(:op\right 8 1988 by the Genetics Society of America

Phenotypic Characterization of Paranoiac and Related Mutants in Paramecium tetraurelia

Manuscript received August 17, 1987

ABSTRACT Paranoiac and related mutants of Paramecium tetraurelia display altered membrane excitability. of We describe an extension of behavioral characterizations the paranoiac, fast-2, and tetraethylamin detail theirreactions to sodiumstimulation under monium-insensitivemutants,comparing standard culture conditions, when grown at various temperatures and when starved. We also use freeze-fracture electron microscopic techniques to analyze these in stocks the morphology of organizedarrays of membrane particles,theciliaryplaques.This group of mutants is diverse, showingdifferences in behavior under standard culture conditions and differentreactions to temperature and starvation stresses. Ciliary plaque morphology is altered in some, but not all, of the mutants. The possibility is discussed that these plaques may be sites of potassium or sodium transport. NDERSTANDINGnerveimpulseconduction and control of muscle contraction requires understanding the structure and function of excitable membranes. We must learn what ion transport gates, channelsandpumps exist and how they function togethertodetermine cellular electrophysiological activity. While significant progress has beenmade using membrane active drugs to interfere with normal excitation in higher organisms, perhaps the most useful model system for a genetic dissection of the activity of excitable membranes has been the ciliated protozoan,Paramecium.Pioneering work by Jennings (1906) on the swimming behavior and by KAMADA (1929) and KAMADA and KINOSITA(1940) on the electrophysiology of paramecia led to KUNGS (197 la, b)firstreports of behavioral mutants in Paramecium tetraurelia. This and subsequent work has been reviewed in detail by ALLEN (1978), BYRNE and KUNG BYRNE (1978a), CRONKITE (1979), (1979), ECKERT and BREHM (1979), and KUNGand SAIMI (1982, 1985). Normal swimming in a left forward spiral (LFS) is the result of coordinated beating of the cilia in metachronal waves with the posteriorly directed effective stroke propelling the cell forward. Mechanical or chemical stimulation of wild-type cells can induce an avoiding reaction with reversal of the direction of the effective stroke,propelling the cell backward. Ciliary reversal typically lasts 0.8 to 1 sec and may be repeated at brief intervals in a pattern of periodic ciliary reversal (PCR); the cell alternately darts backward and forward (JENNINGS 1906). Paranoiacs are characterized by extended periods
Genetics 118: 619-626 (April, 1988).

of backward swimming as the result of continuous ciliary reversal (CCR), especially inreaction to sodium stimulation (KUNG1971b). Mutations at six different loci (PaA, paB, PaC, PaD, PaE, and f n a ) have produced the paranoiac phenotype (VANHOUTEN, and CHANG and KUNG 1977;BYRNE,TANNER MuILENBURG 1981). Several of these mutants have abnormal potassium and/or sodium ion permeabilities as do the fast-2 (fna) and tetraethylammonium KUNG (TEA)-insensitive ( t e d ) mutants (HANSMA and 1976; SATOW KUNG1976a; SATOW, and HANSMA and KUNG 1976; SAIMI 1986). These mutants, potentially important in elucidating the complex interactions of ions, receptors and membrane currents that dictate swimming patterns,are difficult to study because their phenotype is an exaggeration of wild-type behavior, rather than a novel reaction to a stimulus or the complete absence of anormal behavior. Our detailed analysis of the swimming behavior and membrane plaque morphology of the paranoiacs and the usefulness related fna and teaA mutants extends their in genetic, physiological and biochemical studies. MATERIALS AND METHODS
All stocks reported here were derived from the P. tetraurelia wild-type stock 51, nonkiller strain. The mutant stocks d90 (PaA), d578 ( P a A l ) , dl47 (PUB), dl50 (PaC), d564 (PaD), dl49 (fnaP), d91 (fm) dl52 ( t e a A ) are and from C. KUNGScollection at the University of Wisconsin. PaA and PaAl are allelic as are fm and fnaP. We isolated stock d250 (PaE) KUNGS in laboratory. These mutantswere collected following treatment of wild-type cells with Nmethyl-N-nitro-nitrosoguanidine. They are essentially isogenic with stock 51 and with each other. All the mutants

620

B. J. Byrne, A. P. Tanner and P. M. Dietz

Duration of backward swimming in response to sodium stimulation

Genotypes
+I+
teaAlteaA

paBlpaB

PaDIPaD

PaAIPaA

PaAliPaAI

PaCiPaC

PaEIPaE

fnaPifnaP

20 20 5 5 10" 5d 5d

5.0 20 20 20 8 8 8 8

4.7 3.6 6.2 2.1 2.0 1.4

0 4.4 0 ? 2.8 0 f 5.6 0 t 1.9 7.5 +. 2.1 f 4.5 1.0 f 2.2 f 3.1 0
f 4.4
f

0.6 f 1.2 3.8 f 6.3 14.8 f 6.4 6.5 f 5.7 0 0 0 0 1.3 f 1.4 8.8 f 3.3 4.0 f 4.2 10.0 f 9.3 1.8 f 2.7 5.6 ? 3.2

7.4 t 5.8 59.8 f 36.2

16.7 f 4.6 74.0 f 61.4 37.6 f 10.4

29.9 f 25.3 28.8 f 36.9 f 43.8 72.4 f 18.8 ? NT 33.7 f NT 26.2 f 73.0 f 53.8 8.0 f 20.0 f 3.7 2.0 f

18.9 34.9 10.4 14.7 5.5 5.7 4.5

33.9 k 24.7 36.2 k 48.7 21.9 k 24.4 22.9 2 15.6 5.3 ? 4.4 10.0 2 9.7 9.5 5 3.9

,yrb
67.5 39.3 75.3 14.6 20.0 71.0
f 30.9

37.4 38.2 ? 14.5 t 5.0 f 44.2

The duration measured is the number of seconds f one standard deviation for the initial period of backward swimming upon transfer of individual cells to the sodium test solutions described in MATERIALS AND METHODS. Thefnalfnu cells showed no ciliary reversal in any of the observations. a n = number of individual cells tested. NT = not tested. In this experiment n = 10 except for + (n = 24), PuE (n = 35), andfnuP ( n = 25). In this experiment n = 5 except for PuE a n d f d (n = 10).

exceptfna and teaA exhibit a paranoiac phenotype. Mutant fm was included because the mutation is allelic to the paranoiac,fnaP, and because fna shows potassium permeability abnormalities as do some paranoiacs (HANSMA and KUNC 1976;SATOWand KUNC 1976a).Mutant t e a A was shows potassium permeability included because it also abnormalities (SATOW KUNC 1976b). and Standard culture techniques (SONNEBORN were 1970) used. Cells were cultured in Cerophyl rye grass infusion, a complex, undefined medium, previously inoculated with Klebsiella pneumoniue. Swimming behavior was observed in culture fluid and upon transfer to one of two test solutions: (1) the sodium CHANC and KUNC(1977) (20 test solution of VANHOUTEN, m NaCI, 0.3 m CaC12,ImM Tris; pH 7.2 adjusted with M M M HCl); (2) the same as solution 1 except containing 8 m NaCl. Unless specified otherwise, the tests were done with log phase cells grown at 25-27" using test solutions at the same temperature. Responsetosodiumstimulation was measured as mean duration of the first episode of backward swimming after introduction into the test solution. For electron microscopy cells were fixed with glutaraland dehyde and frozen in glycerin as detailed by BYRNE BYRNE (197813). Preparations were fractured in a Balzers BA 360M or 500 freeze-fracture apparatus and examined with a Philips EM300. Electron micrographs were routinely examined as negatives under a stereomicroscope at X 10 magnification.Particle arrays on A fracture faceswere recorded and the results analyzedfor statistically significant differences between mutants and wild type. Examination of the B fracture faces of the plaque regions revealedvery few particles adhering inany of the stocks under any cultural conditions.

the cultures are disturbed by gentle shaking or pipetting. Under our growth conditions, PaD, fnaP paB, a n d PaE are sometimes, but not always, distinguishable from wild type on this basis with fnaP and PaE more likely than PaB and PaD to show spontaneous CCR. T h e f n a and teaA mutants, not paranoiacs, are indistinguishable from wild type the on basis of frequency of CCR in culture fluid although the fast2 mutant generally be can identified by its fast forward swimming with spurts of even faster swimming. Wild-type cells grownat 27", testedinsodium solutions, initially showed CCR that lasted from less than one second to as long as 15 sec. Variation in mean length of reversal and standard deviation is apparent in different experiments (Table 1). There is n o consistent correlation between this variation and the test solution used. Cells grown at 17" showed a longer and more variable duration of initial CCR in response to sodium than did grown at27" (Table cells 2). On the other hand, cells grown at 27" until they hadreachedstationaryphaseandbeguntostarve showed a shorter, less variabledurationof initial CCR (Table 3) than the control, well-fed cells. Repeated observations of duration of initial CCR in response to sodium for the mutants grown at 27" and tested while well fed are recorded in Table 1. T h e various stocks showed different patterns which canreasonablybedividedintothree classes. Fna showed no discernible avoiding reaction; wild type, RESULTS teaA, PUB and PuD typically showed initial CCR with Swimming behavior a mean duration under 10 sec, and the other paranoiacs typically showed initial CCR with a mean Ranges of behaviors in culture fluid and test soduration longer than 10 sec. lutions: We have repeated the observations of KUNG Growth at 17" had a marked effect on the length (1971b) and VANHOUTEN, CHANG and KUNG (1977) of CCR in most of the stocks tested (Table 2). Like thatinnormalculturefluid PaA,PaAl and PaC wild type, PUG, PaE and fnaP showed longer CCR regularly show a high frequency of CCR (continuous when grown at the lower temperature with the greatciliary reversal). This is especially pronounced when

Paranoiac Mutants in Paramecium

62 I

Genotype

27"

(n)

17"

(n)

+/+ PaCIPaC paEIPaE fnaplfnap paBlpaB PaDIPaD PaAIPaA teaAlteaA fnatfm

2.1 f(24) 1.9 26.2 f 5.5 (10) 5.3 f 4.4 (35) 14.6 f 14.5 (25) 1.3 2 1.4 (10) 8.8 2 3.3 (10) 16.7 f 4.6 (10) 7.5 f 2.1 (10) 0 (10)

10.2 f 12.4 (21) 36.1 f 25.8 (10) 16.5 f 11.4 (30) 0.01 55.2 f 21.4 (22) 1.0 2 1.8 (10) 6.0 f 4.5 (10) 10.7 f 5.0 (10) 0 (10) 0 (10)

0.01 0.1 0.01

0.05

The duration is measured as in Table 1. Sodium test solution 2 (8 m NaCI)was used. Equality of themeans was tested by M comparing at-valuecalculatedassumingunequalvariance, to a critical value for t based on an average weighted for sample size (SOKAL ROHLF,1969). and 'Not significant.
TABLE 3 Effect of starvation on duration of backward swimming
Level of significance of difference of

Genotype

Well fed

means

Starved

+/+ PaAl/PaAl PaCIPaC PaEIPaE PaBIPaB PaD)IPaD PaAIPaA

5.0 2 4.4 29.9 2 25.3 28.8 f 18.9 33.9 f 24.7 4.1 0.6 f 1.2 1.0 3.8 f 6.3 7.4 f 5.8

2.7 f 1.6 19.1 f 10.3 19.6 f 11.3 2 3.0 f 1.2 3.3 2 1.7 17.0 f 15.3

0.05

0.1 0.1 0.01

NS

0.05

The duration is measured as in Table 1. In each case n = 20, and the test solution was sodium solution 1 (20 m NaC1). Equality M of the means was tested as in Table 2. Not significant.

est increase for fnaP. PaA showed a small decrease in duration of ciliary reversal, and teuA declined dramatically from a mean length of nearly 8 sec at 27" to no CCR at all at 17". Differences for PUB and PaD are not significant, and fm showed no reversal at any temperature. When stationary phase, starving cells werecomPUB and PuD showed virtually pared to well-fed cells, no change in length of CCR but, like wild type, PaAI, PuC and PaE showed a significant drop (Table 3). The decrease for PaE was especially great and has been observed repeatedly. PuA was quite unlike the other stocks, showing a large increase in length of CCR with starvation. Stocks f n a and teuA were not tested when starved. As with wild type, variation in the length of initial CCR occurred even among different cells from the same sample. Behavior of F1 cells is notalwayspredictable: PaAl and PaD are dominantmutations. The behavior

of F l cells from crosses of thesemutants by wild type is indistinguishable from the behavior of the mutant and parents (VANHOUTEN,CHANC KUNG 1977). In the same paper paB and fnaP are described as receswild type is sive mutations. The F1 from PaC by described as partial paranoiac with prologned backward swimming, but less frequently and for shorter durationsthan fullparanoiacs. These Fls did not react asstronglytosodium stimulation as the full paranoiacs. KUNG (197lb) described the same partial paranoiac phenotype for the F1 of PaA by wild type. We found a similar pattern of behavior for the F1 cells of PaE by wild-type crosses.The periods of CCR in response to sodium were typically longer than the wild type but shorter than, though sometimes approaching, the PaE parent. Thesestocks are classified as codominant or semidominant. The f n a and teaA mutations are recessive (VANHOUTEN,CHANC and KUNG1977; SATOW KUNC 1976a, b). and Crosses of two different paranoiacs by f n u produce Fls with unexpected phenotypes. VANHOUTEN, CHANC KUNG(1977) report wild-type phenotype and for the fna/fnaP heterozygote. In our crosses of the same stocks we saw behavior that is not fast nor paranoiac, nor wild type. In sodium test solutionsthe cells initially showeda fast-2 phenotype as they swam forward rapidly with no avoiding reaction. Within a few seconds they began to show periodic bursts of CCR interspersed with fast forward swimming. In our crosses of PaE to f n a , the F1 cells' response to sodium stimulation was not distinguishable from that of wild type, and the Fls derived from crosses of PaE by fnaP were all full paranoiacs.In both cases the phenotypes are unexpected since PaE is semidominant, and f n a and fnaP are both recessive mutations. The expected FI phenotype is partial para(1971b)has describedthe F1 from crosses noiac. KUNC of f n a to PaA as "slightly paranoiac," the same description given for the FI from crosses ofPaA to wild type, and consistent with the designation of PaA as a semidominant mutation. The F1 phenotype for crossesof other paranoiacs by f n a has not been described, so we do not know the full range of interactions between the f n a and paranoiac loci.

Ciliary plaque morphology Wild-type cells display limited range of plaque morphologies: No stockof Paramecium, including wild type, displays an invariable pattern of membrane particles within ciliary membrane plaques. A "stand a r d plaque is an array of particlesconsisting of three columns and about five rows. Nine plaques are arranged around the base of each cilium, overlying the microtubular doublets of the axoneme (DUTEand KUNC1978; PLATTNER, MILLER BACHMAN and 1973). Differences in plaque phenotype between wild-type and mutant stocks are differences in mean numbers

622
TABLE 4

B. J. Byrne, A. P. Tanner and P. M. Dietz

Characterization of plaques and behaviorof wild-type stock 51 cells

particles

Behavior in sodium test solution

Genotype

no

No. of rows*

particles missing 27.3 28.5 21.1 22.1 18.3 7.8 8.6 10.9 0.6

0-row

plaques 63.6 12.5 9.6 1.6 11.5 1.1 4.5 1.9 0

14 13 14 19 16 16 17 14 19 21 16

4.3 4.6 4.6 4.7

f 0.7
f

1.2

f 0.5 f 1.0

11.9 1.7 4.1 24.7 50% 2.6 9.2 17.7 7.4 2.8 3.2 4.5

4.7 f 0.9 4.7 f 1.1 4.8 f 0.6 4.9 f 0.9 5.1 k 0.9 5.4 f 0.7 5.5 f 0.9

No long CCRc 70% PCR; 30% CCR 100% PCR PCR; 30% LFS; 20% CCR tested Not 70% PCR; 30% CCR 50% short CCR; 50% PCR tested Not tested Not 100% PCR Not tested'

PaAIIPaAI fnayna PaElPaE

fnaPlfnaP +I+ PaDIPaD PaClPaC paBlpaB

66 97 73 63 80 179 67 52 40

4.16 4.55 4.66 4.69 4.73 4.87 4.87 4.88 5.35

f 1.0
f 1.4 f 1.1 f 1.1

f 1.1

0.9 0.8 It 0.9 f 0.9

Number of plaques, including 0-row plaques,examined. Mean f one standard deviation;excludes 0-row plaques.Note that including 0-row plaques in this value would generally accentuate the differences between the stocks in the top half the table of and those in the bottom half.

one standard deviation. In these experiments the final cell density was under 1000Im1, an abnormally low value.
f

' The mean is the mean numberof horizontal rows of particles,

This column records the number of ciliary plaques counted.

of particles in adimension of theplaquematrix, differences in mean number of rows per plaque (plaque length), and differences in the constancy of length as reflected by the standard deviation. We examined both the length of plaques and the frequency of missing plaque particles, noting variability between wild-type cells prepared at different times (Table 4). In sum,thenumber of rows of particles per plaque ranged from 2 to 8 with a mean of 4.89 and a nearly normal distribution around 5. In two cases an area in which a plaque was expected showed no organized rows of particles. These areas, classified as 0-row plaques, accounted for 5% of the total in one sample, 7% in another, and were not represented in the remaining samples, giving an average frequency of 1.1% (2/175) 0-row or missing plaques. In different samples the mean number of rows per plaque, excluding the 0-row plaques, ranged from 4.31 to 5.50 with no consistent correlation between mean plaque length and either cell density or behavioral reactiontosodiumpriortofixation and freezing. The frequency of points within a plaque where a particle was expected but not observed (missing particles) also showed variation between samples, ranging from 1.7 to 24.7%. Our data do not demonstrate a correlation between frequency of missing particles in wild type cells andrecordeddifferencesin cell density or behavioral patterns. Inherent difficulties in maintaining at similar stages in the cell cycle the very large numbers of cells required for freeze fracture analyses make it impractical to seek more precise correlations.

Mutantsshow four discretepatterns of plaque morphology: In Table 5 the stocks studied are listed in order of increasing mean length plaque. Neither of PaC nor PaD was distinguishable from wild type in mean plaque length or variability of plaque length. Both mutants, like wild type, showed nearly normal distribution around the mean with each stock having somewhat more plaques on the low side of the mode of 5 than on the high side. Neither mutant showed a high frequency 0-row plaques or missing particles of within the plaques. Two by two contingency tests (wild type with PaC and PaD) for identity of frequency of missing particles generated values that are not significant at the 0.05 level except in one preparation of PaC where the frequency is significantly higher than instock 5 1 (Table 6). We cannot attribute this one deviation to known factors. from all other stocks Paranoiac PaAl differed examined. Consistent with our earlier report (BYRNE and BYRNE1978b), the frequency of 0-row plaques was very high, and those plaques present were much shorter and had many more missing particles than wild-type plaques. The distribution of number of rows per plaque was non-normal, even ignoring the 0-row plaques. The difference in frequency of missing particles compared withwild type is significant (Table 6). A morphologically intermediate group of stocks includes fna, fnaP,PaE, and teaA. PaA, characterized previously (BYRNE and BYRNE 1978b), also falls into this class. Plaques in these stocks had fewer rows than those of wild type. All these stocks except teaA had a much higher frequency of 0-row plaques than wild type, and all had many more missing particles (Tables 5 and 6), althoughthe particles that were present appeared normal in size. There were important deviations from normal distribution of plaque size

Paramecium

in
TABLE 6

Mutants Paranoiac

623

C o m e s o n of frequency of missing particles for wild type and mutants

Stock with higher frequency

x* from 2 X 2 contingency t e s t
57.61 32.01 20.12 90.22 52.54 27.34 13.92 120.88 40.70 0 35.80 7.30 0.23 46.55 28.79 1.64 8.05 3.35 1.30 0 14.45 1.85 1.70 4.86 3.70 2.73

Level of significance 1 df

PaAI

Mutant Mutant Mutant 0.01 Mutant Mutant Mutant Mutant 0.01 Mutant Mutant Neither 0.01 Mutant Mutant Wild type Mutant Mutant Mutant Wild type Wild type Mutant Neither Mutant Mutant Mutant Wild type Wild type Wild type

0.01 0.01 0.01 0.01 0.01 0.01 0.01

f m

wild type with a cumulative mean of 5.35 rows as opposed to 4.87 rows for wild type (Table 5 ) . There was a distribution of plaque size around the mode of 5 , but therewas a noticeable excess oflonger plaques. This differs from all other stocks where plaques with fewer than five rows outnumbered those with more than five. There were no missing plaques in any of the samples, and only 0.62% of the particles were missing. As seen in Table 6, thoughthe level of significance is marginal, the frequency of missing particles was consistently lower in PUB than in wild type in marked contrast to the other mutants.
DISCUSSION

PaE

ted

0.01
NS

f d
PaD

0.01 0.01
NS

0.01
NS
NS

NS

PaC

0.01
NS

NS

PUB

0.05
NS NS

Only cilia with one or more plaques present were used for these calculations. With the PaA mutant this minimizes the differences withwild type because some cilia have no plaques at all. Each sample includes at least 4 cilia and a minimum of 75 loci where particles are expected. Not significant.

among these mutant stocks. Especially obvious is the high proportion of samples in the zero class. T e d , fna and PaE resembled wild type in other respects with a mode of five and more samples on the low side of five than the high. The bias toward fewer than five rows per plaque was more accentuated in these stocks than in wild type. StockfnaP had a mode of four rows per plaque with a bias toward higher values than the mode so that its mean was close to that of the other mutants. Two by two contingency tests (Table 6) showed that in most experiments the frequency of missing particles was significantly higher for each of these mutants than for wild type. In the preparations which did not show significant differences, thefrequency of missing particles for wild type was unusually high. The inhomogeneity within thecontrol populationsprecludescombining data from different runs. PUB displayed unique plaque morphology. every In experiment the mean plaque length was greater than

Swimming behavior: When studyingany organism ascomplexasParamecium, it should come as no surprise that normal swimming behavior and responses to external stimuli vary considerably. Such differences were reported by JENNINGS (1906) and have since been documented repeatedly. The variation has both genetic and nongenetic components. Large differences in behavior typify different behavioralmutants,documentingthe effects of several genes. In addition, individual cells of a single, genetically homogeneous line show differences in behavior even when dl cultural conditions are as uniform as possible because a single Parameciums behavior is its net response to complex membrane activities. Differences in the pattern of swimming can reliably beproduced by changingsuch factors as degree of starvation, temperature, culture medium and ions in the external milieu. Paramecia are grown in undefined culture medium under nonsterile conditions. Even small variation in the composition of growth medium could affect membrane characteristics and thus alter behavior. The paramecia feed on the bacteria in the medium which may support the growth of contaminating microorganisms as well as Klebsiella pneumoniae. T h e the inoculated food source, paramecia may themselves be of different ages, mating types, or serotypes, with subtle effects on behavior. Therefore, it is essential to record and appreciate therange of normal behavior and responsesto stimuli under specific culture conditions and define ranges for mutants ifwe are to realize the goal of correlatingchangesatdifferentphenotypic levels and relating them to models of excitable membrane functioning. We have focused on characterizing the swimming response to high external concentrations of sodium. I n wild-type cells sodium causes a brief episode of backward swimming. T h e impression from reviewing published reports is of a more uniform response than we have seen. DOUGHTY DRYL(1981) have reand viewed many earlier reports and descriptions of

624

B. J. Byrne, A. P. Tanner and P. M. Dietz

behaviors of Paramecium in response to a variety of stimuli and pointed out the dearth of precise data on frequency andduration of altered swimming patterns. Specific information is found in: (1) KUNCS(1976) report that, in response HANSMA and to sodium test solutions with sodium concentration ranging from 0 to 20 mM, wild-type backward swimming episodes usually lasted less than 1 sec and seldom more than 3 sec with the duration increasing with increasing sodium concentration; (2) HANSMAS (1979) reportthatthe mean durationofthe first episodeof backward swimming by wild-type cells M upon transfer to 10 m Na solutions ranged from 1 to 3 sec with Ca2+ concentrations of 0- 10 mM; (3) SAIMI and KUNGS (1980) report that transferof wildM type cells to 1 m Ca2+, 8 m Na solution induced M variable responses, often including backward swimming for up to 30 sec, followed by jerky swimming in more than one direction. Other reports classify wild-type responses less specifically as short or brief backward swimming. In thecourse of our observations of wild-type cells, using the test solutions described above, we saw initial CCR ranging from zero to as long as 16 sec if the cells tobe tested were well-fed and grownat 27. Different sets of observations produced mean duration of initial CCR between less than l sec and greater than 6 sec. Some cells did not swim backward at all in the 30 sec period of observation. When wild-type cells were starved before testing, the mean length of initial CCR was well under the mean for the group of well-fed cells observed at the same time. Growth at 17 consistently and markedly increased the length of initial CCR. These two effects are probably not unrelated. At 17 cells grow more slowly than at 27, and therefore were at an earlier point on their growth curve when tested. However, all cells were still actively dividing so that starvation cannot be invoked as the explanation for the difference.A most important lesson is the necessityof maintaining repeatable culture conditions as far as possible, of repeating observations on several groups of cells, and of using wild-type cells as controls in each group of observations. With thesecaveats in mind,we turn to the mutants. VANHOUTEN,CHANGand KUNC (1977) described mutants PaAI, PaC, PaD and fnaP as indistinguishable from each other in behavior, and presumably indistinguishablefrom PaA. PaB was described as having shorter, less frequent periods of spontaneous backward swimming in culture fluid and as not responding to sodium stimulation backward swimby ming. HANSMA and KUNC (1976) reported some differences in length of initial backward swimming in high sodium concentrations PaA, PaC and fnaP. by They also demonstrated differences between stocks in intracellular potassium and sodium concentrations

with varying external sodium concentration. HANSMA (1979) correlatedthe longerbackward swimming and higher internal sodium concentration PaC than of of fnuP cellswith a higher rate of sodium uptake by PaC than by fnaP. The teaA mutants, with abnormal potassium permeability, show few, weak avoiding reactions to sodium test solutions (KUNGet al. 1975; CHANC and KUNG1976). KUNG(1971b) and KUNGet al. (1975) characterized f n a cells as showing no immediate avoiding reaction in sodium. HANSMA (1979) found that these cells did show avoiding reaction to sodium in the first 30 sec after introduction to the test solution, but the numberof reversals was significantly lower than wild type and correlated with a lower rate of sodium uptake. We divide the paranoiacs and related mutants into three classes based on the length of initial backward swimming in sodium solutions. The f n a mutant, in a class by itself, showed no backward swimming under our test conditions. Wild-type cells and mutants PaB, PaD, and teuA, usually showed a short initial CCR, defined as less than 10 sec. Mutants that always or usually showed a long initial CCR, over 10 sec, were PaA, PaAI, PaC, PaE, and fnaP. Unidentifiable culture conditionshadamarked effect onlength of ciliary reversal. For instance, in one set of 20 cells of stock PaB, the mean length of reversal was 0.65 sec while the next day for a similar group of 20 cells it was 14.85 sec. Over the same 2 days, PaA and PaC cells showed similar variability but wild-type, PaAl, PaD and PaE cells didnot. Important effects of starvation and growth at low temperatures were superimposed on this unexplained variation. Like wild type, PaAI, PaC and PaE showed a shorter period of backward swimming when starved than when well fed, but the sample of starved PaA cells showed a longer period of backward swimming than the well-fed cells. There was no obvious effect of starvation on behavior of PUB or PuD cells. Growth at low temperature consistently increased the length of ciliary reversal for stocks PaC, PaE and fnaP as well as wild type. The difference in fnaP cellswas especially striking. Conversely, growth at 17 decreased the mean for PaA cells, and reduced backward swimming tozero for teaA cells. Itdidnot significantly affect the mean for PUB and PaD, and had no effectonfna cells which did not show reversal at either temperature. The variable phenotypes of these stocks and their overlap with wild-type phenotype make them challenging to work with. However, they are a valuable reservoir of material with significant membrane alterations. Further investigation of the biochemical and physiological bases of thesedifferences could add important dimensions to our understanding of membrane function. Ciliary plaques: Ciliary plaque morphology is rela-

Paranoiac Mutants in Paramecium

625

tively uniform in wild-type paramecia. The number of columns per plaque wasalways three, and most particle is points in a plaquewhereamembrane expected were, indeed, occupiedby a particle. Variation in the percentage of missing particles and the mean number of rows per plaque becomes apparent when cells from different experiments are compared. Attempts to correlate unusually high frequencies of missing particles with unusualgrowth or behavior were inconclusive. Plaque morphology is diverse within the paranoiac mutant group. divide the mutants We into four classes on the basisofsize and regularity of their ciliary plaques. In the first class, PaC and PaD were indistinguishable from wild type. The second and largest class includes PaA, fnaP, and PaE, and also fna and teaA. All had plaques with fewer rows, a higher frequency of plaques missing altogether (exceptteaA) and a higher frequency of missing particles than wild type.Thesedifferencesare statistically significant and were usually noticeable upon cursory examination of electron micrographs of a number of cells. T h e third class, to date occupied only by PaAl, has drastically reduced,abnormalplaques(BYRNEand BYRNE 1978b). The last class includes only stock paB, and is distinguished by having longer plaques with fewer missing particles than wild type. Notethat there is no regular correlation between these classes and those assigned on the basis of swimming behavior. What are likely functions of the plaque particles? They are probably not the site of calcium channels; there are no plaque alterations in pawn cells (B. J. BYRNE,unpublisheddata)and ciliary regeneration experiments indicate that calcium channels are situated along the entire length of the cilium, not concentrated at the ciliary base where the plaques are found (DUNLAP 1977). On the other hand, plaques are changedin fna, teaA, and most of the paranoiacs. Both fna and fnuP show potassium permeability abKUNG 1976; SATOW and normalities (HANSMA and KUNC 1976a), a situation consistent with a class of potassium channels associated with the ciliary plaques. Also consistent with this suggestion is the correlation of increased K' conductance in strain teaA (SATOW KUNC 1976b) and abnormal plaques and in the same strain. Plaques may also be sites of some class of sodium channels. Stock PaA, with abnormal K' and Na' permeabilities (HANSMA and KUNG 1976; SATOW, HANSMA and KUNG 1976), has abnormal ciliary plaques. SAIMI and KUNG(1980) demonstrated a greatly accentuated slow, calcium-induced inward Na' current in this stock. An alteration of Na+ channels is one possible cause of such a change in membrane currents. If this were the case in this mutant line, we could infer that Na' channels were also located at the ciliary plaques. This must remain

highly speculative unless further biochemical and electrophysiological studies clarify the nature of the lesions in this stock. ADOUTTE al. (1980) report two different memet brane protein alterations, detected by isoelectric focusing on acrylamide and SDS-PAGE associated with mutations PaA and fm. Mutants fnaP and PUB showed neither variation. A complete scrutiny of the membrane proteins of paranoiacs and related mutants using isoelectric focusing might further define the phenotypes of the mutants but would probably not contribute directly to our understandingof their biochemical bases unless alteredproteinscouldbe further characterized and localized in vivo. Both PaAl and PaA, allelic mutants,arestrong paranoiacs, but only PaAl shows grossly altered plaques. Electrophysiological studies of P a A l , comKUNG, parable to those done on PaA (HANSMA and 1976; SATOW, HANSMA and KUNG 1976;SAIMI and KUNG1980) could produce important new informanow, though tion. PaAl cells are unabletomate KUNG (1977) successfully VANHOUTEN, CHANG and crossed the line soon after it was isolated. Although this phenotype is especially frustrating because no genetic analyses can be done, the correlation of abnormal swimming behavior, abnormal mating,and abnormal ciliary plaques indicates a close connection between the mechanisms that control swimming and mating. CRONKITE (1976) reported such a connection when he demonstrated that pawn mutants, unable to swim backward because calcium does not cross the cell membrane, also cannot be chemically induced to mate. T h e roles of potassium, sodium and calcium ions in mating behavior and the relations between mating and swimming behavior deserve further investigation. If the plaques are sites of K + and/or Na' channels, is it possible to explain the stocks such as PaC and PaD which show paranoiac behavior, abnormal ion fluxes, andnormal plaques? Obviously they could carry mutations which affect the same ion channels, but at a level not visible ultrastructurally. Alternatively, these stocks could carry mutationsof other K + or Na' channels or of a Ca2+ channel or pump as suggested by HANSMA (1979) and SAIMIand KUNG (1980). Stock PaB, with large, stable ciliary plaques and atypical paranoiac behavior,has not been studied electrophysiologically or biochemically so that any speculation about membrane transport components is futile. T h e complex set of stimuli and response mechanisms that lead to swimming behavior in Paramecium result in ample opportunity for variety. Substantial and reliable differences are induced by mutations, but overlying thesechanges are a raft of poorly understood environmental and developmental influences. We have definedarange of behaviors and

626

B. J. Byrne, A. P. Tanner and P. M. Dietz

conditions where phenotypes these are relatively consistent. Since swimming behavior of Paramecium mirrorsthe electrophysiological properties of the membrane, careful observation of behavior can contribute to understanding membrane function. While the weaknesses of this type of analysis lie inthe variability of the reactions andtheneed toinfer membrane activity from the net response of 5000 cilia, great strength in the method arises from the numbers of cells that can beobserved and thediverse mutants available. The multiple alleles at the PaA and fna loci, and the close linkage of similar genes (PaA and PaC; PaE and f n a ) is reminiscent of characteristics of gene families reported frequently in eukaryotes. Thehuman globin gene families, reet viewed in detail by MANIATIS al. (1980) may be the best studied example. We suggest that at least some potassium and sodium channels are associated with the ciliary membrane plaques. If more extensive biochemical and physiological studies pinpoint precisely the lesions in stocks with abnormal plaques, we will have succeeded in assigning a specific funtion(s) to recognizable membrane components in situ.

DUNLAP, 1977 K., Localization of Ca channels in Paramecium caudutum. J. Physiol. 271: 119-133. DUTE,R., and C. KUNG,1978 Ultrastructure of the proximal region of somatic cilia in Paramecium tetraurelia. J. Cell Biol. 78: 451-464. ECKERT, and P. BREHM, R., 1979 Ionic mechanisms of excitation in Paramecium. Annu. Rev. Biophys. Bioeng. 8: 353-383. HANSMA, 1979 Sodium uptake and membrane excitation in H., Paramecium. J. Cell Biol. 81: 374-381. HANSMA, and C. KUNG,1976 Defective ion regulation in a H., class of membrane excitation mutants in Paramecium. Biochim. Biophys. Acta 436: 128-139. JENNINGS, H. 1906 Behavior$LowerOrganism. Indiana UniS., versity Press, Bloomington, Ind. KAMADA, 1929 Control of galvanotropism in Paramecium. J. T., Fac. Sci. Univ. Tokyo (Sect. 4) 2: 123-139. KAMADA, and H. KINOSITA, T., 1940 Calcium-potassium factor in ciliary reversal of Paramecium. Proc. Imp. Acad. Tokyo 16: 125-130. KUNG, 1971a Genic mutants with altered system of excitation C., in Paramecium aurelia. I. Phenotypes of the behavioral mutants. Z. Vergl. Physiol. 71: 142-164. KUNG, 1971b Genic mutants with altered system of excitation C., in Paramecium aurelia. 11. Mutagenesis, screening and genetic analysis of the mutants. Genetics 69: 29-45. KUNG, 1979 The biology and genetics ofParamecium behavior. C., pp. 1-26. In: Neurogenetics: Genetic Approaches to the Nervous ElseviedNorth Holland, System, Edited by X. 0. BREAKEFIELD. New York. We thank CHING KUNGfor kindly supplying stocks and helpful KUNG, and Y. SAIMI, C., 1982 The physiological basis of taxes in suggestions, ARNOLD SHILEPSKY THOMAS and A. VAWTERfor help Paramecium. Annu. Rev. Physiol. 44: 519-534. with statistical analyses, and BRUCE BYRNE assistance with data for 1985 Ca2+ channels of Paramecium: A KUNG, and Y. SAIMI, C., collection and manuscript preparation. The work was supported multidisciplinary study. Curr. Top. Memb. and Transp. 23: in part by a grant from Research Corporation and by U.S. Public 45-66. Health Service grant R 0 1 GM25298. Y. J. and KUNG, S. Y. CHANG, SATOW, VANHOUTEN H.HANSMA, C., 1975 Genetic dissection of behavior in Paramecium. Science LITERATURE CITED 188: 898-904. MANIATIS, E. F. FRITSCH, LAUER R. M. LAWN, T., J. and 1980 The ADOUTTE,A.,R. RAMANATHAN, LEWIS, R. DUTE,K.Y. R. M. R. molecular genetics of human hemoglobins. Annu. Rev. Genet. 1980 Biochemical studies LING, KUNGand D.L. NELSON, C. 14: 145-178. of the excitable membrane of Paramecium tetraurelia. 111. ProPLATTNER, F. MILLERand L. BACHMAN, H., 1973 Membrane teins ofcilia and ciliary membranes. J. Cell.Biol. 84: 717specializations in the form of regular membrane to membrane 738. attachment sites in Paramecium. A correlated freeze-etching ALLEN, D., 1978 Membranes of ciliates: ultrastructure, bioR. and ultrathin-sectioning analysis. J. Cell Sci. 13: 687-719. chemistry and fusion. pp. 657-763. In: Membrane Fusion, Y., SAIMI, 1986 Calcium-dependent sodium currents in ParameEdited by G. POSTEand G. L. NICOLSON. ElseviedNorthcium: Mutational manipulations and effects of hyper- and Holland, Amsterdam. depolarization. J. Membr. Biol. 92: 227-236. 1978a Behavior and the excitable BYRNE, J., and B. C. BYRNE, B. SAIMI,Y., and C. KUNG,1980 A Ca-induced Na-current in membrane in Paramecium. CRC Crit. Rev. Microbiol. 6: 53Paramecium. J. Exp. Biol. 88: 305-325. 108. SATOW, and C. KUNG, 1976a A mutant of Paramecium with Y., 1978b An ultrastructural correlate BYRNE, J., and B. C. BYRNE, B. increased relative resting potassium permeability.J. Neurobiol. of the membrane mutant Paranoiac in Paramecium. Science 7: 325-338. 199: 1091-1093. SATOW, and C. KUNG, 1976b A TEA+-insensitivemutant Y., and 1981 The sixth BYRNE, J., A. TANNER P. MUILENBURG, B. with increased potassium conductance in Paramecium aurelia. paranoiac locus in Paramecium tetraurelia. Genetics 97: 516a. J. Exp. Biol. 65: 51-63. S. CHANG, Y., and C. KUNG,1976 Selection and analysis ofa and SATOW, H. G. HANSMA C. KUNG,1976 The effect of Y., mutant of Paramecium tetraurelia lacking behavioral response sodium on Paranoiac-a membrane mutant of Paramecium. to tetraethylammonium. Genet. Res. 27: 97-107. Comp. Biochem. Physiol. 54A: 323-329. CRONKITE, L., 1976 A role of calcium ions in chemical inducD. SOKAL, F., and F. J. ROLF,1969 Biomety. pp. 374-375. W. H. R. tion of mating in Paramecium tetraurelia.J. Protozool. 23: 431Freeman, San Francisco. 433. SONNEBORN, M., 1970 Methods in Paramecium research. pp. T. CRONKITE, L., 1979 The genetics of swimming and mating D. 241-339. In: Methods in Cell Physiology Vol. 4, Edited by D. behavior in Paramecium. pp. 221-273. In: Biochemistry and PRESCOTT.Academic Press, New York. Physiology ofProtozoa, Ed. 2, Vol. 2,Edited by M. LEVANDOWSKY J., C. VANHOUTEN,S. Y. CHANG and KUNG,1977 Genetic analyses and S. H. HUTNER. Academic Press, New York. of Paranoiac mutants of Paramecium tetraurelia. Genetics 86: 1981 Control of ciliary activity in DOUGHTY, J., and S. DRYL, M. 113-120. Paramecium: An analysis of chemosensory transduction in a Communicating editor: S. L. ALLEN eukaryotic unicellular organism. Progr. Neurobiol. 16: 1-1 15.