics. Appl. Environ. Microbiol. 70:1583-1592. [PMC free article] [PubMed]
23. Jones, O. A. H., N. Voulvoulis, and J. N. Lester. 2002. Aquatic environmental assessment of the top 25 English prescription pharmaceuticals. Water Res. 36:5013-5022. [PubMed]
24. Krzanowski, W. J. 1988. Principles of multivariate analysis: a user's perspective. Oxford University Press, New York, NY.
25. Lees, P., P. M. Taylor, F. M. Landoni, A. K. Arifah, and C. Waters. 2003. Ketoprofen in the cat: pharmacodynamics and chiral pharmacokinetics. Vet. J. 165:21-35. [PubMed]
26. Manly, B. F. J. 1994. Multivariate statistical methods: a primer. Chapman and Hall, London, United Kingdom.
27. Martens, H., and T. Naes. 1989. Multivariate calibration. John Wiley & Sons, Chichester, United Kingdom.
28. Martens, H., J. P. Nielsen, and S. B. Engelsen. 2003. Light scattering and light absorbance separated by extended multiplicative signal correction. Application to near-infrared transmission analysis of powder mixtures. Anal. Chem. 75:394-404. [PubMed]
29. Martínez-Bueno, M. A., R. Tobes, M. Rey, and J.-L. Ramos. 2002. Detection of multiple extracytoplasmic function (ECF) sigma factors in the genome of Pseudomonas putida KT2440 and their counterparts in Pseudomonas aeruginosa PA01. Environ. Microbiol. 4:842-855. [PubMed]
30. Mehvar, R., and D. R. Brocks. 2001. Stereospecific pharmacokinetics and pharmacodynamics of beta-adrenergic blockers in humans. J. Pharm. Pharm. Sci. 4:185-200. [PubMed]
31. Naumann, D., D. Helm, and H. Labischinski. 1991. Microbiological characterizations by FT-IR spectroscopy. Nature 351:81-82. [PubMed]
32. Nelson, K. E., C. Weinel, I. T. Paulsen, R. J. Dodson, H. Hilbert, V. A. P. Martins dos Santos, D. E. Fouts, S. R. Gill, M. Pop, M. Holmes, L. Brinkac, M. Beanan, R. T. DeBoy, S. Daugherty, J. Kolonay, R. Madupu, W. Nelson, O. White, J. Peterson, H. Khouri, I. Hance, P. C. Lee, E. Holtzapple, D. Scanlan, K. Tran, A. Moazzez, T. Utterback, M. Rizzo, K. Lee, D. Kosack, D. Moestl, H. Wedler, J. Lauber, D. Stjepandic, J. Hoheisel, M. Straetz, S. Heim, C. Kiewitz, J. Eisen, K. N. Timmis, A. Düsterhöft, B. Tümmler, and C. M. Fraser. 2002. Complete genome sequence and comparative analysis of the metabolically versatile Pseudomonas putida KT2440. Environ. Microbiol. 4:799-808. [PubMed]
33. Nikolai, L. N., E. L. McClure, S. L. MacLeod, and C. S. Wong. 2006. Stereoisomer quantification of the β-blocker drugs atenolol, metoprolol, and propranolol in wastewaters by chiral high-performance liquid chromatography-tandem mass spectrometry. J. Chromatogr. A 1131:103-109. [PubMed]
34. Patel, S. A., F. Currie, N. Thakker, and R. Goodacre. 2008. Spatial metabolic fingerprinting using FT-IR spectroscopy: investigating abiotic stresses on Micrasterias hardyi. Analyst 133:1707-1713. [PubMed]
35. Pearson, A. A., T. E. Gaffney, T. Walle, and P. J. Privitera. 1989. A stereoselective central hypotensive action of atenolol. J. Pharmacol. Exp. Ther. 250:759-763. [PubMed]
36. Potter, L. T., and C. B. Sweetland. 1967. Uptake of propranolol by isolated guinea-pig atria. J. Pharmacol. Exp. Ther. 155:91-100. [PubMed]
37. Pouliquen, H., H. Le Bris, and L. Pinault. 1992. Experimental study of the therapeutic application of oxytetracycline, its attenuation in sediment and sea water, and implications for farm culture of benthic organisms. Mar. Ecol. Prog. Ser. 89:93-98.
38. Reasoner, D. J., and E. E. Geldreich. 1985. A new medium for the enumeration and subculture of bacteria from potable water. Appl. Environ. Microbiol. 49:1-7. [PMC free article] [PubMed]
39. Rice, K. C., and K. W. Bayles. 2003. Death's toolbox: examining the molecular components of bacterial programmed cell death. Mol. Microbiol. 50:729-738. [PubMed]
40. Rickard, A. H., S. A. Leach, C. M. Buswell, N. J. High, and P. S. Handley. 2000. Coaggregation between aquatic bacteria is mediated by specific-growth-phase-dependent lectin-saccharide interactions. Appl. Environ. Microbiol. 66:431-434. [PMC free article] [PubMed]
41. Roberts, P. H., and K. V. Thomas. 2006. The occurrence of selected pharmaceuticals in wastewater effluent and surface waters of the lower Tyne catchment. Sci. Total Environ. 356:143-153. [PubMed]
42. Secor, R. M. 1963. Resolution of optical isomers by crystallization procedures. Chem. Rev. 63:297-309.
43. Seiler, J. P. 2002. Pharmacodynamic activity of drugs and ecotoxicology—can the two be connected? Toxicol. Lett. 131:105-115. [PubMed]
44. Sly, L. I., and M. M. Cahill. 1997. Transfer of Blastobacter natatorius (Sly 1985) to the genus Blastomonas gen. nov. as Blastomonas natatoria comb. nov. Int. J. Syst. Bacteriol. 47:566-568. [PubMed]
45. Stoschitzky, K., G. Egginger, G. Zernig, W. Klein, and W. Lindner. 1993. Stereoselective features of (R)- and (S)-atenolol: clinical pharmacological, pharmacokinetic, and radioligand binding studies. Chirality 5:15-19. [PubMed]
46. Stover, C. K., X. Q. Pham, A. L. Erwin, S. D. Mizoguchi, P. Warrener, M. J. Hickey, F. S. L. Brinkman, W. O. Hufnagle, D. J. Kowalik, M. Lagrou, R. L. Garber, L. Goltry, E. Tolentino, S. Westbrock-Wadman, Y. Yuan, L. L. Brody, S. N. Coulter, K. R. Folger, A. Kas, K. Larbig, R. Lim, K. Smith, D. Spencer, G. K. S. Wong, Z. Wu, I. T. Paulsen, J. Reizer, M. H. Saier, R. E. W. Hancock, S. Lory, and M. V. Olson. 2000. Complete genome sequence of Pseudomonas aeruginosa PAO1, an opportunistic pathogen. Nature 406:959-964. [PubMed]
47. Tan, M.-W., S. Mahajan-Miklos, and F. M. Ausubel. 1999. Killing of Caenorhabditis elegans by Pseudomonas aeruginosa used to model mammalian bacterial pathogenesis. Proc. Natl. Acad. Sci. U. S. A. 96:715-720. [PubMed]
48. Ternes, T. A. 1998. Occurrence of drugs in German sewage treatment plants and rivers. Water Res. 32:3245-3260.
49. Thacker, P. D. 2005. Pharmaceutical data elude researchers. Environ. Sci. Technol. 39:193A-194A. [PubMed]
50. Timmins, E. M., S. A. Howell, B. K. Alsberg, W. C. Noble, and R. Goodacre. 1998. Rapid differentiation of closely related Candida species and strains by pyrolysis-mass spectrometry and Fourier transform-infrared spectroscopy. J. Clin. Microbiol. 36:367-374. [PMC free article] [PubMed]
51. Walle, T., U. K. Walle, M. J. Wilson, T. C. Fagan, and T. E. Gaffney. 1984. Stereoselective ring oxidation of propranolol in man. Br. J. Clin. Pharmacol. 18:741-748. [PMC free article] [PubMed]
52. Winder, C. L., E. Carr, R. Goodacre, and R. Seviour. 2004. The rapid identification of Acinetobacter species using Fourier transform infrared spectroscopy. J. Appl. Microbiol. 96:328-339. [PubMed]
53. Winder, C. L., S. V. Gordon, J. Dale, R. G. Hewinson, and R. Goodacre. 2006. Metabolic fingerprints of Mycobacterium bovis cluster with molecular type: implications for genotype-phenotype links. Microbiology 152:2757-2765. [PubMed]
54. Yang, Y., B. Su, Q. Yan, and Q. Ren. 2005. Separation of naproxen enantiomers by supercritical/subcritical fluid chromatography. J. Pharm. Biomed. Anal. 39:815-818. [PubMed]

Articles from Applied and Environmental Microbiology are provided here courtesy of American Society for Microbiology (ASM)