Concentrations of Na+, K+ and Ca2+ in the growth medium were varied within limits normally found in vivo to determine how cation concentrations affect the sensitivity of ruminal bacteria to the ionophores, monensin (a Na+/H+ and K+/H+ exchanger) and tetronasin (Ca2+/H+). High [Na+] (172 mM cf. 137 mM in control medium) enhanced the efficacy of monensin towards Eubacterium ruminantium 2388, Streptococcus bovis C277, Lactobacillus casei LB17 and Prevotella albensis M384. High [K+] (35 mM cf. 19 mM) alone caused a decreased potency of both ionophores, except with L. casei. Added Ca2+ (7.4 cf. 2.8 mM) increased the potency of tetronasin when [Na+] was low. High [Na+] alone also potentiated the efficacy of tetronasin. Monensin caused intracellular [Na+] and [K+] to be decreased in the most sensitive of these organisms, E. ruminantium, whereas only intracellular [Ca2+] fell with tetronasin. The changes were small; however, Δp fell by only 20 mV after 2 h when ionophores caused immediate cessation of growth. ATP concentrations fell by 77% and 75% with monensin and tetronasin, respectively. Thus, altering cation concentrations might be used to potentiate the efficacy of ionophores, by increasing the rate of energy expenditure to maintain ionic homoeostasis in sensitive bacteria.
Monensin and tetronasin are feedlot ionophores that improve feed efficiency in cattle (Goodrich et al., 1984; Bartle et al., 1988). Although banned in Europe since 2006, they remain in widespread use elsewhere in the world, and research on their efficacy and mode of action continues (Dubuc et al., 2009; Felix & Loerch, 2011; Packer et al., 2011), particularly in the context of their ability to lower methane emissions (Martin et al., 2010). Their nutritional effects are due largely to changes in the fermentation stoichiometry and the metabolism of dietary nitrogen by ruminal microorganisms (Bergen & Bates, 1984; Russell & Strobel, 1989; Duffield et al., 2012). These changes arise partly from the elimination of many Gram-positive bacteria (Chen & Wolin, 1979; Henderson et al., 1981; Nagaraja & Taylor, 1987; Newbold et al., 1988) and partly from adaptations which resistant Gram-negative bacteria undergo when grown in the presence of ionophores (Morehead & Dawson, 1992; Newbold et al., 1992; Callaway & Russell, 1999).
There has been much speculation about the molecular mode of action of feedlot ionophores, mainly by analogy with the action of ionophores on nonruminal species of bacteria (Bergen & Bates, 1984; Russell, 1987; Russell & Strobel, 1989). How ionophores affect ruminal bacteria has important implications for the possible enhancement of their potency in vivo by altering the dietary content of the cations that they translocate (Rumpler et al., 1986; Chirase et al., 1987; Dawson & Boling, 1987; Schwingel et al., 1989). This strategy may be particularly relevant to tetronasin, because it has a much greater affinity for divalent, particularly Ca2+, than monovalent ions, in contrast to other feedlot ionophores, including monensin and lasalocid (Grandjean & Laszlo, 1983). Ca2+ ions are present at much lower concentrations (0.7–11.2 mM) than Na+ (77–157 mM) or K+ (22–68 mM) in the rumen (Durand & Kawashima, 1979); therefore, it seems possible that the potency of an ionophore that carries Ca2+ ions may be more readily enhanced than those that carry the more abundant monovalent ions.
The aim of the experiments described in this paper was to determine how varying the ionic composition of the medium affects the toxicity of monensin and tetronasin to selected species of ruminal bacteria and ion gradients in sensitive bacteria.
Materials and methods
Organisms and growth media
Prevotella albensis M384 (DSM 11370), Lactobacillus casei LB17 and Streptococcus bovis C277 were isolated from the rumen of sheep and are maintained in the culture collection at the Rowett Institute. Eubacterium ruminantium 2388 was originally obtained from the National Collection of Dairy Organisms, Reading. The liquid form of general-purpose, ruminal fluid–containing medium 2 of Hobson (Hobson, 1969) was used as the basal medium for growth experiments with all four bacteria. The C sources contained in this medium are glucose, maltose, cellobiose and lactate. Modifications to the mineral content were made by adding more K+ as phosphate salts and Na+ and Ca+ as chloride salts. The final concentrations of the cations in the control and amended media, respectively, were as follows: Na+, 137 and 172 mM; K+, 19 and 35 mM; Ca2+, 2.8 and 7.4 mM. In experiments to determine Δp and ion gradients in E. ruminantium, cation concentrations in the medium were 19 mM K+, 149 mM Na+ and 2.8 mM Ca2+. Media were prepared, and cultures were maintained, under O2-free CO2. Growth and incubation temperature was 39 °C.
Influence of ionophores and mineral cations on growth
A fresh overnight culture was used to inoculate (7%, v/v) media in Hungate tubes to which ionophores had been added in ethanolic solution (1 μL mL−1) before autoclaving. The concentration of ionophores was serially doubled in these tubes, as described previously (Newbold et al., 1988). Growth was measured by optical density at 650 nm after 48 h. The toxicity of the ionophore was assessed by determining the concentration of ionophore at which growth was inhibited by 50% (IC50).
Ion gradients, protonmotive force and ATP pools in E. ruminantium
Tetronasin or monensin was added to late-exponential phase cultures of E. ruminantium or cultures that had been in stationary phase for 30 h as ethanolic solutions at 0.064 and 0.256 μg mL−1. Ethanol (1 μL mL−1) was added to control incubations. Intracellular pH was determined 2 h after the addition of ionophore by the distribution of radiolabelled benzoic acid (Rottenberg, 1979). Culture (1 mL) was incubated under CO2 with [carboxy-14C] benzoate (0.25 μCi, 22 mCi mmol−1) and 3H2O (2.5 μCi, 125 μCi mL−1) for 10 min, then centrifuged at 12 000 g for 3 min. The supernatant was removed, and radioactivity in cells and supernatant was counted by liquid scintillation spectrometry. Internal pH was calculated from the distribution of 14C and 3H between the pellet and the supernatant. The accumulation of benzoic acid in E. ruminantium was abolished by pretreatment of the cells with 10 μM tetrachlorosalicylanilide, a protonophore (Hamilton, 1968), suggesting there was little or no active uptake or binding of benzoic acid by the cells. The chemical potential gradient (ZΔpH) generated by the pH gradient across the cell membrane was calculated from the Nernst relationship: Z = 2.3 RT/F or 62 mV at 39 °C. Intracellular volume was calculated using a separate aliquot of culture. One millilitre of culture was incubated with 3H2O (7 μCi, 125 μCi mL−1 and hydroxy [14C] methyl inulin (0.7 μCi, 11.1 mCi mmol−1) for 10 min, before centrifuging as before. The distribution of 14C-inulin and 3H2O in the pellet and the supernatant allowed the exclusion volume of inulin compared to H2O to be calculated and hence the intracellular volume (Rottenberg, 1979).
The electrical potential (Δψ) was calculated from the uptake of the lipophilic cation [phenyl-14C]tetraphenylphosphonium bromide (TPP+). One millilitre of culture was incubated under CO2 with TPP+ (0.05 μCi, 31 mCi mmol−1) and 3H3O (0.5 μCi, 16 μCi mL−1), then centrifuged and counted as before. Δψ was calculated from the distribution of TPP+ between the intra- and extracellular space (Rottenberg, 1979). The total transmembrane potential (Δp) was calculated as Δp = Δψ − ZΔpH. Nonspecific uptake/binding of TPP+ was corrected by subtracting the apparent uptake in cells that had been treated with toluene (1% v/v, 1 h).
Intracellular K+, Na+ and Ca2+ concentrations were measured in cells that had been centrifuged and resuspended in 25% TCA, then diluted in deionized water. Hydroxyl [14C] methyl inulin (0.7 μCi mL−1, 11.1 mCi mmol−1) was added to the cultures before centrifugation to allow corrections to be made for extracellular medium trapped in the cell pellet. Na+ and K+ were analysed by atomic emission spectrometry on a Pye Unicam SP9 atomic absorbance spectrometer, while Ca2+ was determined by atomic absorbance on the same instrument.
ATP pools were measured by a luciferase method (Wallace & West, 1982) in cells 2 h after the addition of the ionophores. Protein was determined using Folin reagent (Herbert et al., 1971).
Tetronasin was a gift from Coopers Animal Health Limited, Berkhamsted, Herts. Monensin was from Sigma. TCS and TPP were gifts from I.R. Booth, University of Aberdeen. [Carboxy-14C]-benzoate was from New England Nuclear, Stevenage, Herefordshire. All other radiochemicals were from Amersham.
Effect of changing the ionic composition of the medium on the potency of tetronasin and monensin
The potency of monensin and tetronasin against E. ruminantium, S. bovis, P. albensis and L. casei was determined by inoculating bacteria into media in which the concentration of ionophore was serially doubled. When cell density after 48-h growth was plotted against concentration, a sigmoidal curve was obtained. The potencies of the ionophores, when added alone or in combination with added Na+, K+ or Ca2+, were compared by determining the concentration at which cell density after 48-h incubation decreased by 50% (IC50; Table 1).
Table 1. Effect of varying the ionic composition of the growth medium on the toxicity of monensin and tetronasin to selected species of ruminal bacteria
The clearest and most consistent effect was that increasing the concentration of Na+ increased the potency of both ionophores. The concentration of monensin required to inhibit bacterial growth by 50% was, on average, 35% lower on high compared to low-Na+ medium. With E. ruminantium, S. bovis and P. albensis, the effect of adding Na+ was greatest when K+ was high. K+ alone tended to protect these bacteria from monensin, while the opposite was true with L. casei. Ca2+ had little influence on the potency of monensin, except with L. casei at low [Na+] and [K+] and S. bovis at high [Na+].
Altering the ionic composition of the medium had little influence on the potency of tetronasin with E. ruminantium (Table 1), but with the other bacteria increasing Ca2+ alone enhanced potency by more than 100%. K+ increased the potency of tetronasin with L. casei but protected P. albensis and S. bovis. When increased cation concentrations were combined, the Ca2+ effect was dominant with L. casei, and Ca2+ remained effective in enhancing potency with these three species in the presence of high [K+]. However, the effects of combining high Na+ and Ca2+ were more complex and species dependent: high [Na+] when [Ca2+] was high did not affect the sensitivity of S. bovis, while the combination was less effective than either cation alone with P. albensis.
Ion gradients and protonmotive force in E. ruminantium
The effects of monensin and tetronasin on protonmotive force and ion gradients were determined with late-exponential phase and also with cells that had been in stationary phase for 30 h. The results were similar for both cultures. However, Ca2+ analysis was performed only on stationary-phase cells (Table 2). The concentrations of ionophores used in this experiment were 0.256 μg monensin and 0.064 μg tetronasin mL−1, concentrations sufficient to cause severe inhibition of growth (Table 1). When these concentrations of monensin and tetronasin were added to exponentially growing E. ruminantium, growth ceased within 30 min (results not shown).
Table 2. Influence of monensin and tetronasin on the protonmotive force and intracellular ion concentrations in Eubacterium ruminantium
| (mM)||149|| ||149|| ||149|| |
| (mM)||19|| ||19|| ||19|| |
| (mM)||2.8|| ||2.8|| ||2.8|| |
|ATP (nmol mg protein−1)||2.06||0.068||0.47||0.011||0.51||0.025|
Eubacterium ruminantium had an internal volume of 3.4 ± 0.47 μL mg protein−1, as calculated from the inulin exclusion volume. This was not affected by the presence of either ionophore. Both monensin and tetronasin caused increases in the internal pH of E. ruminantium, coupled with a slight decrease in the electrical potential (Δψ) (Table 2). The total Δp therefore fell by about 20 mV with both ionophores over the 2-h incubation period.
Both ionophores resulted in the movement of Na+ and K+. Intracellular [Na+] and [K+] were approximately 4- and 12-fold higher, respectively, than external concentrations in control cells. Monensin caused efflux of both Na+ and K+. The change in electrical potential that would arise from the efflux of these cations, calculated from the Nernst equation, would be about 22 mV, that is, close to the observed change in Δp. Tetronasin had no influence on intracellular [Na+] or [K+], but caused the efflux of Ca2+. The changed [Ca2+] was equivalent to a decreased electrical potential of about 5 mV.
ATP pools were decreased by 77% and 75% in the presence of monensin and tetronasin, respectively (Table 2).
The selective toxicity of ionophores towards certain ruminal bacteria is a function of their ability to permeate the cell envelopes of some bacteria but not others (Chen & Wolin, 1979; Henderson et al., 1981; Bergen & Bates, 1984; Nagaraja & Taylor, 1987; Newbold et al., 1988; Russell & Strobel, 1989). Ionophores by definition translocate ions through biological membranes (Pressman, 1968), and this has been assumed to be their mode of action at the cellular level: ionophores that permeate the cell envelope will then disrupt transmembrane ionic gradients in accordance with their ion-translocating properties and cause toxicity. Monensin exchanges Na+ and, with a lower affinity, K+ for H+ (Pressman, 1968), and tetronasin facilitates Ca2+/H+ exchange across membranes (Grandjean & Laszlo, 1983). It therefore seems reasonable to suggest that the toxicity of these ionophores might be enhanced by altering the ionic composition of the medium (or diet), particularly of those ions for which the ionophores have highest affinity.
The bacterial species used in this study consisted of one Gram-negative and three Gram-positive species. Prevotella albensis belongs to normally the most numerous genus in the Gram-negative Bacteroidetes found in the rumen (Avgustin et al., 1997). Eubacterium ruminantium is a typical representative of the ruminal Firmicutes (Edwards et al., 2004). Streptococcus bovis and L. casei were chosen because of their important roles in the lactic acidosis spiral (Russell & Hino, 1985), a potentially fatal ruminal dysfunction for which monensin is prophylactic (Nagaraja et al., 1982). As found previously (Newbold et al., 1988), E. ruminantium was much more sensitive to both ionophores than the other bacteria, which is the reason that it was selected for further study.
Some potentiation of monensin and tetronasin was observed when cations were added to the growth medium of the four bacteria. Na+ ions were most potent in enhancing the effects of monensin, and increasing [K+] actually protected the bacteria slightly from monensin. These trends are therefore consistent with the model drawn up by Russell (1987), where it was postulated that monensin caused an efflux of K+ and an influx of Na+, both linked to the flux of H+ in the opposite direction. Not surprisingly in view of the low affinity of monensin for divalent ions (Pressman, 1968), Ca2+ had no influence on the potency of monensin. Ca2+ increased the efficacy of tetronasin, as would be predicted, but Na+ was almost as effective, despite the affinity of tetronasin for Na+ being < 5% that for Ca2+ (Grandjean & Laszlo, 1983).
In general, however, the effects of changing cation concentrations were relatively small and some could not be explained simply by the reported ion specificity of the ionophores. One possible cause of the small response was most likely the relatively small changes in concentration and therefore ionic gradient that were considered feasible, based on what might be achieved in vivo. The increase in [Na+] was only 26%, which would have a small effect on the transmembrane Na+ gradient. However, the change in [Ca2+] was substantial, a 2.6-fold increase, yet potentiation of tetronasin was still small. Several studies have been made previously, with some success, to apply the principle of cation enhancement of ionophores with ruminal bacteria and ruminal fermentation. Rumpler et al. (1986) found that adding Na+ to the diet of steers receiving monensin or lasalocid caused methane production to be decreased. This result was therefore consistent with the main mode of action of monensin as it is presently understood (Russell, 1987), but not with the K+/H+ exchange mechanism proposed for lasalocid (Schwingel et al., 1989). Increasing [K+] increased the potency of monensin towards ruminal bacteria in vitro (Dawson & Boling, 1987), which might not be expected to occur if the direction of induced K+ flux was outward, as in the Russell (1987) scheme. Chirase et al. (1987) observed a significant interaction between K+ and lasalocid in continuous cultures, but also Mg2+ and monensin or lasalocid despite the low affinity of these ionophores for divalent ions. Thus, although interactions undoubtedly occur between the concentrations of individual cations and the efficacy of ionophores, their magnitude and direction do not always appear to correspond to known ionophore specificity and the magnitude and direction of transmembrane ion gradients that have been measured in ruminal bacteria. Furthermore, the effects of combinations of cations and ionophores appeared to be species dependent, possibly indicating that transmembrane ion gradients are different in different rumen bacterial species.
The measurements of protonmotive force and ATP pools in E. ruminantium may help to explain some of these observations. Despite a rapid inhibition of cell growth, only relatively minor changes in intracellular cation concentrations were seen when monensin or tetronasin was added to the culture. Some efflux of Na+ and K+ was induced by monensin and Ca2+ by tetronasin. Undoubtedly, the measured ion concentrations in whole cells may not reflect the concentration of ions free in solution; cell walls, proteins and nucleic acids would be expected to bind Na+, K+ and Ca2+. Simple measurements of Na+, K+ or Ca2+ content can therefore not give a true indication of ion gradients. However, the measured values of TPP+ distribution also indicated that ionophores had only a minor influence on membrane potential. Some acidification of the cytoplasm occurred, but the total protonmotive force was only decreased by about 20%. In contrast, the ATP pool fell by 75%. It should be noted that the experiments were performed with late-exponential or stationary-phase cells to reflect the conditions that pertain predominantly in the rumen (Hobson & Wallace, 1982). It seems improbable that the mechanisms differ in more active bacteria, as in mid-exponential phase, although the magnitude of the gradients and pools may be different.
Russell and his colleagues have made similar observations with other species of ruminal bacteria. The high apparent intracellular concentrations of Na+ and K+ were similar to those measured in S. bovis (Russell, 1987). Ruminal bacteria have been described as mildly halophilic, based on their requirements of Na+ for growth (Caldwell et al., 1973; Caldwell & Hudson, 1974). The membrane potential fell by < 10% when monensin was added to S. bovis, although intracellular pH was affected to a greater extent (Russell, 1987). The protonmotive force of a ruminal Peptostreptococcus was unaffected by monensin, yet the ATP pool fell by two-thirds (Chen & Russell, 1989). It therefore appears that it is not the collapse of transmembrane ion gradients that causes the toxic effect of ionophores on intact bacteria, but the energy expenditure required to support the increased energy demand of homoeostatic mechanisms maintaining the gradients. Any extra demands induced by adding different cations may therefore have an influence on the efficacy of an ionophore, even if the ion is not translocated by the ionophore.
In conclusion, it may be possible to enhance the efficacy of ionophores by adding salts of mineral cations to the diet. However, the spectrum of antibacterial activity against different species, upon which ionophore depends for its nutritional effects, may well be different when the added cations are present, depending on the ion gradients present in different species. Thus, the nutritional effects of the ionophores (Chen & Russell, 1989) may not be the same at different cation concentrations. The present results also have implications for mechanisms by which ruminal bacteria may become resistant to ionophores. Adaptive resistance to ionophores involves changes in the permeability of the cell envelope (Newbold et al., 1992; Callaway & Russell, 1999), which may well affect changes in transmembrane ion gradients. One of the fears concerning the use of antimicrobials in livestock production is that transmissible resistance factors will arise and by transfer to human pathogens will render antibiotic therapy ineffective (Goodrich et al., 1984). However, there is no evidence that such resistance arises by exposure to ionophores such as monensin (Russell & Houlihan, 2003; Phillips, 2007).
This work was partly supported by Coopers Animal Health Limited. The Rowett Institute of Nutrition and Health is funded by the Rural and Environment Science and Analytical Services Division (RESAS) of the Scottish Government. We thank S. James for his support and advice.