|Home | About | Journals | Submit | Contact Us | Français|
Elevated concentrations of isovaleric, methylmalonic, and propionic acid are associated with impaired consciousness in genetic diseases (organic acidemias). We conjectured that part of the central nervous system depression observed in these disorders was due to anesthetic effects of these metabolites. We tested three hypotheses. First, that these metabolites would have anesthetic-sparing effects, possibly being anesthetics by themselves. Second, that these compounds would modulate glycine and GABAA receptor function, increasing chloride currents through these channels as potent clinical inhaled anesthetics do. Third, that these compounds would affect physical properties of lipids.
Anesthetic EC50’s were measured in Xenopus laevis tadpoles. Glycine and GABAA receptors were expressed in Xenopus laevis oocytes and studied using two-electrode voltage clamping. Pressure-area isotherms of 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) monolayers were measured with and without added organic acids.
Isovaleric acid was an anesthetic in tadpoles, while methylmalonic and propionic acid decreased isoflurane’s EC50 by half. All three organic acids concentration-dependently increased current through α1 glycine receptors. There were minimal effects on α1β2γ2s GABAA receptors. The organic acids increased total lateral pressure (surface pressure) of DPPC monolayers, including at mean molecular areas typical of bilayers.
Isovaleric, methylmalonic, and propionic acid have anesthetic affects in tadpoles, positively modulate glycine receptor fuction, and affect physical properties of DPPC monolayers.
General anesthesia is characterized by central nervous system depression leading to a variety of behavioral responses including immobility in response to noxious stimuli, amnesia, and unconsciousness, which are reversible upon discontinuation of the anesthetic (1). We have conjectured that certain metabolites, which are elevated in diseases associated with impaired consciousness, may have anesthetic properties (2–5). In previous studies, we focused on metabolites elevated in end-stage organ failure. We showed that beta hydroxybutyric acid was an anesthetic in tadpoles, and that ammonia reduced isoflurane minimum alveolar concentration (MAC) by 60% in rats. Both compounds modulated the function of particular anesthetic sensitive channels in a manner qualitatively similar to volatile anesthetics (2, 5).
This study focuses on anesthetic properties of organic acids. Organic acidemias are genetic metabolic diseases arising from mutations in genes coding for enzymes involved in amino acid metabolism (6). Isovaleric acidemia is caused by a mutation in the gene coding for isovaleryl CoA dehydrogenase (7). Methylmalonic acidemias are caused by mutations leading to a deficiency of methyl malonyl CoA (8). Propionic acidemia is caused by mutations in propionyl-CoA carboxylase (9, 10). Severe forms of all three disorders are associated with lethargy and unconsciousness.
We hypothesized that part of the central nervous system depression observed in these disorders was due to anesthetic effects of metabolites. Our reasoning was as follows. We noted that interfacial activity (the accumulation of a compound at interfaces, defined as the boundary between two immiscible fluid phases, such as the bilayer/water interface) has been shown to be a necessary condition of anesthetic action for a wide range of volatile anesthetics via molecular dynamic simulations and nuclear magnetic resonance spectroscopy, which showed that volatile anesthetics accumulated at interfaces but volatile compounds that were not anesthetics did not accumulate at interfaces (11–17). We reasoned that interfacial activity may also be a sufficient condition, and if so then other interfacially active compounds may have anesthetic properties. Our study showing that surfactants modulate receptor function in a manner similar to inhaled anesthetics supported this hypothesis (3).
In addition, we noted that a mechanism by which anesthetics affect lipid physical properties which are mechanically coupled to ion channel function has been proposed (18), and this could explain anesthetic effects of interfacially active molecules. This theory notes that because the membrane bilayer is in contact with water, there is tension at the bilayer/water interface. Interfacial tension decreases the area per lipid in the membrane owing to the free energy cost of increasing the area of membrane in contact with water. Opposing this is the entropic drive toward random acyl chain conformations, which increases the area per lipid molecule. As a result, ion channels will see depth-dependent laterally directed pressures (stresses) trying to squeeze the channel shut in some places, and pull it open in others. When an ion channel changes between its closed and open conformation, it will do work against the stresses imposed by the bilayer. This work contributes to the free energy difference between the open and closed conformations of the channel. Normally, this effect is not apparent because channels are adapted to the stresses on them. Volatile anesthetics, and other interfacially active compounds, will affect the packing of lipids at the bilayer/water interface, thereby altering the lateral pressures there (and, indirectly, throughout the bilayer). Channels will accordingly do a different amount of work in going between their closed and open states when volatile anesthetics, and other interfacially active compounds, are present, because the channels are doing work against a changed lateral pressure distribution. This provides a mechanism by which interfacially active molecules could have anesthetic actions. We expected that organic acids would interact with the interfacial region of the bilayer, and for the preceding reasons might have anesthetic effects.
We tested three hypotheses about isovaleric, methylmalonic, and propionic acid. First, that they would have anesthetic-sparing effects, possibly being anesthetics by themselves. Second, that these compounds would modulate glycine and GABAA receptor function, increasing chloride currents through these channels as potent clinical inhaled anesthetics do, and that they would have a reduced modulatory effect on glycine and GABAA receptors containing mutations which reduced the modulatory effect of isoflurane and ethanol. Third, that these compounds would affect physical properties of lipids. We tested this hypothesis by applying isovaleric, methylmalonic, and propionic acid to lipid monolayers (19) while measuring the change in surface pressure (total lateral pressure) of the monolayer. We used 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) monolayers because they are particularly well characterized and the phosphocholine headgroup is the most common headgroup in eukaryotic cell membranes (20, 21).
Studies on animals were approved by the institutional animal care and use committee of the University of California, San Francisco.
Wild-type and mutant α1(S270I)β2γ2s GABAA receptor and wild-type and mutant α1(S267I) glycine receptor clones were a gift of Professor R. Adron Harris (University of Texas, Austin). The cDNAs encoding GABAA receptor α1 wild-type and mutant subunits were cloned into a pBK-CMV vector. cDNAs encoding GABAA receptor β2 and γ2s and glycine receptor α1 wild-type and mutant subunits were cloned into a pCIS2 vector. Collagenase type 1, for defolliculating oocytes, was purchased from Worthington Biochemical Corporation (Lakewood, NJ). The propionic acid, methylmalonic acid, isovaleric acid and other reagents were purchased from Sigma-Aldrich (St. Louis, MO).
A female Xenopus laevis frog was anesthetized in 0.35% tricaine methane sulphonate solution (w/v). A portion of an ovary was surgically removed and defolliculated by gentle rotation in OR2 buffer (82.5 mM NaCl, 2.0 mM KCl, 1 mM MgCl2, 5 mM HEPES, pH 7.5) containing 500 U/ml collagenase type 1 for 1 to 1.5 h at room temperature. The individual oocytes were washed in OR2 buffer 8 to 10 times, and maintained in modified Barth buffer (88 mM NaCl, 1 mM KCl, 2.4 mM NaHCO3, 0.82 mM MgSO4, 0.33 mM Ca(NO3)2, 20 mM HEPES, 0.41 mM CaCl2, 5 mM sodium pyruvate, pH 7.5) containing 100 U/ml penicillin, 100 mg/ml streptomycin and 50 mg/ml gentamycin at 15°C.
2.5 ng cDNA of the glycine receptor α1 or mutant α1 (S267I) subunit, or a total of 2.5 ng cDNA of the GABAA receptor wild type or mutant (in a 1:1:6 ratio of α:β:γ subunits) were injected into the Xenopus oocyte nuclei via blind injection into the animal pole of the oocytes. Oocytes were kept in Barth’s buffer at 15°C and studied 1 to 2 days after cDNA injection.
Two-electrode voltage clamping was performed following previously described methods (5,37). EC5 concentrations of glycine, and EC10 concentrations of GABA, were used in these studies. For each oocyte, stable inward currents in response to application of agonist were verified by application of agonist for 20 seconds followed by a 5 minute washout, repeated three times. Each study compound was then applied to oocytes for 100 seconds, followed by coapplication of agonist(s) with the study compound for 20 seconds. Return to baseline response to agonist after washout of the test compound for 5 minutes was confirmed.
Anesthetic potency was tested in tadpoles one week after fertilization using standard methods (24). Tadpoles were placed in 0.1× Marc’s modified ringer’s (MMR) (10mM NaCl, 0.18mM KCl, 0.2mM CaCl2, 0.1mM MgCl2, 0.5 mM HEPES, adjusted to pH 7.0) with the drugs under study for 1.5 hours, then prodded on their side for thirty seconds with a glass rod, or until they moved, whichever occurred first. Tadpoles that responded to any amount of prodding were classified as moving. Tadpoles that did not respond to prodding were classified as non-moving. For tadpoles tested with propionic or methylmalonic acid, 0.3% isoflurane was bubbled into the 0.1× MMR prior to testing. Equilibration of the buffer with isoflurane was confirmed by gas chromatography. Non-moving tadpoles were removed and placed in 0.1X MMR for recovery; only data for animals that recovered were analyzed.
Isotherms were measured using a KSV Langmuir-Blodgett minitrough (Monroe, CT). The Teflon trough and Delrin barriers were cleaned prior to use with HPLC grade butyl acetate followed by ethanol and MilliQ water. The surface pressure balance used a platinum Wilhelmy plate which was cleaned with the same procedure as above and then heated in a flame before use. MilliQ water, with or without organic acid, was used for the subphase. The absence of contaminants in the subphase which affected surface pressure was confirmed by measuring surface pressure during compression of the barriers prior to spreading the monolayer.
The spreading agent was HPLC grade chloroform. DPPC (Avanti Polar Lipids, Alabaster, AL) at a concentration of 1 mg/ml was applied to the surface of the subphase using a Hamilton syringe. Once the spreading agent had fully evaporated (15 to 20 minutes), the monolayer was compressed at a rate of approximately 2 Å2/molecule/minute. A Wilhelmy balance recorded the resulting surface pressure every second. Two isotherms were obtained for each concentration of each acid, averaged, and compared to a reference isotherm for pure DPPC.
Change in currents was calculated in oocytes clamped at −80mV during drug and agonist coadministration vs. agonist alone. Data for oocytes are presented as mean ± SE. Statistical significance was determined using Student’s t-test. The fraction of tadpoles not moving as a function of drug concentration was analyzed by nonlinear regression to a Hill equation. EC50s and Hill coefficients were calculated and used to compare the curves. The surface pressure at a mean molecular area of 60 Å2 (similar to that found in bilayers (25)) was calculated at each concentration of organic acid. Linear regression was performed to determine if there was an increase in surface pressure with organic acid concentration. P<0.05 was considered significant.
There was a concentration dependent increase in currents through wild-type α1 glycine receptors for all three organic acids (Fig 1A). Propionic acid decreased currents at 1.25 mM.
10 mM methylmalonic acid and isovaleric acid produced a significantly larger increase in currents through wild-type α1 glycine receptors than in mutant α1 (S267I) glycine receptors, as seen with isoflurane and ethanol (26). There was a greater enhancement of the wild-type α1 glycine receptor compared to the mutant α1 glycine (S267I) receptor by 10 mM propionic acid (Fig 1B), but this result did not reach statistical significance.
Representative current tracings are in Fig. 1C.
10 mM methylmalonic acid (MMA) and isovaleric acid (IVA) produced only small increases in current through GABAA α1β2γ2s receptors. There was no significant effect of 10 mM propionic acid. The change in current was smaller, but nevertheless significant, in mutant α1(S270I)β2γ2s receptors for isovaleric and methylmalonic acids. Propionic acid inhibited currents through the mutant receptor. See Table 1.
Isovaleric acid was an anesthetic. Its EC50 ± se was 44.8 ± 3.7 mM (bath concentration), with a Hill number n = 4.1 ± 1.2. Both methylmalonic acid and propionic acid had to be coadministered with isoflurane or the tadpoles would not recover over the concentration range that prevented movement. When coadministered with 0.3% isoflurane (a concentration equal to half isoflurane’s EC50 (22)), 75.0 ± 9.1 mM propionic acid (Hill number n = 1.4 ± 0.4), or 149.6 ± 2.1 mM methylmalonic acid (Hill number n = 12.4 ± 2.0) had to be present in the bath to reversibly anesthetize the tadpoles. See Fig 2.
All three organic acids increased surface pressure of DPPC monolayers at 23.1°C (Fig 3 panels A, B, and C) including at a mean molecular area of 60Å2, which is approximately that obtained in lipid bilayers (25). The slope of the regression lines for pressure vs concentration at 60 Å2 were 0.22 ± 0.05 (p = 0.018), 0.62 ± 0.07 (P < 0.001), and 0.19 ± 0.02 (P < 0.001) milliNewton/(meter*millimolar) for isovaleric, methylmalonic, and propionic acid, respectively. See Fig 3D.
Our results show that isovaleric acid is ananes the ticintad poles. Methylmalonic and propionic acid have anesthetic-sparing effects in tadpoles. All three compounds show a concentration-dependent enhancement of homomeric α1 glycine receptor function at concentrations that are observed in disease (up to 5mM ) (27, 28) as well as at the higher concentrations we explored because of the anesthetic potential of these compounds. Minimal effects were observed on α1β2γ2s GABAA receptors.
To determine whether the allosteric effects were similar to those of volatile anesthetics, we applied the organic acids to mutant α1 glycine and α1β2γ2s GABAA receptors that were less sensitive to the modulatory effect of isoflurane and ethanol than wild type receptors (3). With the exception of propionic acid, which had a decreased modulatory effect on the glycine receptor that was not statistically significant, all of these compounds had a smaller effect on the mutant compared to the wild type channels. This is consistent with the hypothesized shared mechanism of action with isoflurane and ethanol.
We conjectured that the organic acids exerted their anesthetic effects indirectly, by modulating lipid properties. Measurement of surface pressures of lipid monolayers has been performed for over a century (29–31). We tested whether these compounds affected lipid properties by measuring pressure-area isotherms of DPPC monolayers with and without added organic acid. We used monolayers because their physical chemistry has been extensively studied (19, 20), the composition of the monolayer can be controlled, and its properties varied (in particular the mean area per molecule). For the pure DPPC monolayer, the shape of the isotherm can be explained as follows (Fig 4). At high mean molecular areas DPPC molecules interact weakly, forming a so-called gaseous phase. With compression, the molecules eventually interact to form a liquid expanded phase, at which point surface pressure rises. Further compression leads to the formation of a second, rigid, liquid condensed phase. The appearance of this second phase leads to the loss of a degree of freedom (a consequence of Gibb’s phase rule), with the result that surface pressure becomes constant in the two phase coexistence region and a plateau is observed. Continued compression produces a monolayer that is entirely in the liquid condensed phase. This rigid phase is relatively incompressible, and surface pressures rise rapidly. Further compression eventually leads to collapse of the monolayer. A molecular dynamic simulation shows the structure of the monolayer at various points on the isotherm (20).
Addition of each of the organic acids to the water subphase shifted the isotherm to higher pressures, and changed the shape of the isotherm. Increases in surface pressure of monolayers have been reported for volatile anesthetics (32–34). Surfactants can also shift DPPC pressure-area isotherms (35). In this study, the shift to higher surface pressures was most pronounced for methylmalonic acid, and least for propionic acid. In all cases these effects on the isotherms result from incorporation of the acids into the monolayers, which also changes the phase behavior of the DPPC. Of particular relevance, the acid was incorporated into the lipid monolayers at mean molecular areas characteristic of bilayers (fig 3D) of 60 A2 (25). Only at very low mean molecular areas (high surface pressures) were the acids squeezed out of the monolayer, with the isotherms approaching that of a pure DPPC monolayer.
That the organic acids interact with lipids does not rule out the possibility that these compounds may modulate ion channel function by binding directly to the channel protein. It does, however, open the door to a lipid-mediated mechanism for some of their clinical effects. The results are consistent with, but do not prove, that membrane stresses can account for the effect of the organic acids on receptor function.
Although the organic acids produce anesthetic or anesthetic-sparing effects in tadpoles, we do not know the concentrations in the tadpole that produce this effect. Our results only show that these compounds have anesthetic effects. Because we expected the animals would rapidly metabolize the organic acids, these drugs were introduced into aquarium water, which provided a large reservoir for the drug. Since only the unionized form of the acid should cross epithelial barriers, this acid is only a small fraction of the total concentration of the metabolite, and any metabolite entering the tadpole would be quickly metabolized, we expect that the concentrations of organic acid in the tadpoles are much less than the concentration we needed to apply in the aquarium.
Like ammonia (2) and beta hydroxybutyric acid (5), isovaleric acid, methylmalonic acid, and propionic acid are all endogenous compounds. Our results show that all are reversible central nervous system depressants. Why would an organism respond to an endogenous compound in this manner? Responding to an endogenous compound with central nervous system depression should be harmful and, hence, selected against. Yet, this does not seem to have happened. We have speculated that the basic mechanism accounting for this response may have arisen in one-celled organisms responding to interfacially active compounds in the environment (4). The observation that inhaled anesthetics (and surfactants (3)) enhance currents through inhibitory receptors, and inhibit currents through excitatory receptors, seems an odd physiologic response for a neuron. However, for a one-celled organism, this response makes sense: only this pattern of actions on ion channels will protect a one-celled organism’s transmembrane potential, by limiting entry of positive charges into the cell, in response to interfacially active compounds which are ubiquitous in the environment and which can modulate channel function through their effect on bilayer properties. By responding in this manner, anesthetic-sensitive ion channel architectures established in microbes could be passed on, ultimately to their descendants in multicellular organisms. It has been suggested that there is also ongoing selection for this response in animals (36).
We have identified twelve new modulators of glycine receptor function over the past two years. This was accomplished primarily using as guides the known interfacial activity of surfactants (37) and their effects on monolayer properties (35), the predicted effect of neurotransmitters on bilayer properties (36), molecular dynamic simulations (38) and radioligand binding studies (39) which show that various amino acids adsorb onto membranes. These modulators include two anionic surfactants (3), two cationic surfactants (3), one zwiterionic surfactant (3), a plasticizer (40), a neurotransmitter (GABA) (41), and five metabolites including the three organic acids reported here (ammonia (2), beta hydroxybutyric acid (5), isovaleric acid, methylmalonic acid, and propionic acid). We expect that several other small, interfacially active, endogenous molecules elevated in metabolic diseases may also be allosteric modulators of ion channel function.
In summary, we have shown that isovaleric acid, methylmalonic acid, and propionic acid have anesthetic effects in animals, positively modulate glycine receptor function, and affect pressure-area isotherms of DPPC lipid monolayers. These observations may provide mechanistic insight into the central nervous system depression observed in organic acidemias, and identify new structures with anesthetic properties.
This work was supported in part by NIGMS R01 GM069379
Reprints will not be available