Eur. J. Biochem. 163, 569-575 (1987) 0FEBS 1987

RNA-dependent DNA polymerase activity in Paramecium tetraurelia :what for? Michael D. KATINKA Centre de Gknttique Molkculaire, Centre National de la Recherche Scientifique, Gif-sur-Yvette (Received August 4/0ctober 27, 1986) - EJB 86 0843

Protein extracts from the protozoan ciliate Paramecium tetraurelia revealed high levels of RNA-dependent DNA polymerase activity (reverse transcriptase). Stable and constant during the somatic phase of the cell cycle, the reverse transcriptase activity quickly diminished following the completion of the sexual phases of the cell cycle: conjugation and autogamy. The Paramecium reverse transcriptase presented a number of common features with retroviral polymerases: (a) ability to copy synthetic templates such as poly(rCm) . oligo(dG) as well as mRNA; (b) sensitivity to various reverse transcriptase inhibitors such as HPA 23, suramin, phosphonoformate and ethidium bromide; (c) insensitivity to the action of other DNA and RNA polymerase inhibitors and, finally, (d) the requirement for divalent cations before the enzyme can function: either magnesium or manganese. Although the reverse transcriptase activity was not proven to be independent from one of the DNA polymerses in paramecia, its high activity predicts a role in the paramecia cell cycle. From what we are able to concieve today two possible roles could be envisaged. (a) Participation in the anlage macronucleus formation : micronuclear sequences are first transcripted and, after rearrangements of the RNA molecules, these are retrotranscribed into the macronuclear DNA molecules or (b) association with retrotransposons that participate in the movement of certain macronuclear sequences into the germ-line micronucleus.

The protozoan Paramecium tetraurelia, like all ciliates, presents the chracteristic of nuclear dimorphism with two types of nuclei cohabiting in the same cell: a macronucleus (Mac) and two micronuclei (mic). The somatic polyploid (z800 n) Mac controls the cellular phenotype and, the transcriptionally quasi-inert diploid mic are the germinal reserve. In the sexual phases of the Paramecium cell cycle, autogamy and conjugation, the two mic undergo meiosis and form four haploid progametes. After mitosis, seven of eight progametes degenerate and another round of mitosis results in the formation of two mature gametes per paramecia. The gametes either fuse inside the mother cell (autogamy) or do so after reciprocal exchange of one of them with its counterpart in a sister mating cell (conjugation). During these processes the old Mac are gradually fragmented and degraded. Following two rounds of mitosis the new diploid zygote develops into Mac and rnic (for a review see [I]). The development of the anlage Mac was mostly studied in hypotrichous ciliates [2]and the holotrichous Tetrahymena [ 3 , 41. These studies show that in Mac formation both amplification and rearrangements (deletions and fragmentation) of DNA sequences occur. If the mechanism of the anlage Mac development is certainly complex, two hypothesis might nevertheless be envisaged: either the entire mechanism operates through DNA molecules only, or both RNA and DNA are involved. In this latter possibility RNA would serve Correspondence to M. D. Katinka, Centre de Gtnetique Moltculaire, CNRS, 1 Avenue de la Terrasse, F-91190 Gif-sur-Yvette, France Abbreviations. Mac, macronucleus; mic, micronucleus; PRT, Puramecium reverse transcriptase; RT, reverse transcriptase.

as a transient intermediate between the initial mic chromosomes and the terminal Mac sequences. To copy RNA molecules into DNA requires as a sine qua non condition the expression in Paramecium of an RNAdependent DNA polymerase activity. In this work 1 show that in Paramecium such an activity is present in important amounts. The characteristics of the Paramecium reverse transcriptase (PRT) are remarkably similar to those of retroviral polymerases. During the paramecia vegetative cell cycle the level of PRT activity was constant and high. This level did not augment in the sexual phases and rapidly decreased afterwards, that is, after formation of the anlage Mac. In the light of these results we will discuss the eventual role of PRT in ciliates. MATERIALS AND METHODS Materials

Avian myeloblastosis virus reverse transcriptase was from

P. H. Stehelin, Basel. Pancreatic RNase A, a-amanitin, actinomycin D and all nucleotide triphosphates were purchased from Boehringer Mannheim. [ L X - ~ ~ P I ~ and GTP [cz-~’P]~TTP( ~ 4 0 Ci/mmol) 0 were from Amersham. Aphidicolin, phosphonoformic acid and ethidium bromide were purchased from Sigma, suramin from Bayer (Germanin, Bayer 205). The reverse transcriptase inhibitor HPA 23 was a kind gift from Drs. J. C. Chermann and F. C. Sinoussi at the Pasteur Institute. All templates and 12- 18-nucleotide primers were from PL-Pharmacia. Poly(A)-rich RNA from Paramecium primaurelia, purified as in [5], was generously

570 donated by Dr. F. Caron. Sucrose, RNAase-free grade, was from BRL (Bethesda). All the other chemicals used in this work were from Sigma, BDH or Merck, analytical grade or better. Strains and culture conditions Paramecium tetraurelia wild-type strain, stock d4-2 (mating types VIIS and VII12B), described by Sonneborn [6], was grown according to usual procedures [7] in either (a) scotch-grass infusion, bacterized with Aerobacter aerogenes and supplemented with 2 pg/ml p-sitosterol at 28 "C, or (b) in the crude axenic medium as described by Thiele et al. [8], but for the replacement of 10 g/ml protease peptone and 5 g/ml tryptic agar base by 9 g/ml low-fat milk (NestlC, minoform), at 25 "C. Autogamy and conjugation were monitored by histological coloration methods as described by Dippel [9] or by the orcein/acetic acid technique [lo]. Rapid reverse tvanscriptase ( R T ) assay Cultures paramecia were collected by centrifugation or by filtration through 8-pm nitrocellulose filters. Paramecia grown in axenic medium were washed three times in 10 mM Tris/HCl (pH 7.2), and paramecia grown in scotch-grass infusion, after two washes in the Tris buffer, were left for 2 3 h in the same buffer and further washed twice. The cultures were resuspended in 0.0005 vol. Tris, adjusted to 1 mM ethylenediamine-tetraacetic acid . Naz (EDTA), 1 mM freshly prepared phenylmethylsulfonyl fluoride, 14 mM 2-mercaptoethanol, 0.2% Nonidet P40 (NP40) and approximately 0.1 vol. acid washed 0.45-mm-diameter glass beads. Vortexing for 30 s at 4°C was followed by 2 min sonication at 50% duty cycle in a cooled Cup Horc probe at maximum output (Ultrasonics; model W-225R). After elimination of the cellular debris by centrifugation for 15 min at 13000 x g at 4"C, the supernatant was developed on a 15-50% sucrose gradient in 20 mM Tris/HCl (pH 7.5); 20 mM NaCl; 1 mM EDTA; 0.01% NP40; 14 mM 2-mercaptoethanol and 0.5 mM phenylmethylsulfonyl fluoride. The gradient was centrifuged for 20 h at 40000 rpm and 2°C in a Beckman SW-41 rotor. Fractions of 0.8 ml were collected and directly assayed for RT activity. The assayed used was an adaptation of similar tests devised by Goff et al. [ll] and Mellor et al. [12]: briefly, 10 p1 of each fraction were mixed with 20 pl SO mM Tris/HCl (pH 8.3), 60 mM NaC1, 3 mM MgC12, 1 mM MnC12, 10 mM dithiothreitol, 0.05% NP40; 0.1 unit RNasine; 0.025 A260 unit poly(rA) . oligo(dT) in equimolar amounts and [a32P]dTTP (1000- 5000 counts min- pmol-'). The reaction mixture was incubated for 1 h at 25°C after which it was adsorbed in total on a Whatman DE81 filter, washed four times in 5% NazHP04,twice in water, rinsed in ethanol, dried and counted in a scintillation counter. Protein concentration was determined by a Bio-Rad Bradford assay kit. Estimation of reverse transcriptase activity in autogamy and conjugation Autogamous paramecia were grown in mass culture for 22 - 26 generations and then induced into autogamy simply by starvation. Monitored by histological coloration under a bright-field optical microscope, fractions of the culture were removed at different times and assayed for reverse transcriptase activity as described above. The beginning of autogamy was defined as the observation of the first morpho-

m

r

f

Fraction nvmbei

Fig. 1 . Sucrose gradients of Paramecium total, soluble protein extract, assayed for reverse transcriptase activity. Poly(rA) . oligo(dT) and [ E ~ ' P ] ~ T T Pwere used as substrate, as described in Materials and Methods. Polymerase assays were performed on either vegetative growing paramecia ( 0 )or paramecia that started autogamy 12- 14 h ealier ( W )

logical alterations of the macronuclei [13]. Autogamy was quite impossible to synchronize to a precision of better than 3 h. Conjugation was much more controlable and the precision in following its different stages could be estimated to about 20- 30 min. The different phases of autogamy and conjugation are identical, but for the exchange of gametes in the latter. In conjugation the respective mating types were rendered reactive by starvation of autogamous cells that were allowed to complete four cell divisions. The reactive cells were mixed (beginning of conjugation) and 1.5 h later were fed with medium containing 0.5 g/1 2 - 10 pm reduced iron ore, previously sonicated for 2 min in a Cup Horn probe at maximum output. Only non-conjugating paramecia from digestive vacuoles, and thus could be eliminated from the culture by a magnet 15 min after feeding (S. Grandchamp, unpublished). Otherwise, assays for reverse transcriptase activity were preformed as for autogamy and are described above. RESULTS Reverse transcriptase activity in Paramecium Autogamous paramecia were cultured for 6 - 7 generations and the total protein extract was developed on sucrose gradients as described in Materials and Methods. Using poly(rA) . oligo(dT) as template . primer, the incorporation of [a-32P]dTMPwas measured, the sucrose gradient fractions having served as the source of enzyme. An example is shown in Fig. 1. In this experiment an RNA-dependent DNA polymerase activity was readily detectable around fraction 10 and in fraction 2. The activity in fraction 2 was always 8 - 10fold lower than in fraction 10. When the same experiment was performed with paramecia that were induced into autogamy and were implicated in the process for approximately 12 h, the PRT activity measured was some 5-fold lower (Fig. 1). This difference in activity was measured when the same number of paramecia were used. It was 2 - 3-fold more important when PRT activity was compared relative to the protein concentration. When poly(dA) was used as template instead of poly(rA), incorporation of radioactive nucleotides was 3-fold lower with

571

P

I

1

I

1 Fraction 10

2

3

4

5

6

Time after extraction (days)

Fig. 2. Evolution of reverse transcriptase activity in sucrose gradient fractions 2 and 10. Assays were performed with either: ( 0 )poly(rA) or (0)poly(dA) as template. The 10-fold increase in incorporation as compared with Fig. 1 was solely due to the higher specific activity of the radiolabel used

fraction 10, but only 1.2-fold lower with fraction 2 (see Table 1 and Fig. 2). The experiment was repeated the following day in the same conditions, fractions 2 and 10 being kept at 0°C. The ratio of activity using either polyriboadenylate or polydeoxyriboadenylate templates, rA/dA was constant for fraction 10,at about 3 (Fig. 2) and remained so day after day. For fraction 2 this was not true, and the rA/dA activity ratio varied from 1.2 on the first day to 0.2 on the sixth (Fig. 2). The PRT activity in fraction 10 involves perhaps only one molecular species, that of fraction 2 contained an additional and possibily independent enzymatic activity: a DNA-dependent DNA polymerase. The augmentation of this last activity with time is not explicable at the present time, but a tentative explanation would be the inactivation of an inhibitor or repressor. For commodity reasons only fraction 10 was used in most of the experiments described below. Whether paramecia were cultured in bacterized scotchgrass or in axenic media the measured reverse transcriptase activities were similar. However, enzyme from paramecia grown in the axenic medium was in lesser quantity, had a shorter life span during assays and could not be kept for more then 24 h after extraction. This observation could be due to a higher level of proteases in the too-well-fed paramecia in axenic medium. When protein extraction, sucrose gradients and RT assays were done without the detergent NP40, results were similar to those obtained in its presence (Table 1). This tends to show that PRT is probably not associated with virus-like particles of retroviruses [I41 or yeast Ty [12]. The detergent, necessary to unmask RT activity, by destroying the viral structures with which it is associated, was without effect on PRT activity. RT activity per paramecium in exponentially growing cells was calculated. The unit was defined as for retroviral reverse transcriptase [15].Determined experimentally by comparison to known AMV RT enzyme activity, the PRT activity was found to be as high as 50 pU per paramecium. Use of’dcferent templates and primers

RNA-dependent DNA polymerase activity is not limited to retroviral, viral (hepatitis or cauliflower mosaic viruses) or

to retrotransposon (yeast Ty or Drosophila copia) enzymes. RT activity could be demonstrated with certain DNA-dependent DNA polymerases, such as in E. cofi pol I [16]or in human HeLa cells pol y [17]. Thus, the better to characterize the PRT activity and exclude a ‘hitch-hike’ activity of one of the Paramecium DNA-dependent DNA polymerases, a number of experiments were performed. The use of different templates was tested first. Like all DNA polymerases, PRT requires a primer to copy the template, and without one no incorporation of radioactive dTMP could be noted with poly(rA) as template (Table 1). When the RNA template was destroyed by RNase A prior to the addition of PRT, no activity was detected (Table 1). Different primer . template combinations were used to test the PRT. Activity using poly(rC) . oligo(dG) was only 0.2fold as high as with the poly(rA) . oligo(dT) reference (see Table 1). This result was in good agreement with the one published for Harvey-MuSV RT [18].On the other hand, with AMV RT we have found a higher activity, about a 2.8-fold increase, with respect to the poly(rA) reference template (Table 1). Gerard et al. [I81 have described a template that was specific to retroviral polymerases, and permitted the exclusion of all DNA-dependent polymerases: poly(mrC), with cytosine methylated on the 2’position. PRT did quite well in copying this template (0.15times the reference; Table 1). This result was again in good agreement with the published values for Harvey-MuSV RT [18],but quite different from the value we found for AMV RT (Table 1). Finally PRT copies quite efficiently a poly(dA) template (0.3- 0.4-fold compared to the poly(rA) reference). Retroviral RT copy poorly poly(dA) templates, the rA/dA incorporation ratio being between 15 and 60 for different polymerases (Table 1 and [18]). Retroviral RT readily retrotranscribe polyadenylated mRNA, if the reaction mixture is supplemented with the appropriate oligo(dT) primer. This was verified with PRT assayed with poly(A)-rich Paramecium RNA, primed with oligo(dT) at 20: 1 (w/w) ratio. Qualitatively incorporation of [ u - ~ ~ P I ~ Gwas M Pregistered (Table l), but any quantitative analysis would be illusory. Use of various inhibitors

Another method employed to probe the characteristics of RT-like enzymes is the use of inhibitors. The first used were inhibitors specific to transcription. Actinomycin D inhi bits transcription by binding to the guanine residues of the template. As shown in Table 1, at a concentration of 100 pg/ ml actinomycin D did not inhibit RT activity with either poly(A)-rich RNA, or with poly(rA) . oligo(dT) templates. The slight deficit in radiolabel incorporation was probably due to the presence of up to 3 -4% ethanol used to dissolve actinomycin. The same deficit was similarly noted with AMV RT (Table 1). a-Amanitin, essentially an inhibitor of eukaryotic RNA polymerase B, did not affect radiolabel incorporation when used in concentrations of up to 200 pg/ml. Aphidicolin is an inhibitory of mammalian DNA-dependent DNA a polymerases [19]. When employed at concentrations of up to 50 pM, aphidicolin failed to present any inhibitory effect on PRT activity (Table 1). Numerous substances that specifically inhibit retroviral polymerases have been described in the literature. In this work I studied the effect of four different molecules : (a) ammonium5’-tungsto-2’-antimonate (HPA 23 [20]); (b) suramin [21];(c) ethidium bromide [22] and (d) phosphonoformate [23].HPA 23 and phosphonoformate have a strong affinity for proteins

572 Table 1. Properties of Paramecium reverse transcriptuse Assays were done as described in Materials and Methods. Additions to the reaction mixtures were 100 pg/ml RNase A (added 15 min before the enzyme and in the absence of RNasine); I00 pg/ml actinomycin D; 200 pg/ml a-amanitin and 50 pM aphidicolin. The quantity of template . primer used in each assay was of 0.025 A Z b Ounit, with an equimolar template-to-primer ratio Radioactive nucleo tide

Template primer

Conditions

Relative rate of incorporation fraction 10

fraction 2

AMV RT

1 0.02 f 0.01 0.96 & 0.03 0.05 f 0.02 0.04 f 0.03 0.88 f 0.03 1.05 k 0.05 0.93 f 0.04 0.02 f 0.01 0.3 k 0.05 0.26 f 0.02 0.2 k0.03 0.15 0.04 2 k0.3 1.85 5 0.2

1 (0.2 f 0.05)b n.d. n.d. n.d. n.d. 0.9 f 0.03 1.02 & 0.05 0.91 f 0.05 n.d. 0.9 f 0.08 0.81 k 0.05 n.d. n.d. n. d. n.d.

0.04 f 0.01 0.92 f 0.05 0.03 f 0.01 0.05 k 0.03 0.9 k 0.02 1.1 f 0.06 0.96 f 0.02 n.d. 0.05 k 0.01 n.d. 2.8 f 0.4 2.4 k0.3 2.7 f 0 . 5 1.87 f 0.3

~~

Poly(rA) . oligo(dT)

Poly(dA) . oligo(dT) Poly(rC) . oligo(dG) Poly(mrC) . oligo(dG) Paramecia poly(A)-rich RNA" . oligo(dT)

dTTP

dGTP dTTP dGTP

complete -Mg2+-Mn2+ - NP40 - oligo(dT) + RNas A + actinomycin D +a-amanitin + aphidicolin -dTTP complete + aphidicolin complete

+ actinomycin D

1

a 1 pg poly(A)-rich P . primaurelia RNA and 0.05 pg oligo (dT) were used as template-primer. In addition to [E-~' P]~G TP, 0.1 mM unlabelled dATP, dCTP and dTTP were used in the reaction mixture. The value in parenthesis designates the activity in fraction 2, relative to fraction 10.

[20, 231, while ethidium bromide and suramin as intercalating agents probably inhibit RT activity by binding to the template [21,22]. As shown in Fig. 3, all four molecules were inhibitory for PRT activity. The most potent inhibitor was HPA 23, with which 100% inhibition could be achieved at concentrations as low as 1 pM. The second best was suramin (100% inhibition at 40 - 50 pM) and less efficient inhibition was observed with ethidium bromide and phosphonoformate (Fig. 3). With these last two molecules it was impossible to attain 100% inhibition even at concentrations of 1 - 2 mM. It was shown that not all retroviral RT have the same sensitivity to a certain inhibitor. For example, to obtain 100% inhibition with phosphonoformate, 80 pM was necessary for the Rauscher-MuLV RT and up to 1 mM or more for AMV RT 1231. I studied the dose-response curve for two inhibitors: HPA 23 and ethidium bromide, on AMV and Paramecium RT. AS shown in Fig. 4,0.5 pM ethidium bromide was needed to attain 50% inhibition of AMV RT activity, but up to 90 pM was needed for PRT. By contrast, PRT was more sensitive to HPA 23 than AMV RT: 1 pM and 10 pM inhibited PRT to 50% and 100% respectively, while 50 pM and 150 pM were necessary to attain this level of inhibition with AMV RT (Fig. 4). The utilization of the same four inhibitors in vivo was unsatisfactory. We have no idea if suramin and HPA 23 were able to penetrate the cell otherwise than through the digestive vacuoles. Suramin at concentrations of up to 2 mM did not affect in any way Paramecium growth or the development of the anlage Mac following autogamy. The other three inhibitors, at low concentrations (0.5 pM ethidium bromide; 10 pM HPA 23 and 100 pM phosphonoformate), did not affect in any way normal development of the paramecia vegetative and sexual cell cycles. At higher concentrations 5 pM ethidium bromide rapidly arrested cellular division and, respectively, 100 pM and 10 mM HPA 23 and phosphonoformate induced paramecia to lyse after less than 24 h. It should be noted that exponentially growing, daily cloned paramecia were much

less sensitive to HPA 23 than a stationary culture, and kept dividing at a normal rate for up to 5 days. Divalent cation requirements RT, like all polymerases, require divalent cations to function. PRT in the absence of either magnesium or manganese did not function (Table 1) but activity could be readily restored by their addition. When assayed at different concentrations (Fig. 5) an optimum for PRT activity was observed at 2-4 mM magnesium and 0.3 -0.5 mM manganese. This is in good agreement with the optimal concentrations published for retroviral polymerases [14]. Optima of activity were at manganese concentrations 2 - 60-fold lower than those found for magnesium [14]. Moreover, the maximum of PRT activity in the presence of manganese was some 1.5-fold lower than with magnesium, a result quite similar to the published data for AMV RT (2-fold lower; [14]). As already noted above, different retroviral RT show a variability in their sensitivity to the same inhibitor. This is again true for the relative activity of several RT in either magnesium or manganese. For example, Moloney-MuLV RT presents an activity 4-fold superior at its manganese optimum of 1 mM than at its magnesium optimum of 2 mM; A m , on the other hand, has an activity 2-fold superior at 6 mM magnesium than at 1 mM manganese [141. Cell-cycle and reverse transcriptuse activity As mentioned above, PRT activity was present in exponentially, vegetative growing cells, but quickly declined after autogamy completion. In order to follow PRT activity along the vegetative multiplying cell generations, sucrosegradient-purified protein extracts (fraction 10) from paramecia, refed daily with fresh medium, were assayed. Autogamous paramecia cultured in fresh medium, following of latency of 8 - 12 h, had an average generation period of

573

I

2

3 4"'I0 20 Concentration (Mm)

30

'$0

go

Fig. 5. Effect of various magnesium (@) and manganese (0) concentrations on Paramecium reverse transcriptuse activity rlfter incubation under the conditions described in Fig. 3. The vertical bars are thc variations in activity in three independent experiments

about 6 h. As shown in Fig. 6A, two generations after autogamy, multiplying paramecia exhibited an RT activity already at its maximum level. This level was constant for at least 40 generations. Cells cultivated beyond this number of generations saw their RT activity level slowly decline. This was not very surprising. In paramecia of the tetraurelia species, grown for more than 40 - 50 generations, cell division I 1 J becomes slower and more erratic, autogamy is less and less 30 60 9'0 110 3'0 60 SO 120 efficient, resulting in reduced survival of normal progeny. Our Time (rnin) observations might be correlated to the senescence program Fig. 3. EjjLcct of various concentrations of specific reverse transcriptuse of paramecia [l]. inhibitors on paramecia D N A polymerase activity. Poly(rA) . oligo(dT) To follow the evolution of PRT activity in autogamy, cells and [ E ~ ' P ] ~ T T Pwere used as substrates. Inhibitors were added to were induced into it after 22-26 generations of vegetative the reaction mixture 10 min prior to the enzyme (sucrose gradient growth. Activity was monitored after the first morphological fraction 10). Assay mixtures were incubated for 30, 60, 90 and changes in the Mac appeared. Approximatively 24 h later, 120 min. Concentration ofinhibitors: (-)control; ( W ) 2 mM; (0) extinction of PRT activity was registered. At this stage 0.5 mM; ( A ) 0.1 mM; ( @ ) 0.05 mM; (V)0.01 mM; (A)5 pM; (+) autogamy was terminated and the anlage Mac formed. At no 1 pM and ( 0 )0.4 pM stage during autogamy was any increase in RT activity noted (Fig. 6B). Activity was back to its maximum level 13-16 h after cells were fed with fresh medium (Fig. 6B). The same type of experiment was performed with conjugating cells. As noted in Materials and Methods, conjugation was synchronized with more accuracy than autogamy. The accuracy achieved was of approximatively 20 min and enabled a more intimate assessment of the PRT activity rate during the anlagen micronuclei (mic) and Mac development. As shown in Fig. 6C, no variation in the rate of PRT activity was noted during the process of conjugation. As observed during autogamy, only at least 12 h after the beginning of conjugation was a decrease in activity observed. It should be noted that the time scale in 6B (autogamy) and 6C (conjugation), was not the same. Therefore, to compare the experiments 4-6 h should be added to the time scale of 6C. In conclusion, PRT activity was constant and at high levels during vegetative growth, a level maintained but by no means increased in the sexual phases of autogamy and conjugation. Concentration of inhibitor (M) A rapid decrease and extinction of activity was observed once these processes were completed. Fig. 4. Dose-response effect of two inhibitors: ethidium bromide (@, 0 ) a n d H P A 2 3 ( A , A)onParamecium (@, A ) , o r A M V ( O , A ) reverse transcriptuse. after incubation of the reaction mixture (as in Fig. 3 ) ,for 60 min in the presence of various concentrations of the inhibitors. The vertical bars are the maxima and minima of three independent experiments

DISCUSSION Vegetatively growing Paramecium cells were shown to contain RT-like activity of as much as 50 pM enzyme/cell.

574

5

l j - - : r l P)

m c)

.-P b

0 ~

; n

10 30 Average no. of generations following autogamy

8 12 16 20 24 28 30 34 Time after beginning of autogarny(h)

38

c.

03 E ?

$ 0 c!?

c

7 Q)3 0 N

=a SI

L%

0

3.5

4

5

5.5

5.7

6

12

Time after beginning of conjugation ( h )

Fig. 6. Evolution uf Paramecium reverse transcriptase activity in autogamy and conjugation. All assays were done for 60 min as described in Materials and Methods and in Fig. 3. (A) Activity in vegetative growing paramecia. Cells that had completed autogamy, were mass-cultured, daily passaged in fresh medium and daily assayed for activity. (B) Reverse transcriptase activity after induction of autogamy ( 0 ) .The determination of the beginning of autogamy is as described in the text. 24 h after beginning of autogamy, cells were resuspended in fresh medium and assayed for activity ( 0 ) .(C) Activity in conjugation. The main stages of conjugation coincide with those of autogamy. They are shown in the upper part of the figure, and were compiled from Grandchamp and Beisson [15]. (M) Mac; (m) mic; (pc) paroral cone; (zn) zygotic nucleus formed after fusion between gametes of two mating cells (conjugation) or fusion of the two gametes in the same cell (autogamy); (ms) mitotic spindle; (Mft old fragmented Mac; (Ma) anlage Mac

Stable and constant during the vegetative phases of the cell cycle, the PRT level rapidly decreased following the completion of the sexual phases: autogamy or conjugation. PRT had features in common with retroviral polymerases : capacity to use specific templates, resistance to a number of DNA and RNA polymerases inhibitors, sensitivity to specific RT inhibitors and the requirement of magnesium or manganese to function. One peculiarity of the paramecia enzyme was its capacity to copy poly(dA) at a rate of about 0.35, compared to poly(rA). This difference from retroviral enzymes that poorly copy poly(dA) [0.015-0.06 x the rate of poly(rA)] could be explained by either: (a) contamination by a DNA-dependent DNA polymerase or (b) an intrinsic characteristic of the enzyme. One must be aware that the present study does not unequivocally confirm the presence of an independent RT in paramecia, and only further purification of the activity and the other DNA polymerases would provide a clear answer. For seveal years data have accumulated, showing the large spectrum of phenomena associated with RT [24- 261. Thus, RT were shown to be involved in the replication of certain DNA viruses, such as HBV and C a m , the formation of pseudogenes, distribution of semirepetitive sequences, like the human Alu sequences. They are associated with retrotransposons, like the yeast Ty, Drosophila copia and Dictyostelium DIRS-1 [27]. Most recently RT-like activities were registered in the mould Podospora anserina [28] and the hypotrichous ciliate Stylonichia [29]. Has the RT activity in paramecia a role and if so what is it? One possibility for such a role is represented in the model of Fig. 7. Mac formation, following autogramy or conjugation, involves extensive sequence rearrangements of type I

(DNA breakage) and I1 (DNA splicing) [4]; both explicable by RNA splicing and maturation. Most of the genes isolated from ciliates do not contain introns, the exception being rDNA genes [30] and a conjugation-induced Tetrahymena gene [31]. Has splicing occurred during the Mac development? Amplification of the Mac DNA might also be explained in part by the rate of transcription of the mic chromosome. In part only, because it was shown that Mac sequences are amplified in the anlage Mac, further amplification occurs during the first two rounds of mitosis [32]. The one contradiction between the proposed model and published data arises with the results of Yao et al. [33] for T. thermophila, showing that during Mac formation certain sequences are eliminated stepwise and that DNA intermediates could be isolated. If PRT participates in Mac formation one could expect its induction just prior the sexual phases, or the enzyme level to be constant during the vegetative phases, decreasing rapidly after termination of autogamy or conjugation. Since in P . tetraurelia conjugation can be induced only two generations after autogamy, the constant enzyme level could be explicable; however, only further studies on the role of the enzyme and its purification will provide an answer. An alternative role for PRT activity could be proposed in relation to the recent discovery of transposon-like elements bounded by telomeric C4A4 sequences in the mic of Tetrahymena [34] and Oxytricha [35]. Thus RT might participate in retrotransposon movement into the mic, introducing certain Mac sequences, and thus allowing gene shuffle to occur in the germinal mic. Moreover, the evolutionary distance between the various ciliates is so big that the RT activity might participate in the, one gene-sized (2-20 kbp; [2])

575 cM i -T .

DNA

Transcription1 5'

I

' (Au to$\>,jl

3'

II)

i / s p Iicing

I I

Reverse transcriptase

RNase H 5 ' 3

~

I -

DNA polymerase (reverse transcriptase?)

DNase

T-Mac

3. Brunk, C. F. (1986) Rev. Cytol. 99,49-83. 4. White, T. C. & Lyman-Allen, S. (1986) J . Protozool. 33, 30-38. 5. Meyer, E., Caron, F. & Guiard, B. (1984) Biochimie (Paris) 66, 403-412. 6. Sonneborn, T. M. (1975) Trans. Am. Micros. SOC.94, 155- 178. 7. Sonneborn, T. M. (1970) Methods Celf. Physiol. 4, 241 -339. 8. Thiele, J., Hormer-Schmid, O., Wahl, J., Kleefeld, G. & Schultz, J. (1 980) J . Protozool. 27, 1 1 8 - 121. 9. Dippel, R. (1955) Stain Technol. 30, 60-71. 10. Gabe, M. (1968) Techniques histologiques, p. 704, Masson, Paris. 11. Goff, S., Traktman, P. & Baltimore, D. (1981) J . Virol. 38, 239 248. 12. Mellor, J., Malim, M. H., Gull, K., Tuite, M. F., McReady, S., Dibbayawan, T., Kingsman, S. M. & Kingsman, A. J. (1985) Nature (Lond.) 318, 583 - 586. 13. Grandchamp, S. & Beisson, J. (2981) Dev.Bid. 81, 336-341. 14. Verma, I. M. (1977) Biochim. Biophys. Actu 473, 1 - 38. 15. Houts, G. E., Miyagi, M., Ellis, C., Beard, D. & Beard, J . M. (1979) J. Virol. 29, 517-522. 16. Karkas, J. D. (1973) Proc. Natf Acad. Sci. USA 70, 3834-3838. 17. Spadari, S. & Weissbach, A. (1974) J . Biol. Chew!. 249, 58095815. 18. Gerard, G. F., Rottman, F. &Green, M. (1974) Biuchemisfry 13, 1632-1541. 19. Wist, E. & Prydz, H. (1979) Nucleic Acids Rex. 6, 1583- 1590. 20. Chermann, J. C., Sinoussi, F. C. & Jasmin, C . (1975) Biochim. Biophys. Res. Commun. 65, 1229- 1236. 21. Declerq, E. (1979) Cancer Lett. 8, 9 - 22. 22. Fridlander, B. & Weissbach, A. (1971) Proc. Nut1 Acud. Sci. USA 68, 3116-3119. 23. Sundquist, B. & Oberg, B. (1979) J . Gen Virol. 45, 273-281. 24. Baltimore, D. (1985) Cell 40, 481 -482. 25. Rogers, J. (1986) Nature (Lond.) 31Y, 725. 26. Varnius, H . E. (1985) Nature (Lond.) 314, 583-584. 27. Cappello, J., Handelsman, K. & Lodish, H. F. (1985) Cell 43, 105- 113. 28. Steinhibler, W. & Cummings, D. J. (1985) Curr. Genet. 10, 389392. 29. Lipps, H. (1985) Curr. Genet. 10, 239-243. 30. Cech, T. R. (1986) Cell44,207-210. 31. Martindale, D. W., Martindale, H. M. & Bruns, P. J. (1986) Nucleic Acids Res. 14, 1341- 1354. 32. Berger, J. D. (1973) Chromosoma (Berl.) 42, 247-268. 33. Austerberry, C. F., Alles, C. D. & Yao, M.-C. (1984) Proc. Natl. Acad. S C ~U. S A 81,7383-7387. 34. Cherry, J. M. & Blackburn, E. H. (1986) Cell43, 747-758. 35. Herrick, G., Cartinhour, S., Dawson, D., Ang, D., Sheets, R., Lee, A. & Williams, K. (1985) CeN43, 759-768.

!

Telomers

DNA

Further amplification?

Fig. 7. Model .for Paramecium macronuclear sequence formation. Micronuclear chromosomes are transcribed between specific initiation and termination sites. After RNA rearrangements (splicing and fragmentation) the molecules are retrotranscribed into doublestranded DNA molecules, which can undergo further amlification. These molecules can constitute the mature polyploid macronuclear chromosomes. T, telomeric C4A2 sequences. See the text for further discussion

hypotrichous Mac, development [29] and possess a totally different role in holotrichous ciliates. Only further research will tell of the possibility that a role for RT in the natural development program of a eukaryotic cell, such as paramecia, might be highlighted. I wish to thank Jannine Beisson for constant interest throughout this work, Simone Grandchamp for introducing me to the subtilities of the Paramecium cell cycle and culture, and Annik Prat, Franqois Caron and Bernard Guiard for helpful discussions and for reviewing the manuscript.

REFERENCES 1. Nanney, N. (1980) Experimental ciliatology, Wiley, New York. 2. Kraut, H., Lipps, H. J. & Prescott, D. (1986) Rev. Cytol. 99, 1 28.