Isolation and Characterization of Magbane, a Magnesium-Lethal Mutant of Paramecium

Corresponding author: Robin R. Preston, Department of Pharmacology and Physiology, MCP Hahnemann University, 245 N. 15th St., Mailstop 488, Philadelphia, PA 19102., robin.preston{at}drexel.edu (E-mail)

Communicating editor: S. L. ALLEN

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Discerning the mechanisms responsible for membrane excitation and ionic control in Paramecium has been facilitated by the availability of genetic mutants that are defective in these pathways. Such mutants typically are selected on the basis of behavioral anomalies or resistance to ions. There have been few attempts to isolate ion-sensitive strains, despite the insights that might be gained from studies of their phenotypes. Here, we report isolation of "magbane," an ion-sensitive strain that is susceptible to Mg²⁺. Whereas the wild type tolerated the addition of 20 mM MgCl₂ to the culture medium before growth was slowed and ultimately suppressed (at >40 mM), mgx mutation slowed growth at 10 mM. Genetic analysis indicated that the phenotype resulted from a recessive single-gene mutation that had not been described previously. We additionally noted that a mutant that was well described previously (restless) is also highly sensitive to Mg²⁺. This mutant is characterized by an inability to control membrane potential when extracellular K⁺ concentrations are lowered, due to inappropriate regulation of a Ca²⁺-dependent K⁺ current. However, comparing the mgx and rst mutant phenotypes suggested that two independent mechanisms might be responsible for their Mg²⁺ lethality. The possibility that mgx mutation may adversely affect a transporter that is required for maintaining low intracellular Mg²⁺ is considered.

INTEREST in the biological role of the magnesium ion has reached new levels in recent years, fueled largely by discoveries that intracellular free Mg²⁺ concentration ([Mg²⁺]_i) is both tightly regulated and can be modulated by extracellular ligands. Angiotensin II and vasopressin modulate intracellular Mg²⁺ levels in vascular smooth muscle cells (TOUYZ and SCHIFFRIN 1996 ), for example, while mitogens increase [Mg²⁺]_i in fibroblasts (ISHIJIMA et al. 1991 ). Although the cellular consequences and functions of such changes are still being investigated, several authors have suggested that abnormalities in the pathways that regulate [Mg²⁺]_i could be involved in the etiology of several diseases, including non-insulin-dependent diabetes mellitus (NIDDM) and hypertension (PAOLISSO et al. 1990 ; ARSENIAN 1992 ; WHITE and CAMPBELL 1993 ; ALTURA and ALTURA 1995 ; TOSIELLO 1996 ). Both conditions are associated with decreased cellular and serum Mg²⁺ levels and there are indications that development of NIDDM in a rat model can be delayed by a high Mg²⁺ diet. This suggests that a defect in intracellular Mg²⁺ handling could be an underlying cause (BALON et al. 1994 , BALON et al. 1995 ). However, despite the prevalence and severe consequences of NIDDM and hypertension, we still have little information about the mechanisms by which cells regulate [Mg²⁺]_i. We know that most cells maintain [Mg²⁺]_i at low levels against a steep electrochemical gradient and there is now substantial evidence suggesting that a Na⁺:Mg²⁺ exchanger may be responsible, but we still lack any molecular information about the pathways involved (BEYENBACH 1990 ; GUNTHER and EBEL 1990 ; FLATMAN 1991 ; MURPHY et al. 1991 ; ROMANI and SCARPA 1992 ).

In contrast, we have considerable information about mechanisms of Mg²⁺ homeostasis in prokaryotes. This is due largely to the work of Maguire and colleagues who have demonstrated three distinct Mg²⁺ transport systems encoded by the corA, mgtA, and mgtB genes (MONCRIEF and MAGUIRE 1999 ). The first of these was identified in Escherichia coli decades ago as a gene that conferred resistance to cobalt when disrupted by mutation (NELSON and KENNEDY 1972 ; PARK et al. 1976 ). It is now recognized to encode a transporter that is the major pathway for Mg²⁺ influx and efflux in most, if not all, prokaryotic species (SMITH and MAGUIRE 1995 ). The mgtA and mgtB genes encode P-type ATPases that take up Mg²⁺ from the extracellular medium (HMIEL et al. 1989 ; SNAVELY et al. 1991B ). When these genes are disrupted by mutation, cells become dependent on the presence of millimolar concentrations of Mg²⁺ in the growth medium for continued growth (MAGUIRE 1992 ).

We took a similar mutagenetic approach in attempts to identify pathways involved in regulating intracellular [Mg²⁺] in eukaryotes. Paramecium tetraurelia is a ciliated unicell that proved to be uniquely suited to investigations of [Mg²⁺]_i by virtue of its utilizing a Mg²⁺-specific influx pathway to regulate cell behavior (PRESTON 1990 , PRESTON 1998 ). Thus, by selecting for strains with abnormal swimming responses to Mg²⁺, we identified genes that either suppress or potentiate this influx pathway when mutated (xntA and Cha, respectively; PRESTON and KUNG 1994A ; PRESTON and HAMMOND 1997 ). Our ultimate goal is to characterize these genes at the molecular level through complementation cloning, a technique that has successfully identified several P. tetraurelia genes since its introduction 3 years ago (HAYNES et al. 1996 , HAYNES et al. 1998 , HAYNES et al. 2000 ; KELLER and COHEN 2000 ; KUNG et al. 2000 ). While the pathways that regulate this Mg²⁺ conductance are interesting in their own right, we also identify genes encoding Mg²⁺ transporters in Paramecium. Two basic approaches are available. The first is to isolate strains that are resistant to toxic ions that might be expected to enter the cell via a Mg²⁺ transport system. Ni²⁺, for example, is readily taken up by the corA Mg²⁺ transport system in bacteria (SNAVELY et al. 1991A ) and is highly toxic to many organisms, including Paramecium (ANDRIVON 1972 ). When put into practice, this approach yielded multiple lesions in the xntA gene (PRESTON and KUNG 1994B ). An alternate approach is to look for Mg²⁺-sensitive strains, the rationale being that cells lacking a functional transporter might be overwhelmed and killed by extracellular Mg²⁺ concentrations that are well tolerated by the wild type. However, the absence of automated techniques for replica plating Paramecium clones makes the isolation of sensitive mutants more labor intensive than resistants and hence has seldom been attempted. Two such mutants (K⁺ sensitives) were described by Cronkite almost 20 years ago (CRONKITE and BURG 1982 ; CRONKITE 1983 ), but none have been reported since. The potential importance of a Mg²⁺-lethal strain is great, however, and since Cronkite and colleagues showed the technique to be both feasible and practical, we attempted a similar search. Here we report the result, which is the successful isolation of a Mg²⁺-sensitive strain that we refer to as "magbane."

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Stocks and culture conditions:
Studies were conducted using P. tetraurelia, stock 51s, and the following mutants derived from this stock: d4-90 Paranoiac A (PaA/PaA; VAN HOUTEN et al. 1977 ); d4-91 fast-2 (cam¹¹/cam¹¹, formerly fna/fna; KUNG 1971 ; KINK et al. 1990 ); d4-150 Paranoiac C (PaC/PaC; VAN HOUTEN et al. 1977 ); d4-152 TEA-insensitive A (teaA/teaA; CHANG and KUNG 1976 ); d4-623 Dancer (Dn¹/Dn¹; HINRICHSEN et al. 1984 ); d4-644 k-shy A (ksA¹/ksA¹; EVANS and NELSON 1989 ); d4-646 paranoiac F (see SAIMI and KUNG 1987 ); d4-647 restless (rst/rst; RICHARD et al. 1985 ); d4-700 eccentric A (xntA¹/xntA¹; PRESTON and KUNG 1994B ); pa-711 Chameleon (Cha/Cha; PRESTON and HAMMOND 1997 ); and pa-717 magbane (mgx/mgx). The latter was isolated during the present studies. A trichocyst nondischarge mutation (nd6/nd6; LEFORT-TRAN et al. 1981 ) was used as a genetic marker in all of the crosses described below. There is no evidence for genetic linkage between nd6 and any of the mutations used here. All stocks were raised at 22–28° on either a wheat grass infusion (for behavioral and genetic testing) or C7 medium (for electrophysiological recording) inoculated with Enterobacter aerogenes as described (SONNEBORN 1970 ; PRESTON and HAMMOND 1998 ). The only significant differences between cells grown in the two media were cell size and final density.

Solutions:
All solutions contained 1 mM CaCl₂, 1 mM HEPES, 0.01 mM EDTA, pH 7.2 (referred to below as "K⁺-free solution"). Chloride salts of barium, magnesium, nickel, potassium, or sodium were added to this solution as required and at the concentrations stated. "Mg²⁺ solution" contained 5 mM MgCl₂ and 10 mM tetraethylammonium (TEA⁺) chloride, "Na⁺ solution" contained 10 mM NaCl, "Ba²⁺ solution" contained 6 mM BaCl₂, "30 K⁺ solution" contained 30 mM KCl, "1 K⁺ solution" contained 1 mM KCl, and "Ca²⁺ solution" contained 10 mM TEA⁺. "Resting solution" contained 4 mM KCl.

Mutagenesis and mutant selection:
Mutations were induced in nd6 mutant cells using N-methyl-N'-nitro-N-nitrosoguanidine, as described by KUNG 1971 . Following mutagenesis, cells were starved to induce homozygosity via autogamy, and then 4000 cells were transferred individually to 0.25 ml of growth medium in each well of a sterile, polystyrene, 24-well culture plate (Corning Inc., Corning, NY). Approximately 1600 of these clones subsequently died or grew poorly, leaving 2400 clones available for screening over a period of several months. The clones were fed every 2–3 weeks by suctioning out all but 50 µl of the contents of the wells using a pipette (flushed with boiling water to kill paramecia between wells) and then adding 800 µl of fresh culture medium. Mutant clones were maintained at 15° between feedings to slow growth. We screened all clones for sensitivity to 15 mM Mg²⁺, 20 mM K⁺, and 15 µM Ni²⁺, using similar techniques for each screen. We added 1 ml of ion-supplemented bacterized culture medium to each well of a set of screening plates numbered to match those containing the mutant cell clones. We then transferred 100 µl of each clone to the corresponding position on the screening plates and placed them at room temperature for either 24 hr (for K⁺ and Ni²⁺ screens) or 1 week (Mg²⁺). The wells were then examined under low-power magnification and their contents noted. These screens were repeated on two further occasions. After the final screen, clones that failed to thrive in ion-supplemented medium on at least two occasions were identified and transferred to culture tubes for further analysis. This procedure yielded 31 K⁺-sensitive clones, 31 Ni²⁺-sensitive clones, and 82 Mg²⁺-sensitive clones. Of the latter, 8 proved to have also been isolated by virtue of K⁺ (2 clones) or Ni²⁺ sensitivity (6 clones). All clones were then rescreened for sensitivity to K⁺ (0–25 mM), Ni²⁺ (0–20 µM), or Mg²⁺ (0–20 mM) and were tested behaviorally (see below).

Behavioral tests:
Ten to 20 cells were transferred from culture medium to resting solution and left undisturbed for 10–15 min. Individual cells were then selected with a micropipette and ejected forcibly into Mg²⁺, Na⁺, Ba²⁺, or 30 K⁺ solution. These solutions typically elicited repeated turning or backward swimming, the duration of which was recorded with a stopwatch.

Electrophysiology:
The techniques used to record the membrane currents of Paramecium under two-electrode voltage clamp have been described (PRESTON et al. 1992 ). Mg²⁺ currents were elicited from cells held at -30 mV in Mg²⁺ solution using intracellular capillary microelectrodes filled with 3 M CsCl. K⁺ currents were elicited from cells bathed in 1 K⁺ solution and clamped using electrodes filled with 3 M KCl. All data are presented as means (±SD), with the significance of differences between means calculated using a Student's t-test. Recordings were carried out at room temperature (22° ± 2°).

Ion dependence of cell growth:
Growth studies were conducted using wheat grass medium supplemented with NaCl, KCl, or MgCl₂ at the concentrations stated. At day 0, cells were added to 50 ml of culture fluid in a 250-ml conical flask to a final concentration of 50 cells/ml. Samples were removed from each flask at regular intervals thereafter and the number of cells was counted and averaged. Final cell density at stationary phase varied from one batch of culture to the next, probably as a reflection of the initial bacterial cell density. This made it difficult to pool growth data, so, although we may present original data from only one such growth experiment (see RESULTS), all experiments were repeated on three or more occasions.

Ni²⁺ resistance:
Twenty cells suspended in 10–20 µl of culture fluid were transferred to a glass well containing 1 ml of a 1 K⁺ solution supplemented with Ni²⁺ at the concentrations stated. After 2 hr, each well was examined using a dissecting microscope and its contents were scored for cell survival as described (PRESTON and KUNG 1994B ). iC₂ values were determined from plots of survival score against Ni²⁺ concentration and represent the amount of Ni²⁺ required to immobilize cells following the 2-hr exposure period.

Genetic analyses:
Standard techniques were used to establish genetic relationships between different strains of Paramecium (SONNEBORN 1950 ). Two homozygous strains were crossed by conjugation to yield heterozygous F₁ progeny. The F₁ were then starved to induce autogamy, producing an F₂ generation that was again homozygous at all loci. The nd6 mutation was used as a genetic marker in all crosses to ensure that cross-fertilization had occurred between two mating cells and that autogamy had been induced successfully in the F₁.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Isolation of ion-sensitive mutants:
We screened 2400 mutant cell clones for strains that failed to thrive in growth medium supplemented with 20 mM K⁺, 15 µM Ni²⁺, or 15 mM Mg²⁺. After excluding clones that grew poorly even under control conditions, we identified 8 that were specifically K⁺ sensitive, 1 that was Ni²⁺ sensitive, and 5 that were Mg²⁺ sensitive. These were then tested for behavioral responses to solutions that probe the functioning of several key ion currents, the rationale being that ion-sensitive or ion-resistant phenotypes frequently cosegregate with a conductance abnormality (SCHEIN 1976 ; SHUSTERMAN et al. 1978 ; PRESTON and KUNG 1994B ). Three of these (K⁺-sensitive P27 and Mg²⁺-sensitive M38 and M60) indeed showed enhanced responses to Mg²⁺, Na⁺, and Ba²⁺ solutions that could indicate a Ca²⁺-current defect, but these were not investigated further. Several of the remaining clones proved to have strong ion-lethal phenotypes. For example, whereas it required 22.5 mM K⁺ to kill wild-type cells by 48 hr, strain P13 died in only 5 mM K⁺ and strain P4 died in 7.5 mM K⁺. Strain N13 was unable to survive exposure to 2 µM Ni²⁺, whereas 14 µM Ni²⁺ was needed to kill the wild type. Strain N13 also had a distinctive behavioral phenotype; it swam in a greatly exaggerated helical path when transferred to resting solution. Finally, strain M16 was killed by <8 mM Mg²⁺, about one-half that required to kill the wild type under similar conditions. This mutant phenotype, which we refer to as magbane in recognition of its unique susceptibility to Mg²⁺ poisoning, is the focus of this report.

Death by magnesium:
Before quantifying magbane's sensitivity to Mg²⁺, we first investigated the effects of this cation on the wild type. Cells in logarithmic growth phase were added to fresh culture fluid containing various amounts of MgCl₂ (0–40 mM) to a final density of 50 cells/ml (day 0). In the absence of added Mg²⁺, cultures exhausted their food supply at around day 3 and reached stationary phase at day 4 (Fig 1A, solid circles). Modest Mg²⁺ concentrations (15 mM) had no significant effect on growth, but higher concentrations both slowed growth significantly and reduced the maximum cell density. When 40 mM Mg²⁺ was added to the culture medium, the cells usually underwent a few fissions but then died (Fig 1A, open diamonds). To quantify these effects, cell generation times were determined from the logarithmic growth phase of each culture (days 0–3). In the absence of added Mg²⁺, cell generation time was 9.3 hr (±0.4 hr, n = 5), but Mg²⁺ concentrations of 25 mM increased this time significantly. This is demonstrated in Fig 1C (solid circles), where the reciprocal of generation time is plotted against Mg²⁺ concentration. Magbane gene (mgx) mutation clearly made cells more susceptible to Mg²⁺ such that even 5 mM reduced maximum cell density and 25 mM inhibited growth entirely. Effects on generation time are shown in Fig 1C (open circles): Significant lengthening was obtained at 10 mM Mg²⁺ (from 9.6 ± 0.9 hr in control cultures containing no added Mg²⁺ to 14.7 ± 1.4 hr at 10 mM, n = 5).

View larger version (14K): In this window In a new window Download PPT slide   Figure 1. Growth of wild-type and mgx mutant cells in cultures supplemented with MgCl2. (A) Cells in logarithmic growth were added to fresh culture medium to a final density of 50 cells/ml on day 0. The medium was supplemented with MgCl2 as follows: 0 mM (•), 5 mM (), 10 mM (), 15 mM (), 20 mM (), 25 mM (), 30 mM (), or 40 mM (). Samples were removed at daily intervals thereafter and the number of cells was counted. Cultures were maintained at 23°. Data are representative of five experiments. (B) Growth rates of wild-type (•) and mgx () mutant cells in culture supplemented with concentrations of MgCl2 as indicated. Rates were determined from the exponential phase of growth and are expressed as the reciprocal of generation time. Data points are means ± SD of five determinations.

We also examined the sensitivity of mgx mutant cells to KCl and NaCl to determine whether the effects of increasing extracellular ion concentrations on cell growth were specific for Mg²⁺. Wild-type cultures typically showed reduced maximum cell densities and increased generation times at 25 mM KCl whereas 30 mM was lethal (not shown). Magbane cell cultures were indistinguishable from the wild type in this respect. The mutant also showed no significant difference from the wild type in sensitivity to NaCl. Both stocks tolerated up to 80 mM NaCl before densities were reduced and generation times slowed, and both succumbed completely at 100 mM.

Genetic analysis of mgx:
Magbane was crossed first to nd6, a recessive mutation that prevents trichocyst discharge and is used frequently as a genetic marker in studies of the relationships between various Paramecium mutants. While magbane was isolated originally as a strain that was unable to thrive or even survive for a week in 15 mM Mg²⁺, practicality required that we find a more expeditious means of identifying the mutant phenotype among the progeny of genetic crosses. Systematic testing suggested that 24-hr exposure to growth medium supplemented with 30 mM Mg²⁺ was a reliable way of differentiating magbane mutant cells from the wild type. Thus, 99% of wild-type clones (77 in total) continued to thrive when transferred to growth medium containing 30 mM Mg²⁺, whereas 100% of 365 magbane clones showed extensive death after 24 hr (i.e., 50% of cells originally transferred to the Mg²⁺-rich medium).

When magbane was crossed to nd6, the F₁ was wild type with respect to trichocyst discharge and Mg²⁺ survival. Starving the F₁ produced an F₂ through autogamy. The F₂ contained an 1:1:1:1 ratio of wild type, nondischargers, magbane, and double-mutant clones (56:44: 59:63, P > 0.1), suggesting that the mgx mutant phenotype results from a recessive single-gene mutation.

We next crossed magbane to a number of mutants with well-defined behavioral phenotypes, including pwA, pwB, Dn, teaA, cam¹, cam¹¹, PaA, paF, and KsA. The results of these crosses suggested that they were independent mutations. Analysis of double-mutant clones found them to be indistinguishable from magbane in terms of Mg²⁺ toxicity, whereas their behavior was indistinguishable from that of the previously characterized single-mutant parent. Three crosses were of interest, however, and they are described in more detail below.

Crosses to mutations that affect the Mg²⁺ current:
The conductance that allows for Mg²⁺-induced backward swimming in Paramecium is a significant pathway for Mg²⁺ entry into the cell. It is conceivable, therefore, that abnormal regulation of this conductance in magbane could cause intracellular Mg²⁺ levels to rise to toxic levels and thereby produce the lethal phenotype. Thus, we were interested to know if the Mg²⁺ conductance (I_Mg) functioned normally in magbane and also the relationship between this mutant and existing Mg²⁺ current mutants.

First, we determined the duration of backward swimming induced by Mg²⁺ solution, a behavior that is correlated with activation of I_Mg. Wild-type cells swam backward for 9 sec in Mg²⁺ and similar responses were recorded using magbane mutant cells. We also examined the two cell lines under voltage clamp to determine the magnitude and kinetics of I_Mg but found no significant differences between the two (not shown). Finally, we examined the sensitivity of magbane to Ni²⁺, a heavy metal that enters cells via the Mg²⁺ conductance pathway. It then immobilizes and poisons the cells in a time- and concentration-dependent manner that is determined, in part, by the magnitude of the Mg²⁺ conductance (PRESTON and KUNG 1994B ). The concentration of Ni²⁺ required to immobilize wild-type cells by 2 hr was 39 µM ± 10 µM (n = 5); similar values were determined using mgx mutant cells (51 ± 18 µM, n = 5).

We next examined the effects of manipulating I_Mg on sensitivity to Mg²⁺. Chameleon gene (Cha) mutation slows the rate at which I_Mg deactivates following excitation, thereby enhancing overall Mg²⁺ influx (PRESTON and HAMMOND 1997 ). It also appeared to slightly enhance Mg²⁺ sensitivity in our growth assays. Whereas wild-type cultures were usually capable of limited growth in 40 mM MgCl₂, Cha mutant cells were not (Fig 2A). We then mated Cha with mgx. The F₁ were wild type in terms of behavioral responses and sensitivity to Mg²⁺ while the F₂ produced through autogamy segregated in an approximately equal ratio of wild type, single-mutant parents, and double mutants. When the double mutant was monitored during growth in Mg²⁺-supplemented media, we found that it was more sensitive than either parent. Growth was slowed significantly at 10 mM Mg²⁺ and was inhibited fully at 20 mM (Fig 2A).

View larger version (19K): In this window In a new window Download PPT slide   Figure 2. Effects of mutational manipulation of Mg2+ and K+ permeability on survival in MgCl2-supplemented media. (A) Growth of wild type (•), mgx (), Cha (), and mgx;Cha (half solid/half open square) double-mutant cells in various concentrations of MgCl2. Double-mutant cells typically grow slowly compared with the wild-type or single-mutant parents, so generation times were calculated for each strain, the reciprocal was taken, and this value was normalized to controls (no added MgCl2). Data are means ± SD determinations from four experiments. (B) Growth of wild-type (•), mgx (), rst (), and mgx;rst (half solid/half open square) double-mutant cells in media supplemented with MgCl2. Points are means ± SD, n = 3.

We next crossed magbane to eccentric A, a mutant that fails to swim backward in Mg²⁺ because it lacks I_Mg (xntA; PRESTON and KUNG 1994B ) The xntA mutant parent was indistinguishable from the wild type in terms of Mg²⁺ sensitivity (not shown). The F₁ progeny from this cross showed normal behavioral responses to Mg²⁺ and readily survived a 24-hr incubation with 30 mM Mg²⁺. Starving the F₁ induced autogamy and produced an F₂ generation that segregated in an approximately equal ratio of wild type, eccentric, magbane, and double mutants. The double-mutant clones failed to respond behaviorally to Mg²⁺ as expected, but xntA mutation coincidentally reduced the effect of mgx mutation on growth. Thus, 15 mM Mg²⁺ was required to slow growth rates significantly in the double mutant compared with only 10 mM in the magbane single-mutant parent.

The effects of rst mutation on Mg²⁺ sensitivity:
Mutation of the restless gene (rst) causes cells to swim unusually fast under "resting" conditions, a phenotype resulting from inappropriate regulation of a Ca²⁺-dependent K⁺ conductance (see below). When we began analyses of the genetic relationship between rst and mgx, we were surprised to discover that rst mutant cells were also highly sensitive to Mg²⁺, as shown in Fig 2B. Magbane cells were typically able to survive the addition of 15 mM Mg²⁺ to the culture medium, although growth rates were slowed significantly. By contrast, some restless cultures failed to survive even in this concentration of Mg²⁺. This raised the possibility that magbane and restless were related genetically.

Before attempting to mate these two mutants, we first examined the responses of magbane to solutions that are used routinely to test for the rst mutant phenotype. Transferring wild-type cells to 10 mM Na⁺ caused backward swimming for 2–4 sec, whereas rst mutant cells swam fast forward with no evidence of even a weak reversal response. Magbane responded to this solution with backward swimming, much like the wild type. A second way of testing for the rst mutant phenotype was to transfer cells to K⁺-free solutions. Whereas wild-type cells survived for hours in K⁺-free solution, rst mutant cells died within 30 min due to excessive "bleeding" of K⁺ via the abnormal conductance pathway. Magbane produced intermediate results. Cells from logarithmic growth phase cultures readily survived 24 hr in K⁺-free solution, but when cultures became nutrient depleted and the cells starved, they died 2 hr after transfer to K⁺-free saline. Thus, Na⁺ responses proved to be the most reliable means of distinguishing the two mutant strains. Crossing mgx and rst yielded an F₁ generation that responded to 10 mM Na⁺ with repeated turns rather than continuous backward swimming. They were also sensitive to K⁺ withdrawal, with death typically following within 1 hr. They were also highly Mg²⁺ sensitive, with all clones showing complete or near complete death after 24 hr in 30 mM Mg²⁺. The F₁ were then starved to induce autogamy and individual cells were isolated and cloned for analysis of the F₂. The result was an approximate 1:1:2 ratio of wild-type:magbane:restless clones (55:58:93, P > 0.5). Three double-mutant lines were identified by randomly crossing clones that exhibited a rst mutant phenotype to the wild type. The double-mutant strains proved to be highly intolerant of Mg²⁺: Their growth was slowed by the addition of even 1 mM MgCl₂ to the growth medium whereas 5 mM proved fatal (Fig 2B). However, the fact that there were double mutants among the F₂ suggested that magbane and restless phenotypes result from mutations in two independent genes.

This result prompted a number of questions, foremost among which was the above-mentioned cause of the low-K⁺ and high-Mg²⁺ lethality of the mgx/+ ; +/rst heterozygotes. We had already found mgx to behave like a recessive allele with respect to the wild type. Work by others had reached a similar conclusion regarding rst (RICHARD et al. 1985 ), so we expected the heterozygote to be wild type in terms of Mg²⁺ survival and sensitivity to K⁺ withdrawal. In seeking an explanation for the mgx/rst phenotype, we crossed rst to the wild type and tested the behavior of the resultant F₁. The heterozygote responded to Na⁺ solution with repeated avoidance reactions, as reported previously (RICHARD et al. 1985 ). However, it also showed 100% mortality in both our 30 mM Mg²⁺ survival tests and upon KCl withdrawal. The ex-autogamous F₂ segregated normally (18:16 ratio of + to rst), confirming the veracity of the F₁.

Ca²⁺-dependent K⁺ currents in restless and magbane:
Although the analysis above suggested that restless and magbane mutants were unrelated genetically, their shared sensitivity to K⁺ withdrawal and high Mg²⁺ concentrations raised the possibility that the two mutations affect a single physiological process. The rst mutant phenotype was attributed to improper regulation of a Ca²⁺-dependent K⁺ current (RICHARD et al. 1986 ; PRESTON et al. 1990B ). In the wild type, this current typically activates slowly over a period of 400–800 msec during membrane hyperpolarization (see below). In restless cells, however, the current is both enhanced greatly and activates within a few milliseconds of hyperpolarization, causing uncontrolled K⁺ efflux and deepening of membrane potential (V_m) in solutions containing <4 mM K⁺. Thus, we first examined resting V_m in a solution containing low KCl (1 mM). The membrane potential of wild-type cells rested at -43 mV (±3 mV, n = 4). This compared with a V_m of -79 mV (±6 mV, n = 5) in rst mutant cells. V_m in magbane was -42 mV (±6 mV, n = 6), which was not significantly different from the wild type. We then placed cells under voltage clamp and examined K⁺ currents directly. The Ca²⁺-dependent K⁺ current abnormality in restless cells was best demonstrated using a series of step hyperpolarizations from -40 to -110 mV of gradually increasing duration (Fig 3A, top left). In the wild type, this voltage protocol elicits an inward current that contains contributions from a leak current(s), a Ca²⁺-specific current, a voltage-dependent K⁺ current, and the Ca²⁺-dependent K⁺ current (PRESTON et al. 1990A ). To gauge the contribution of the Ca²⁺-dependent current to this complex, we focused on the outward tail current elicited upon returning to -40 mV (Fig 3A). This tail represented K⁺ escaping from the cell via two K⁺ conductances in the brief period before the channels close. A voltage-dependent K⁺ current contributed a tail current that decays with a time constant of 2.5 msec, whereas the slower Ca²⁺-dependent component decayed with a constant of 20 msec (Fig 3A, arrows). This difference made the relative contribution of the two components to the tail easy to discern, despite the common charge carrier. When the amplitude of the Ca²⁺-dependent component was plotted as a function of the duration of the activating step, we noted that it required a >40 msec step to elicit this current in the wild type (Fig 3B, solid circles). It then gradually increased in amplitude toward a plateau of 3–4 nA at 500 msec. In restless, however, this current was already prominent following a 10-msec step to -110 mV and climbed rapidly in amplitude toward a peak of 16–17 nA at 40 msec (Fig 3B, open squares). Similar voltage-clamp protocols were applied to mgx mutant cells. The time course of Ca²⁺-dependent K⁺ current activation, as reflected in tail-current amplitudes, appeared similar to that of the wild type. A slowly decaying tail component was first observed following steps of >40 msec duration and its amplitude then climbed toward a plateau of 6–7 nA at 500 msec (Fig 3B, open circles). Although the mean tail amplitude appeared larger than that of the wild type, the difference was not significant. We also determined amplitudes of other K⁺ currents (Ca²⁺-dependent current upon depolarization and the two voltage-dependent K⁺ currents) in magbane and the wild type, but saw no significant differences between the two cell stocks (not shown).

View larger version (14K): In this window In a new window Download PPT slide   Figure 3. K+ currents in wild-type and mutant cells. (A) Cells were held at -40 mV using two-electrode voltage clamp in a solution containing 1 mM K+ . Membrane potential was then stepped to -110 mV to elicit the currents shown. The initial step was of 10-msec duration with subsequent steps each increasing by 15 msec. The voltage protocol is shown at top left. The steps elicit an inward current complex followed by an outward K+ tail current that is made up of a fast, voltage-dependent component and a slower, Ca2+-dependent component. The latter is indicated by arrowheads. In restless, the slow component is dominant, even following steps of short 10-msec duration. (B) Amplitudes of the slow, Ca2+-dependent K+ tail current (Itail) plotted against the duration of the step hyperpolarization used to elicit the tail current. Data are means ± SD from 7 wild-type (•), 10 magbane (), and 4 rst mutant () cells. For clarity of presentation, SD is given for one data point only but other values were similarly variable.

These studies failed to uncover any significant K⁺ conductance abnormality in magbane, suggesting that the Mg²⁺-lethal phenotype common to both this mutant and restless may result from unrelated cellular defects. In exploring this idea more fully, we hypothesized that perhaps the susceptibility of rst mutant cells to Mg²⁺ was related to the deepened V_m through an increase in the driving force for Mg²⁺ influx. If so, compensating for the V_m defect might simultaneously lessen sensitivity to Mg²⁺. We tested this by supplementing the culture medium with 10 mM KCl, a concentration that was shown previously to normalize V_m in restless (RICHARD et al. 1985 ). This had little effect on the growth of wild-type, magbane, or restless cell cultures (Fig 4, left). We then determined the effects of adding 15 mM MgCl₂ to the growth medium. The effects were very similar to that shown previously: Both magbane and restless cultures died, whereas the wild type was not affected significantly (Fig 4, center). We then tested the effects of simultaneously adding KCl and MgCl₂ to cultures. Growth of the wild-type cultures was reduced by 10% compared with controls whereas magbane again died in the face of high Mg²⁺. However, the addition of KCl to the culture medium permitted restless not only to survive exposure to MgCl₂ but also to actually thrive (Fig 4, right).

View larger version (45K): In this window In a new window Download PPT slide   Figure 4. Effects of KCl on survival of mgx and rst mutant cells in MgCl2-supplemented media. Wild-type, mgx, and rst mutant cells were added to cultures supplemented with 10 mM KCl (left), 15 mM MgCl2 (middle), or both KCl and MgCl2 (right). After 3 days, the number of cells in each culture was counted and normalized to the density of control cultures to which neither KCl nor MgCl2 had been added. Data are means ± SD of three determinations.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Magbane represents a new class of Paramecium mutant with a specific Mg²⁺-lethal phenotype. Evidence suggests that it results from a recessive, single-gene mutation that was not characterized previously, although these studies did find that one of the better-characterized behavioral mutants of P. tetraurelia is also highly sensitive to Mg²⁺.

Isolation of sensitive mutants:
The large numbers of behavioral and ion-resistant mutants of P. tetraurelia now available stand testament to the ease with which such phenotypes can be isolated. In contrast, the need to systematically test hundreds of individual mutant clones to identify ion-sensitive strains means that relatively few such lines have been isolated. Techniques for replica-plating Paramecium clones en masse that might help in this task are described (e.g., SONNEBORN 1970 ), but they are generally unreliable and hence have not been widely adopted. However, the isolation of potassium sensitives ks-1 and ks-2 (CRONKITE and BURG 1982 ; CRONKITE 1983 ) clearly demonstrated that such mutants were both readily obtainable and of great importance in furthering our understanding of intracellular ion handling in Paramecium, so we began sifting through thousands of individual mutant clones for a Mg²⁺-sensitive strain. In practice, such phenotypes appeared with a relatively high frequency. After testing 2400 clones, we identified at least 5 that appeared to have strong and specific ion-sensitive phenotypes, although only 2 of these have been examined rigorously to date. The frequency of such phenotypes among the mutant cell population reflects in part the severity of treatment with mutagen and perhaps also the complexity of ion handling in unicells, in terms of both the number of cellular pathways that are directly involved in ionic regulation and also the number of pathways that could affect homeostasis indirectly when disrupted.

Relationship between mgx and rst:
Of the various known behavioral mutants tested for susceptibility to Mg²⁺, only restless showed a significant sensitivity. Side-by-side trials of rst and mgx in high Mg²⁺ concentrations frequently ended with the latter being the sole survivor, suggesting that rst may be slightly more sensitive, although the two mutants appeared identical in normalized growth curves (Fig 2B). This raised the possibility that the two genes or their intracellular targets were related—or perhaps even identical. Further support for this idea came from finding that magbane died in low K⁺ solutions when stressed by starvation and that mating mgx and rst produced an F₁ generation that died in high Mg²⁺ and low K⁺. However, while the similarities in phenotypes between the two mutants are intriguing, several lines of evidence indicate that they are unrelated. First, the ex-autogamous F₂ from a mgx x rst mating yielded four classes of cells that included a double mutant, indicative of two independent alleles. Second, the restless Mg²⁺-lethal phenotype could be suppressed with 10 mM KCl whereas magbane was not protected by high K⁺. Finally, electrophysiological analysis of magbane failed to find any evidence of a K⁺-current anomaly that is believed to underlie a rst mutant behavioral phenotype. Thus, it is likely that the two mutations cause Mg²⁺ sensitivity through two independent pathways.

The possible causes of magbane's Mg²⁺-sensitive phenotype are considered in more detail below, but it is plausible that the Mg²⁺ intolerance of rst mutant cells is caused by the same K⁺ current anomaly that causes the behavioral phenotype. When wild-type cells are bathed in low K⁺ media such as the growth medium used here, they have a V_m of -40 mV. This provides an inward driving force for Mg²⁺ influx of sufficient magnitude to drive intracellular levels toward 700 mM when extracellular concentration is 30 mM Mg²⁺ (the concentration used to differentiate the mgx mutant phenotype from that of the wild type when scoring genetic crosses). [Mg²⁺]_i in Paramecium is held at 0.4 mM (PRESTON 1990 , PRESTON 1998 ). This is made possible by the activity of transporters whose identity is so far unknown, but the reduced rates of growth at 20 mM Mg²⁺ (Fig 1B) suggest that these transporters are stressed at such high levels of [Mg²⁺]_o. Restless had a V_m of -80 mV in low K⁺ media, a potential that would be predicted to drive [Mg²⁺]_i to >17 M at equilibrium. Reducing the threat to [Mg²⁺]_i to wild-type levels would require [Mg²⁺]_o to be dropped to 1.25 mM, so it is easy to imagine that a very negative V_m caused [Mg²⁺]_i homeostatic mechanisms in rst to be overwhelmed at a lower [Mg²⁺]_o compared with the wild type. Support for this idea comes from the effects of supplementing media with 10 mM KCl to raise V_m of rst mutant cells to wild-type levels ( -20 mV; RICHARD et al. 1985 ). This largely eliminated the difference in Mg²⁺ sensitivity between rst and the wild type (Fig 4) while having no effect on mgx mutant survival.

Mg²⁺ sensitivity of magbane:
The specificity of the magbane phenotype for Mg²⁺ over other ions (Na⁺ and K⁺) suggests that the mutation may affect one of the cellular components responsible for regulating [Mg²⁺]_i in Paramecium. Since we have little information about the molecular mechanisms of Mg²⁺ homeostasis in eukaryotes, any discussion of potential sites of lesion in magbane must necessarily be speculative.

Unlike most eukaryotic cells, P. tetraurelia has a relatively high membrane permeability to Mg²⁺. This is by virtue of its expressing a Mg²⁺-specific membrane current, I_Mg. We considered that magbane may have become overly sensitive to Mg²⁺ through inappropriate upregulation of this current, an idea that gained credibility when a well-characterized mutant with an enhanced I_Mg was found to be slightly more sensitive to Mg²⁺ than the wild type (Cha; Fig 2A). However, examination of magbane under voltage clamp found that I_Mg had normal amplitude and kinetics (not shown). It is possible that mgx mutation enhances a Mg²⁺ permeability that has not been characterized previously, but this should have been apparent as a "leak" current in Mg²⁺ solutions. If Mg²⁺-influx pathways are normal in mgx, then perhaps the mutant has a problem removing Mg²⁺ from the cytoplasm. CRONKITE and BURG 1982 noted that two strains of K⁺ sensitives were unable to control [K⁺]_i in the face of rising extracellular concentrations and similarly considered a permeability or pump defect as the underlying cause. Many eukaryotic cells expel Mg²⁺ in a Na⁺-dependent manner, suggesting that a Mg²⁺:Na⁺ exchanger may be involved (reviewed by BEYENBACH 1990 ). It is conceivable that mgx mutation reduces the efficacy of such an exchanger and thereby causes [Mg²⁺]_i to increase to lethal levels when extracellular concentrations are raised beyond a certain level. Thus, it may be interesting to determine the effects of mgx mutation on [Mg²⁺]_i using ion-sensitive fluorescent dye techniques (PRESTON 1998 ). Note that it recently became possible to identify mutant alleles of P. tetraurelia through complementation cloning, a technique that involves injecting mutant cells with fragments of wild-type DNA in an attempt to find a sequence that restores a normal (or near normal) phenotype (HAYNES et al. 1998 ). This technique has already yielded two genes that are required for Ca²⁺-channel function (HAYNES et al. 1998 , HAYNES et al. 2000 ), a novel tubulin gene, and several genes involved in secretion (KELLER and COHEN 2000 ). We hope to use similar methods to characterize the mgx gene and perhaps provide new insights into the molecular mechanisms of intracellular Mg²⁺ homeostasis in eukaryotes.

	ACKNOWLEDGMENTS

We are grateful to Serita Reels (Friends High School, Philadelphia) and Alea Harmon (Central High School, Philadelphia) for their help during the early mutant screening procedures, to Dr. Barbara Mroz for christening magbane, Dr. Don Cronkite for his encouragement and advice on isolating sensitive mutants, and to Dr. Ching Kung for providing many of the behavioral mutants used in these studies and for helpful discussions. We were supported by the National Institutes of Health (GM 51498).

Manuscript received December 15, 2000; Accepted for publication April 9, 2001.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ALTURA, B. M. and B. T. ALTURA, 1995  Magnesium in cardiovascular biology. Sci. Am. 37:28-37.

ANDRIVON, C., 1972  The stopping of ciliary movements by nickel salts in Paramecium caudatum: the antagonism of K⁺ and Ca⁺⁺ ions. Acta Protozool. 11:373-386.

ARSENIAN, M. A., 1992  Magnesium and cardiovascular disease. Prog. Cardiovasc. Dis. 19:271-310.

BALON, T. W., A. JASMAN, S. SCOTT, W. P. MEEHAN, and R. K. RUDE et al., 1994  Dietary magnesium prevents fructose-induced insulin insensitivity in rats. Hypertension 23:1036-1039[Abstract/Free Full Text].

BALON, T. W., J.-L. GU, Y. TOKUYAMA, A. P. JASMAN, and J. L. NADLER, 1995  Magnesium supplementation reduces development of diabetes in a rat model of spontaneous NIDDM. Am. J. Physiol. 269:E745-E752[Abstract/Free Full Text].

BEYENBACH, K. W., 1990  Transport of magnesium across biological membranes. Magnesium Trace Elem. 9:233-254[Medline].

CHANG, S.-Y. and C. KUNG, 1976  Selection and analysis of a mutant Paramecium tetraurelia lacking behavioral response to tetraethylammonium. Genet. Res. 27:97-107[Medline].

CRONKITE, D. L., 1983  Gene interactions between potassium-sensitive and potassium-resistant mutations of Paramecium tetraurelia. Genetics 103:153-160[Abstract/Free Full Text].

CRONKITE, D. L. and M. BURG, 1982  Ion regulation in potassium-sensitive mutants of Paramecium tetraurelia. J. Cell. Physiol. 110:271-276[Medline].

EVANS, T. C. and D. L. NELSON, 1989  New mutants of Paramecium tetraurelia defective in a calcium control mechanism: genetic and behavioral characterizations. Genetics 121:491-500[Abstract/Free Full Text].

FLATMAN, P. W., 1991  Mechanisms of magnesium transport. Annu. Rev. Physiol. 53:259-271[Medline].

GÜNTHER, T., and H. EBEL, 1990 Membrane transport of magnesium, pp. 215–225 in Metal Ions in Biological Systems, edited by H. SIGEL and A. SIGEL. Marcel Dekker, New York.

HAYNES, W. J., K. Y. LING, Y. SAIMI, and C. KUNG, 1996  Toward cloning genes by complementation in paramecium. J. Neurogenet. 11:81-98[Medline].

HAYNES, W. J., B. VAILLANT, R. R. PRESTON, Y. SAIMI, and C. KUNG, 1998  The cloning by complementation of the pawn-A gene in Paramecium. Genetics 149:947-957[Abstract/Free Full Text].

HAYNES, W. J., K. Y. LING, R. R. PRESTON, Y. SAIMI, and C. KUNG, 2000  The cloning and molecular analysis of pawn-B in Paramecium. Genetics 155:1105-1117[Abstract/Free Full Text].

HINRICHSEN, R. D., Y. SAIMI, and C. KUNG, 1984  Mutants with altered Ca²⁺-channel properties in Paramecium tetraurelia: isolation, characterization and genetic analysis. Genetics 108:545-558[Abstract/Free Full Text].

HMIEL, S. P., M. D. SNAVELY, J. B. FLORER, M. E. MAGUIRE, and C. G. MILLER, 1989  Magnesium transport in Salmonella typhimurium: genetic characterization and cloning of three magnesium transport loci. J. Bacteriol. 171:4742-4751[Abstract/Free Full Text].

ISHIJIMA, S., T. SONODA, and M. TATIBANA, 1991  Mitogen-induced early increase in cytosolic free Mg²⁺ concentration in single Swiss 3T3 fibroblasts. Am. J. Physiol. 261:C1074-C1080[Abstract/Free Full Text].

KELLER, A.-M. and J. COHEN, 2000  An indexed genomic library for Paramecium complementation cloning. J. Euk. Microbiol. 47:1-6[Medline].

KINK, J. A., M. MALEY, R. R. PRESTON, K.-Y. LING, and M. A. WALLEN-FRIEDMAN et al., 1990  Mutations in Paramecium calmodulin indicate functional differences between the C-terminal and N-terminal lobes in vivo. Cell 62:165-174[Medline].

KUNG, C., 1971  Genic mutants with altered system of excitation in Paramecium aurelia. II. Mutagenesis, screening and genetic analysis of the mutants. Genetics 69:29-45[Free Full Text].

KUNG, C., Y. SAIMI, W. J. HAYNES, K.-Y. LING, and R. KISSMEHL, 2000  Recent advances in the molecular genetics of Paramecium. J. Euk. Microbiol. 47:11-14[Medline].

LEFORT-TRAN, M., K. AUFDERHEIDE, M. POUPHILE, M. ROSSIGNOL, and J. BEISSON, 1981  Control of exocytotic processes: cytological and physiological studies of trichocyst mutants in Paramecium tetraurelia. J. Cell Biol. 88:301-311[Abstract/Free Full Text].

MAGUIRE, M. E., 1992  MgtA and MgtB: prokaryotic P-type ATPases that mediate Mg²⁺ influx. J. Bioenerget. Biomembr. 24:319-328[Medline].

MONCRIEF, M. B. and M. E. MAGUIRE, 1999  Magnesium transport in prokaryotes. J. Biol. Inorg. Chem. 4:523-527[Medline].

MURPHY, E., C. C. FREUDENRICH, and M. LIEBERMAN, 1991  Cellular magnesium and Na/Mg exchange in heart cells. Annu. Rev. Physiol. 53:273-287[Medline].

NELSON, D. L. and E. P. KENNEDY, 1972  Transport of magnesium by a repressible and a nonrepressible system in Escherichia coli. Proc. Natl. Acad. Sci. USA 69:1091-1093[Abstract/Free Full Text].

PAOLISSO, G., A. SCHEEN, F. D'ONOFRIO, and P. LEFE'BVRE, 1990  Magnesium and glucose homeostasis. Diabetologia 33:511-514[Medline].

PARK, M. H., B. B. WONG, and J. E. LUSK, 1976  Mutants in three genes affecting transport of magnesium in Escherichia coli: genetics and physiology. J. Bacteriol. 126:1096-1103[Abstract/Free Full Text].

PRESTON, R. R., 1990  A magnesium current in Paramecium. Science 250:285-288[Abstract/Free Full Text].

PRESTON, R. R., 1998  Transmembrane Mg²⁺ currents and intracellular free Mg²⁺ concentration in Paramecium tetraurelia. J. Membr. Biol. 164:11-24[Medline].

PRESTON, R. R. and J. A. HAMMOND, 1997  Phenotypic and genetic analysis of "chameleon," a Paramecium mutant with an enhanced sensitivity to magnesium. Genetics 146:871-880[Abstract].

PRESTON, R. R. and J. A. HAMMOND, 1998  Long-term adaptation of Ca²⁺-dependent behaviour in Paramecium tetraurelia. J. Exp. Biol. 201:1835-1846[Abstract].

PRESTON, R. R. and C. KUNG, 1994a  Inhibition of Mg²⁺ current by single-gene mutation in Paramecium. J. Membr. Biol. 139:203-212[Medline].

PRESTON, R. R. and C. KUNG, 1994b  Isolation and characterization of Paramecium mutants defective in their response to magnesium. Genetics 137:759-769[Abstract].

PRESTON, R. R., Y. SAIMI, and C. KUNG, 1990a  Evidence for two K⁺ currents activated upon hyperpolarization of Paramecium tetraurelia. J. Membr. Biol. 115:41-50[Medline].

PRESTON, R. R., Y. SAIMI, E. AMBERGER, and C. KUNG, 1990b  Interactions between mutants with defects in two Ca²⁺-dependent K⁺ currents of Paramecium tetraurelia. J. Membr. Biol. 115:61-69[Medline].

PRESTON, R. R., Y. SAIMI, and C. KUNG, 1992  Calcium current activated upon hyperpolarization of Paramecium tetraurelia. J. Gen. Physiol. 100:233-251[Abstract/Free Full Text].

RICHARD, E. A., R. D. HINRICHSEN, and C. KUNG, 1985  A single gene mutation that affects a potassium conductance and resting membrane potential in Paramecium. J. Neurogenet. 2:239-252[Medline].

RICHARD, E. A., Y. SAIMI, and C. KUNG, 1986  A mutation that increases a novel calcium-activated potassium conductance of Paramecium tetraurelia. J. Membr. Biol. 91:173-181[Medline].

ROMANI, A. and A. SCARPA, 1992  Regulation of cell magnesium. Arch. Biochem. Biophys. 298:1-12[Medline].

SAIMI, Y. and C. KUNG, 1987  Behavioral genetics of Paramecium. Annu. Rev. Genet. 21:47-65[Medline].

SCHEIN, S. J., 1976  Nonbehavioral selection for pawns, mutants of Paramecium aurelia with decreased excitability. Genetics 84:455-468.

SHUSTERMAN, C. L., E. W. THIEDE, and C. KUNG, 1978  K⁺-resistant mutants and "adaptation" in Paramecium. Proc. Natl. Acad. Sci. USA 75:5645-5649[Abstract/Free Full Text].

SMITH, R. L. and M. E. MAGUIRE, 1995  Distribution of the CorA Mg²⁺ transport system in gram-negative bacteria. J. Bacteriol. 177:1638-1640[Abstract/Free Full Text].

SNAVELY, M. D., S. A. GRAVINA, T. T. CHEUNG, C. G. MILLER, and M. E. MAGUIRE, 1991a  Magnesium transport in Salmonella typhimurium. Regulation of mgtA and mgtB expression. J. Biol. Chem. 266:824-829[Abstract/Free Full Text].

SNAVELY, M. D., C. G. MILLER, and M. E. MAGUIRE, 1991b  The mgtB Mg²⁺ transport locus of Salmonella typhimurium encodes a P-type ATPase. J. Biol. Chem. 266:815-823[Abstract/Free Full Text].

SONNEBORN, T. M., 1950  Methods in the general biology and genetics of Paramecium aurelia. J. Exp. Zool. 113:87-148.

SONNEBORN, T. M., 1970  Methods in Paramecium research. Methods Cell Physiol. 4:241-339.

TOSIELLO, L., 1996  Hypomagnesemia and diabetes mellitus. Arch. Int. Med. 156:1143-1148[Abstract].

TOUYZ, R. M. and E. L. SCHIFFRIN, 1996  Angiotensin II and vasopressin modulate intracellular free magnesium in vascular smooth muscle cells through Na⁺-dependent protein kinase C pathways. J. Biol. Chem. 271:24353-24358[Abstract/Free Full Text].

VAN HOUTEN, J., S.-Y. CHANG, and C. KUNG, 1977  Genetic analysis of "Paranoiac" mutants of Paramecium tetraurelia. Genetics 86:113-120[Abstract/Free Full Text].

WHITE, J. R. J. and R. K. CAMPBELL, 1993  Magnesium and diabetes: a review. Ann. Pharmacother. 27:775-780[Abstract].