Epigenetic Mating-Type Inheritance

ARTICLE doi:10.
1038/nature13318
Genome-defence small RNAs exapted for

epigenetic mating-type inheritance
Deepankar Pratap Singh1,2, Baptiste Saudemont1,2{, Gérard Guglielmi1, Olivier Arnaiz3, Jean-François Goût4{, Malgorzata Prajer5,
Alexey Potekhin6, Ewa Przybòs5, Anne Aubusson-Fleury3, Simran Bhullar1, Khaled Bouhouche1{, Maoussi Lhuillier-Akakpo2,7,
Véronique Tanty1, Corinne Blugeon1, Adriana Alberti8, Karine Labadie8, Jean-Marc Aury8, Linda Sperling3, Sandra Duharcourt7
& Eric Meyer1
In the ciliate Paramecium, transposable elements and their single-copy remnants are deleted during the development of
somatic macronuclei from germline micronuclei, at each sexual generation. Deletions are targeted by scnRNAs, small RNAs
produced from the germ line during meiosis that first scan the maternal macronuclear genome to identify missing sequences,
and then allow the zygotic macronucleus to reproduce the same deletions. Here we show that this process accounts for the
maternal inheritance of mating types in Paramecium tetraurelia, a long-standing problem in epigenetics. Mating type E
depends on expression of the transmembrane protein mtA, and the default type O is determined during development by
scnRNA-dependent excision of the mtA promoter. In the sibling species Paramecium septaurelia, mating type O is determined
by coding-sequence deletions in a different gene, mtB, which is specifically required for mtA expression. These independently
evolved mechanisms suggest frequent exaptation of the scnRNA pathway to regulate cellular genes and mediate transgene-
rational epigenetic inheritance of essential phenotypic polymorphisms.
Ciliates are complex unicellular eukaryotes that use different types of nuclei RNA-mediated genomic subtraction thus reproduces the deletions
within the same cytoplasm to separate germline and somatic functions1. observed in the parental MAC and can account for epigenetic inheritance
The diploid micronuclei (MICs) undergo meiosis to provide gametic nuclei of alternative rearrangement patterns, such as retention of a given IES in
during sexual events, but their genome is not expressed. Genes are ex- the MAC14,15, or deletion of a given gene16, across sexual generations.
pressed from the polyploid macronucleus (MAC), which is not transmit- In Paramecium, conjugation (the reciprocal fertilization of cells of
ted across sexual generations. After meiosis and fertilization, the parental opposite mating types) does not allow any significant exchange of cyto-
MAC is lost and replaced by a new one that develops from a mitotic copy plasm between the mates, so that the pools of scnRNAs produced during
of the zygotic nucleus. Macronuclear development involves extensive meiosis are independently sorted in each cell. After fertilization, each
rearrangements of the germline genome, including the elimination of developing MAC will thus reproduce the particular rearrangements pre-
virtually all transposable elements and other repeats2. Furthermore, in sent in the old MAC of its own cytoplasmic parent. This mechanism
P. tetraurelia ,45,000 short, single-copy internal eliminated sequences might underlie the maternal (cytoplasmic) inheritance of mating types
(IESs) are precisely excised from coding and non-coding sequences3,4 by in P. tetraurelia, one of the earliest cases of transgenerational epigenetic
the domesticated transposase Pgm5. IESs are invariably flanked by two inheritance in any eukaryote.
59-TA-39 dinucleotides which recombine into one after excision4. A short Although mating types were discovered in 1937 (ref. 17), so far the only
consensus adjacent to the TAs (59-TAYAGYNR-39) is reminiscent of the available test relies on the ability of type O (odd) to agglutinate with type E
ends of Tc1/mariner elements4,6, and a recent study provided support for (even) when vegetative cells become sexually reactive, a physiological state
the hypothesis that IESs are degenerate remnants of ancient transposable induced by mild starvation. Agglutination is a prerequisite for conjugation
element insertions3,7. and occurs on contact through adhesion of ciliary membranes. Mutational
All intragenic IESs identified so far must be excised to reconstitute func- analyses showed that several genes are specifically required for expression
tional genes in the MAC. However, the poorly conserved IES end con- of type E18–20; mutations in these genes result in a constitutive O pheno-
sensus is not sufficient to specify the excision pattern genome-wide4, and type, which thus seems to be a default state. In P. tetraurelia, mating types
a specific class of small RNAs is required to identify some IESs on the are not genetically determined in the MIC21. Each new MAC becomes
basis of their absence from the parental MAC8–10. In the current ‘genome determined for one type during its development, and remains the same
scanning’ model, scnRNAs are produced from most of the germline gen- throughout vegetative growth of the derived clone. The O/E alternative is
ome during MIC meiosis11,12 and are then filtered by pairing interactions not random, but maternally inherited; experiments showed that mating-
with nascent transcripts in the parental MAC13, which acts as a sponge to type determination in the developing zygotic MAC is controlled through
remove matching scnRNAs from the active pool. Those that cannot find the cytoplasm22 by the maternal MAC23,24. A pleiotropic mutation enfor-
a match remain free to target homologous sequences in the zygotic MAC cing constitutive determination for type E was later found to impair a limited
when it develops, thereby recruiting the IES excision machinery. This subset of genome rearrangements25, suggesting that type O is normally
1
Ecole Normale Supérieure, Institut de Biologie de l’ENS, IBENS; Inserm, U1024; CNRS, UMR 8197 Paris F-75005, France. 2Sorbonne Universités, UPMC Univ., IFD, 4 place Jussieu, 75252 Paris cedex 05,
France. 3CNRS UPR3404 Centre de Génétique Moléculaire, Gif-sur-Yvette F-91198, and Université Paris-Sud, Département de Biologie, Orsay F-91405, France. 4CNRS UMR5558, Laboratoire de Biométrie
et Biologie Evolutive, Université de Lyon, 43 boulevard du 11 Novembre 1918, Villeurbanne F-69622, France. 5Institute of Systematics and Evolution of Animals, Polish Academy of Sciences, Sĺawkowska
17, 31-016 Krakow, Poland. 6Department of Microbiology, Faculty of Biology, St Petersburg State University, Saint Petersburg 199034, Russia. 7Institut Jacques Monod, CNRS, UMR 7592, Université Paris
Diderot, Sorbonne Paris Cité, Paris F-75205, France. 8Commissariat à l’Energie Atomique (CEA), Institut de Génomique (IG), Genoscope, 2 rue Gaston Crémieux, BP5706, 91057 Evry, France. {Present
addresses: Laboratoire de Biochimie, Unité Mixte de Recherche 8231, École Supérieure de Physique et de Chimie Industrielles, 75231 Paris, France (B.S.); Department of Biology, Indiana University, Bloomington,
Indiana 47405, USA (J.-F.G.); INRA, UMR 1061 Unité de Génétique Moléculaire Animale, Université de Limoges, IFR 145, Faculté des Sciences et Techniques, 87060 Limoges, France (K.B.).
2 2 M AY 2 0 1 4 | VO L 5 0 9 | N AT U R E | 4 4 7
©2014 Macmillan Publishers Limited. All rights reserved
RESEARCH ARTICLE
a b a ATG TGA
Reactive O E MIC
mtA O E V R S V R S
mtAL1 mtA
mtAL2 MAC E
500 amino acids mtAL1
PCR1
Signal peptide Furin-like repeats
Transmembrane helices mtAL2 MAC O 1 kb
PCR1
Figure 1 | Structure and expression of mtA and related proteins.
a, Recognizable protein features. b, Northern blot analysis of expression of mtA, b c
mtAL1 (GSPATG00025159001) and mtAL2 (GSPATG00002922001). The O E O E M O O E E pO pE
3
panels on the left compare total RNA samples from sexually reactive cultures of
0.5
the two mating types (reactive, O and E). In the panels on the right, RNA kb 2 kb
samples were extracted from: V, vegetative, exponentially growing cells; R, 0.2
1.5
mildly starved and sexually reactive; S, starved for 48 h and no longer reactive.
The three mRNAs migrate just above the large rRNA, as can be seen from the
pattern of background hybridization. d e f
determined by the rearrangement of one of the genes required for E

expression26.
mtA rearrangements determine mating type

To identify the putative mating-type-determining gene, we used a whole-
genome microarray to compare the transcriptomes of sexually reactive
cells of both mating types (not shown). The gene found to have the largest
E/O expression ratio, GSPATG00017533001, encodes a 1,275-amino-acid
protein with a signal peptide and cysteine-rich furin-like repeats followed
by 5 transmembrane segments at the carboxy terminus (Fig. 1a and Supple-
mentary Data 1a); the amino-terminal part is predicted to be outside the Figure 2 | Structure of mtA in the MIC and in the MACs of E and O cells.
plasma membrane. This gene was later shown to be mtA (ref. 19) (see a, Black boxes are IESs; the 195-bp segment (grey box) is excised only in O
below). Northern blot analyses confirmed that it is expressed only in MACs. A small unannotated gene downstream of mtA is conserved in other
E cells and further showed that expression is limited to sexual reactivity, species. b, Southern blot of EcoR1-digested total DNA, hybridized with an mtA
as transcripts could not be detected in exponentially growing cells or in probe (PCR5) revealing fragments of 1,945 and 2,140 bp in O and E clones.
over-starved cells (Fig. 1b). Genes encoding structurally similar proteins c, PCR amplification of the mtA 59 end (PCR1) on independent pools of O and
were also found to be specifically expressed in sexually reactive cells, but E clones (266 and 461 bp, respectively). pO and pE, control PCRs on cloned
in both mating types (Fig. 1b). RNA interference (RNAi)-mediated silen- MAC versions; M, size markers. d–f, Localization of an mtA–GFP fusion
protein, with GFP inserted between furin-like repeats and transmembrane
cing of mtA in E cells resulted in the default O phenotype during sexual
segments. d, DIC image. Scale bar, 20 mm. e, Intracellular confocal optical
reactivity (as determined by their capacity to agglutinate with E but not section. Some GFP fluorescence (green) is detected in cilia of the anterior (top),
with O tester lines), indicating that mtA is required for E expression but not posterior, part of the cell. The MAC is stained with Hoechst (blue).
(Supplementary Table 1). f, Confocal image of the ventral cell surface, showing the opening of the oral
To understand the molecular basis for mating-type-E-specific express- apparatus. The fixed, non-permeabilized cell was labelled with anti-GFP and
ion, we sequenced the mtA gene from the MACs of O and E cells, and the secondary antibodies (red).
unrearranged MIC version. The MIC gene is interrupted by four IESs that
are excised in both MAC types; however, a 195-base-pair (bp) segment Expression and inheritance uncoupled
containing the transcription start site and the first 26 bp of the coding Previous mutational analyses showed that the expression of mating types
sequence was found to be excised as an IES in mating type O MACs, but during sexual reactivity can be uncoupled from the mechanism that en-
retained in E MACs (Fig. 2a–c and Supplementary Data 1a). This seg- sures their transgenerational inheritance19–21,23. The recessive mutations
ment contains the mtA promoter, as indicated by microinjection of diffe- mtAO, mtBO and mtCO all preclude expression of type E and restrict cells
rent constructs into the MACs of O cells (Extended Data Fig. 1a–c): its to the default O phenotype, but they do not affect the maternal inher-
presence upstream of the coding sequence was sufficient for transformed itance of mating-type determination: mutant homozygotes formed in an
clones to express mating type E instead of mating type O. An mtA–GFP E cytoplasmic lineage, despite expressing O, will keep the memory of E
fusion protein was detected in cilia of the anteroventral surface, although determination for an indefinite number of sexual generations, as shown
an excess of GFP fluorescence was seen in cytoplasmic structures, prob- by the fact that their progeny switches back to E expression when the
ably the ER (Fig. 2d, e). The fusion protein could be detected on cilia with wild-type allele is reintroduced by conjugation19 (Extended Data Fig. 2a–d).
anti-GFP antibodies in fixed, non-permeabilized cells (Fig. 2f), confirm- To determine whether this memory is attributable to retention of the
ing that its N-terminal part is exposed outside of the ciliary membrane. mtA promoter, we crossed each of the mutants to the wild type. Autogamy
When mtA-transformed clones were taken through autogamy (a self- of F1 heterozygotes of mating-type E yielded homozygous F2 progeny with
fertilization sexual process), the induced type E was robustly transmitted a 1:1 ratio of E- and O-expressing clones, reflecting the Mendelian segrega-
to progeny (Extended Data Fig. 1b). As observed for maternally controlled tion of wild-type and mutant alleles. Both types of clones retained the mtA
IESs14,15, a plasmid containing only the 195-bp segment was sufficient promoter (Extended Data Fig. 2e). That the O-expressing mutant clones
to inhibit excision of the homologous sequence during development of could still transmit E determination was verified by crossing them again to
zygotic MACs, causing the progeny of transformed O cells to switch to E wild-type E cells: the second-round F1 heterozygotes derived from both
(Extended Data Fig. 1d). Retention of the mtA promoter thus recapitulates parents always expressed E and retained the mtA promoter (Extended Data
both aspects of developmental determination for E, namely the capacity to Fig. 2f). Autogamy of these F1 heterozygotes again resulted in the expected
express mating-type E during the vegetative phase and the capacity, after Mendelian segregation of E- and O-expressing clones among second-
sexual events, to direct determination of new MACs for the same type. round F2 homozygotes (Supplementary Table 2). Thus, independently
4 4 8 | N AT U R E | VO L 5 0 9 | 2 2 M AY 2 0 1 4
ARTICLE RESEARCH
a in mtA transcription and E expression in both cases (Fig. 3b and Extended

mtAO x WT F3s mtBO x WT F3s mtCO x WT F3s
Data Fig. 3). Thus, both mtB and mtC seem to be transcription factors
pO pE E E O O O O E E O O O E E E O O O
required for mtA expression in E-determined cells. The mtB and mtC
genes are constitutively expressed at low levels, and both of the encoded
WT WT m m m m WT WT m m m WT WT WT m m m
proteins are predicted to be nuclear. mtC is a 138-amino-acid protein
containing a C2H2 zinc finger, the structure of which is probably affec-
mtA ted by the mtCO mutation (Supplementary Data 2). Although the 310-
amino-acid mtB protein does not contain any recognizable domain, an
mtAL1 mtB–GFP fusion protein was found to localize to the MAC (Fig. 3c). Deep
sequencing of the transcriptomes of mtBO and mtCO mutants further
showed that mtA is the only gene that requires both factors for express-
b mtBO clones mtCO clones ion during sexual reactivity; importantly, it is the only gene found to be
pmtB51-NGFP pmtC51 downregulated in the mtBO mutant (Supplementary Table 3).
C2 1 4 8 9 C1 2 4 5
mtA
mtA promoter excision is regulated by scnRNAs
Excision of the mtA promoter in mating-type O cells occurs between two
59-TA-39 dinucleotides within a reasonable IES end consensus (Supple-
c mentary Data 1a), suggesting that the general IES excision machinery is
GFP DAPI
involved. RNAi-mediated depletion of the Pgm endonuclease5 during
autogamy of O cells indeed impaired excision and resulted in accumula-
tion of unexcised copies in the developing new MAC (Fig. 4a). So did
depletion of proteins known to be involved in scnRNA biogenesis or
action, including the Dicer-like Dcl2 and Dcl3 (ref. 12), the Piwi-like
Ptiwi01 and Ptiwi09 (ref. 11), and the Nowa1 and Nowa2 RNA-binding
proteins27, indicating that the scnRNA pathway is required to target exci-
sion in O cells. Depletion of each pair of proteins causes massive reten-
tion of IESs genome-wide, resulting in non-functional new MACs and
Figure 3 | Molecular analysis of expression mutants. a, mtA expression in post-autogamous lethality. In conditions of partial depletion, obtained
E-determined mtAO, mtBO and mtCO homozygotes. The top panel shows a PCR
analysis (PCR1) of E- and O-expressing F3 lines from the second round of
by silencing only one of the two Piwi (or the two Dicer-like) genes, the
backcross of each mutant to wild-type E cells. pO and pE, control PCRs on cloned progeny is usually viable. Analysis of individual clones showed that they
MAC versions of mtA. All clones retained the mtA promoter. The wild-type frequently retained the mtA promoter in the new MAC, and in all cases
(WT) or mutant (m) genotypes were confirmed by sequencing. Middle and this correlated with a switch to mating type E (Fig. 4b).
bottom panels show northern blots of total RNA samples from sexually reactive No other IES is known to be affected in these conditions; in this regard,
cultures of each F3 line, hybridized with mtA or mtAL1 probes. b, Northern blot excision of the mtA promoter behaves like MAC deletions of cellular
analysis of mtA expression in sexually reactive cultures of E-determined mtBO genes, which were shown to be more sensitive to partial impairment of
and mtCO clones transformed with plasmids pmtB51-NGFP and pmtC51,
the scnRNA pathway11,12. Whereas other IESs are intervening sequences
respectively (see Extended Data Fig. 3). Transformed clones expressed mating
type E, whereas uninjected controls (C2 and C1) expressed mating type O.
c, Fluorescence microscopy of an mtBO cell transformed with pmtB51-NGFP, a A1
3
CE 9
expressing mating type E (GFP and DAPI filters). Scale bar, 20 mm.
+D
+P
W
M
7
NO
ND
PG
C1
D3
C1
C2
D2
D2
C2
P1
P1
P9
of the mating type being expressed, maternal inheritance of mtA promoter
retention may indeed underlie the epigenetic memory of E determination.
To understand how mutant clones can express mating type O despite
retaining the mtA promoter, we tested mtA expression in E- and O- b
DCL2 NOWA1
expressing lines, using F3 populations obtained by an additional autogamy
of second-round F2 homozygotes. In the case of the mtAO mutation,
northern blots showed that mtA (GSPATG00017533001) transcripts were
produced in sexually reactive cultures of O-expressing lines (Fig. 3a), pro- O O E E O E O E O O
mpting us to re-sequence the gene in mutant clones. A substitution was
found to change Arg codon 751 to a stop codon, and the mutation co- PTIWI01 PTIWI09
segregated with mating-type O expression among second-round F2 homo-
zygotes (Supplementary Table 2a). This revealed that GSPATG00017533001
is mtA and confirmed that the encoded protein is required for expression
of mating type E. The mutation lies at a distance from the mtA promoter O E E O E E E E E E E E E E E O O O E O E
and apparently does not affect the regulation of its excision, explain- Figure 4 | Genes required for excision of the mtA promoter in O cells.
ing how O-expressing mutant homozygotes can transmit either O or E a, PCR analysis (PCR5) of mtA promoter retention in mass progenies of O
determination. clones after RNAi-mediated silencing of the indicated genes. Total DNA
In contrast to mtAO lines, mtBO and mtCO O-expressing lines did not samples were prepared from starved post-autogamous cells when the new
produce mtA mRNA on sexual reactivity (Fig. 3a). The mutations were MACs were clearly visible. Because the parental MAC is still present at this
stage, the promoter-excised version is amplified in all cases; the promoter-
identified by whole-genome sequencing, and found to be substitutions
retaining fragment can be detected only if it accumulates in zygotic MACs.
in GSPATG00026812001 and GSPATG00009074001, respectively (Sup- ND7, unrelated-gene negative control; D2, DCL2; D3, DCL3; C1, no-silencing
plementary Data 1b, c). Correct identification of the mutations was negative control; C2, empty-vector RNAi control; P1, PTIWI01; P9, PTIWI09;
confirmed by their co-segregation with O expression among F2 lines (Sup- CE, control PCR5 on E cells. b, PCR analysis (PCR5) and mating-type tests
plementary Table 2c, d), and by transformation of the MACs of E- (below) of individual viable post-autogamous clones from non-lethal silencing
determined, O-expressing mutants with the wild-type alleles, which resulted conditions (including partial silencing of NOWA1).
2 2 M AY 2 0 1 4 | VO L 5 0 9 | N AT U R E | 4 4 9
RESEARCH ARTICLE
that must be excised to reconstitute functional genes, the 195-bp seg- a ATG TGA b
MIC C T
ment is a functional part of the mtA gene. The homologous segment is
not excised in mtA orthologues from other P. aurelia species (see below), MAC E
indicating that excision of a bona fide part of the MAC genome evolved mtA
as a derived character in P. tetraurelia, as random mutations happened MAC O
to create sites suitable for Pgm-mediated excision. O E
Deep sequencing of small RNAs from an early conjugation time point 500 bp
confirmed that scnRNAs are produced from both strands of the mtA pro-
moter, as they are from the rest of the germline genome (Extended Data c
1 2 3 4 5 6 7 8 9 10 11 12 13 14 C1 C2
Fig. 4). On average, scnRNA coverage was ,2-fold higher for IESs than mtA51
for other genome features, as independently found in another study28. pmtA227
Although we cannot exclude that this results from precocious degrada- 51 O – – – – – – – – – – – – – – – –
Testers
tion of MAC-matching scnRNAs, an intrinsically more abundant pro- 51 E – – – +++ – – – – – – – ++++ – – ++++ ++++
duction of scnRNAs from IESs could explain why IES excision is less 227 O ++ + + + +++ +++ ++ +++ +++ ++ ++ – +++ +++ – –
sensitive to partial impairment of the scnRNA pathway than gene dele- 227 E – – – – – – – – – – – – – – – –
tion. Northern blot analyses of autogamy time courses showed that mtA- Figure 5 | mtB rearrangements determine mating types in P. septaurelia.
promoter scnRNAs are produced in similar amounts during meiosis of a, MIC and MAC versions of mtB227. The MIC sequence contains two IESs
O and E cells, and failed to detect any difference in the timing of their dis- (black boxes), conserved in P. tetraurelia strain 51. In the MAC of O cells of
appearance (Extended Data Fig. 5). Thus, mtA-promoter scnRNAs may strain 227, the mtB227 coding sequence suffers two alternative deletions, each
be inactivated in E cells by some other mechanism than the active degra- between TA dinucleotides within IES-end consensus sequences (TAYAG, see
dation proposed in Tetrahymena thermophila29–31, for instance by seques- Supplementary Data 1b). b, Northern blot analysis of mtA expression in a
sexually reactive, pmtB51-transformed O clone of strain 227 (T). C, control
tration in the maternal MAC. Whatever the precise mechanism, one
uninjected O clone. Tested mating types are indicated. c, Transformation of
prediction of the model is that RNAi-mediated destruction of maternal P. tetraurelia clones of mating type O with the P. septaurelia mtA227 gene. A
transcripts of the mtA promoter during autogamy of E cells will prevent duplex PCR was used to assess transgene copy numbers through the relative
inactivation of homologous scnRNAs, licensing them to target excision abundance of a 176-bp transgene-specific product (pmtA227, PCR12),
in the zygotic MAC13. Indeed, feeding E cells with double-stranded RNA compared to the 266-bp product from the promoterless endogenous mtA
homologous to the mtA promoter before autogamy induced its precise (mtA51, PCR1; undetectable in high-copy transformants). C1 and C2,
excision in the new MAC and resulted in O progeny (Supplementary uninjected control clones. Each clone was tested with P. tetraurelia (51) and
Table 4). The induced type O was thereafter inherited for at least three P. septaurelia (227) O and E testers. Number of plus signs represents strength
sexual generations (not shown). of mating reaction; minus indicates no agglutination.
A different switch in P. septaurelia to different nucleotides. As with mtA promoter excision in P. tetraurelia,
P. tetraurelia belongs to a group of 15 sibling species that are morpho- this excision system apparently does not originate from any transposable
logically indistinguishable but sexually incompatible1,32. All have homo- element insertion but seems to have arisen in P. septaurelia after the
logous O and E mating types, as shown by cross-agglutination between chance appearance of good matches to the IES end consensus, which seems
the O types of some species and the E types of others. However, maternal to be conserved among P. aurelia species36.
inheritance of mating types is observed in only half of these species. In Taking advantage of the evidence that the mtB51 protein from
others, mating types are randomly determined during MAC develop- P. tetraurelia can activate the mtA227 promoter in P. septaurelia, we micro-
ment, without any influence of the maternal MAC. The distribution of injected the mtA227 gene into the MACs of P. tetraurelia cells of mating
these systems in the phylogenetic tree of aurelia species suggests multiple type O, and tested the sexual preference of transformed clones with O and
changes during evolution of the complex33,34 (Extended Data Fig. 6). We E testers of both species (Fig. 5c). Consistent with the reported lack
first sequenced mtA, mtB and mtC orthologues in two strains of P. octau- of cross-agglutination between the complementary types of these two
relia, a maternal-inheritance species closely related to P. tetraurelia, with species1, uninjected control clones reacted only with P. tetraurelia E testers.
which it can form viable (though sterile) F1 hybrids35. We found the mtA In contrast, transformed clones agglutinated only with P. septaurelia
promoter to be excised in O clones, with the same boundaries as in O testers, allowing interspecific conjugation to occur. Thus, mtA227 is
P. tetraurelia, but not in E clones (Extended Data Fig. 7 and Supple- sufficient for expression of the P. septaurelia E specificity, incidentally
mentary Data 1a). This excision system thus probably evolved in the confirming that mtA localizes on the external side of ciliary membranes.
common ancestor of the two species. The lack of reaction of transformed clones with P. tetraurelia O testers
We then examined P. septaurelia, another maternal-inheritance spe- indicates that mtA interacts with an unknown O-specific receptor which
cies that groups with random-determination species. Sequencing of the also differs between species. Furthermore, transformed clones did not
mtA, mtB and mtC orthologues in the wild-type strain 227 showed that react with P. tetraurelia E testers, indicating that mtA227 expression, like
the mtA227 promoter was not excised in the MACs of O cells (Supple- mtA51 expression (Extended Data Fig. 1), functionally masks expression
mentary Data 1a). Instead, MAC copies of the mtB227 gene contained of the endogenous O-specific receptor. This would explain the default
either of two alternative deletions of coding-sequence segments between nature of the O type.
IES-like boundaries (Fig. 5a and Supplementary Data 1b), suggesting that Strain 38 is a natural P. septaurelia isolate carrying a Mendelian muta-
the gene was rearranged into non-functional forms. Our stock of strain tion at a single locus (mt) which, in contrast to the P. tetraurelia muta-
227 contained only O cells, but complementation of their MACs by micro- tions, affects both the expression and the epigenetic inheritance of mating
injection of the P. tetraurelia pmtB51 transgene made these cells phenoty- types20. This dual effect suggests that mt is the gene that controls mating
pically E and resulted in mtA transcription during sexual reactivity (Fig. 5b). types through an alternative rearrangement in that species (see Extended
Thus, mating type E is characterized by mtB-dependent mtA expression Data Fig. 8), providing an independent means to confirm the identifica-
in P. septaurelia, as in P. tetraurelia. tion of the P. septaurelia switch. Sequencing of the mtB38 allele revealed
The mtB gene belongs to a family of up to six paralogues in P. aurelia mutations that can explain both effects (Extended Data Fig. 8c and Supple-
species, and sequence conservation indicates that all have been under puri- mentary Data 1b), and the analysis of a cross between strains 38 and 227
fying selection since before speciation (not shown). The excision events showed that mt is identical or closely linked to mtB (Supplementary Table 5).
that inactivate mtB227 in O cells are not observed in any other paralogue To confirm directly that mtB rearrangements are regulated by homo-
or in mtB orthologues from other species, where the 59 TAs correspond logy-dependent maternal effects, the full-length, functional MAC form
4 5 0 | N AT U R E | VO L 5 0 9 | 2 2 M AY 2 0 1 4
ARTICLE RESEARCH
of the mtB227 allele was microinjected into the MACs of O cells of strain transmission to sexual progeny would thus allow continuous adaptation
227. Transformed clones expressed E and produced E post-autogamous of the somatic (MAC) genome, independently of any Mendelian (MIC)
progeny (Extended Data Fig. 9). We conclude that mating types are determ- mutation.
ined in P. septaurelia by maternally inherited alternative rearrange-
ments of the mtB gene, a mechanism that must have evolved independently METHODS SUMMARY
from that of P. tetraurelia. Unless otherwise stated, all experiments were carried out with the entirely homo-
zygous strain 51 of P. tetraurelia; the origins of other strains and species are given in
Discussion the Methods section. Molecular biology experiments used standard procedures. Para-
Seventy-five years after mating types were first described in P. tetraurelia, mecium-specific methods (RNAi-mediated gene silencing by dsRNA feeding, DNA
microinjection into the MAC) have been published and are referenced in the Methods
the expression of a single gene, mtA, was found to make the difference
section. For mating-type tests, testers were prepared from cell lines of known mating
between the two types. mtA transcription in E cells was further shown to types by re-feeding ,1,000 autogamous cells in tubes for 2–3 divisions. The proce-
require the mtB and mtC gene products, and this E-specific pathway seems dures used to test mass post-autogamous progenies or individual clones are presented
to be conserved in other P. aurelia species. Our results indicate that mtA is in the Methods section. Bioinformatic analyses and statistical testing (microarray ex-
a ciliary transmembrane protein directly involved in the species-specific pression data, whole-genome transcriptome profiling by RNA-seq) are described in
recognition of O cells. Although the nature of the O-specific receptor the Methods.
remains unknown, its expression or function appears to be masked by
Online Content Any additional Methods, Extended Data display items and Source
the mtA protein, given the default O phenotype observed in mutants. Data are available in the online version of the paper; references unique to these
Expressing the P. septaurelia mtA orthologue in O cells of P. tetraurelia sections appear only in the online paper.
not only provided them with the P. septaurelia E specificity, but also
blocked expression of the P. tetraurelia O specificity. In the related Received 17 August 2013; accepted 11 April 2014.
P. caudatum, an antigen involved in agglutination was shown to be pro- Published online 7 May 2014.
duced in the cytoplasm of both O and E reactive cells, but to localize in the
1. Sonneborn, T. M. Paramecium aurelia. in Handbook of Genetics (ed. King, R. C.)
ciliary membrane only in O cells37. In E cells of P. aurelia species, the mtA 469–594 (Plenum, 1974).
protein could inhibit the synthesis, processing or transport of the O- 2. Chalker, D. L. & Yao, M. C. DNA elimination in ciliates: transposon domestication
specific receptor. mtA is structurally similar to the proteins recently shown and genome surveillance. Annu. Rev. Genet. 45, 227–246 (2011).
3. Arnaiz, O. et al. The Paramecium germline genome provides a niche for intragenic
to be specific for each of the seven mating types of T. thermophila38, but parasitic DNA: evolutionary dynamics of internal eliminated sequences. PLoS
how these function in mating-type recognition remains to be determined. Genet. 8, e1002984 (2012).
Despite conservation of the mtA on/off switch as the E/O expression 4. Bétermier, M. Large-scale genome remodelling by the developmentally
programmed elimination of germ line sequences in the ciliate Paramecium. Res.
determinant, the mechanisms used to produce O clones differ among Microbiol. 155, 399–408 (2004).
P. aurelia species. Excision of the mtA promoter is shared between the 5. Baudry, C. et al. PiggyMac, a domesticated piggyBac transposase involved in
closely related P. tetraurelia and P. octaurelia, but in P. septaurelia mtA programmed genome rearrangements in the ciliate Paramecium tetraurelia. Genes
expression is prevented by deletions in mtB (Extended Data Fig. 10). This Dev. 23, 2478–2483 (2009).
6. Klobutcher, L. A. & Herrick, G. Consensus inverted terminal repeat sequence of
seems to be functionally equivalent, as mtB is highly specific for the mtA Paramecium IESs: resemblance to termini of Tc1-related and Euplotes Tec
promoter in P. tetraurelia. The P. tetraurelia rearrangement involves the transposons. Nucleic Acids Res. 23, 2006–2013 (1995).
IES excision machinery and is regulated by the scnRNA pathway, ex- 7. Dubois, E. et al. Transposon invasion of the Paramecium germline genome
countered by a domesticated PiggyBac transposase and the NHEJ pathway. Int. J.
plaining the maternal inheritance of mating types in this species, and Evol. Biol. 2012, 436196 (2012).
probably in the other two as well. Ciliate scnRNAs resemble metazoan 8. Chalker, D. L., Meyer, E. & Mochizuki, K. Epigenetics of ciliates. Cold Spring Harb.
piRNAs in many respects, including their central roles in epigenetic Perspect. Biol. 5, a017764 (2013).
reprogramming events that are critical for sexual reproduction39, and 9. Coyne, R. S., Lhuillier-Akakpo, M. & Duharcourt, S. RNA-guided DNA
rearrangements in ciliates: is the best genome defence a good offence? Biol. Cell
both have been described as genomic ‘immune systems’ primarily 104, 309–325 (2012).
involved in the control of transposable elements2,40–42. One difference, 10. Duharcourt, S., Lepere, G. & Meyer, E. Developmental genome rearrangements in
however, is that metazoan piRNA systems keep a memory of ‘non-self’ se- ciliates: a natural genomic subtraction mediated by non-coding transcripts.
Trends Genet. 25, 344–350 (2009).
quences, based on integration into specific small-RNA-producing loci40,42,43. 11. Bouhouche, K., Gout, J. F., Kapusta, A., Betermier, M. & Meyer, E. Functional
The Paramecium system seems to work the opposite way, keeping a specialization of Piwi proteins in Paramecium tetraurelia from post-transcriptional
memory of ‘self’ sequences in the form of the maternal MAC, which is gene silencing to genome remodelling. Nucleic Acids Res. 39, 4249–4264 (2011).
probed by genome-wide scnRNAs at each generation to identify ‘non- 12. Lepere, G. et al. Silencing-associated and meiosis-specific small RNA pathways in
Paramecium tetraurelia. Nucleic Acids Res. 37, 903–915 (2009).
self’ molecules by subtraction. 13. Lepere, G., Betermier, M., Meyer, E. & Duharcourt, S. Maternal noncoding
As a consequence, the Paramecium genome-defence mechanism can transcripts antagonize the targeting of DNA elimination by scanRNAs in
easily be recruited for heritable regulation of bona fide cellular genes, an Paramecium tetraurelia. Genes Dev. 22, 1501–1512 (2008).
14. Duharcourt, S., Butler, A. & Meyer, E. Epigenetic self-regulation of developmental
important question that has been raised in diverse eukaryotes42,44–46. In- excision of an internal eliminated sequence on Paramecium tetraurelia. Genes Dev.
deed, the sequences being excised from the mtA and mtB genes differ 9, 2065–2077 (1995).
from other IESs3,7 in that they are functional parts of cellular genes and 15. Duharcourt, S., Keller, A. M. & Meyer, E. Homology-dependent maternal inhibition
apparently do not derive from the insertion of transposable elements; of developmental excision of internal eliminated sequences in Paramecium
tetraurelia. Mol. Cell. Biol. 18, 7075–7085 (1998).
moreover, the mtA rearrangement behaves like deletions of cellular genes 16. Garnier, O., Serrano, V., Duharcourt, S. & Meyer, E. RNA-mediated programming of
in its mechanistic requirements, which show subtle differences with IES developmental genome rearrangements in Paramecium tetraurelia. Mol. Cell. Biol.
excision11,12. The independent evolution of the P. tetraurelia and P. sep- 24, 7370–7379 (2004).
17. Sonneborn, T. M. Sex, sex inheritance and sex determination in Paramecium
taurelia switches suggests that exaptation of the scnRNA pathway is a aurelia. Proc. Natl Acad. Sci. USA 23, 378–385 (1937).
general mechanism for transgenerational epigenetic inheritance of dif- 18. Butzel, H. M. Mating type mutations in variety 1 of Paramecium aurelia, and their
ferentiated states. In plants and mammals, epigenetic regulation of cel- bearing upon the problem of mating type determination. Genetics 40, 321–330
lular genes often relies on transposon-induced epigenetic changes44,45, as (1955).
19. Byrne, B. C. Mutational analysis of mating type inheritance in Syngen 4 of
is the case of sex determination in melon47. The capacity of the Parame- Paramecium aurelia. Genetics 74, 63–80 (1973).
cium mechanism to inactivate any single-copy gene makes it even more 20. Taub, S. R. The genetic control of mating type differentiation in Paramecium.
flexible, dispensing with the need for transposon insertion. In keeping Genetics 48, 815–834 (1963).
21. Sonneborn, T. M. Genetics of cellular differentiation: stable nuclear differentiation
with the low information content of the IES end consensus, acciden- in eucaryotic unicells. Annu. Rev. Genet. 11, 349–367 (1977).
tal rearrangement errors are not rare genome-wide3,36,48, and may occa- 22. Sonneborn, T. M. Recent advances in the genetics of Paramecium and Euplotes.
sionally provide a selective advantage in specific conditions; maternal Adv. Genet. 1, 263–358 (1947).
2 2 M AY 2 0 1 4 | VO L 5 0 9 | N AT U R E | 4 5 1
RESEARCH ARTICLE
23. Nanney, D. L. Mating type inheritance at conjugation in variety 4 of Paramecium 47. Martin, A. et al. A transposon-induced epigenetic change leads to sex
aurelia. J. Protozool. 4, 89–95 (1957). determination in melon. Nature 461, 1135–1138 (2009).
24. Sonneborn, T. M. Patterns of nucleo-cytoplasmic integration in Paramecium. 48. Duret, L. et al. Analysis of sequence variability in the macronuclear DNA of
Caryologia 6 (suppl.), 307–325 (1954). Paramecium tetraurelia: a somatic view of the germline. Genome Res. 18, 585–596
25. Meyer, E. & Keller, A. M. A Mendelian mutation affecting mating-type (2008).
determination also affects developmental genomic rearrangements in 49. Edgar, R., Domrachev, M. & Lash, A. E. Gene expression omnibus: NCBI gene
Paramecium tetraurelia. Genetics 143, 191–202 (1996). expression and hybridization array data repository. Nucleic Acids Res. 30,
26. Meyer, E. & Garnier, O. Non-mendelian inheritance and homology-dependent 207–210 (2002).
effects in ciliates. Adv. Genet. 46, 305–337 (2002).
27. Nowacki, M., Zagorski-Ostoja, W. & Meyer, E. Nowa1p and Nowa2p: novel putative Supplementary Information is available in the online version of the paper.
RNA binding proteins involved in trans-nuclear crosstalk in Paramecium
Acknowledgements We thank S. Malinsky, C. Ciaudo and M.-A. Félix for critical reading
tetraurelia. Curr. Biol. 15, 1616–1628 (2005).
of the manuscript, and S. Marker and all other laboratory members for continuous
28. Sandoval, P. Y., Swart, E. C., Arambasic, M. & Nowacki, M. Functional diversification
support and discussions. This work was supported by the ‘Investissements d’Avenir’
of Dicer-like proteins and small RNAs required for genome sculpting. Dev. Cell 28,
program ANR-10-LABX-54 MEMO LIFE/ANR-11-IDEX-0001-02 Paris Sciences et
174–188 (2014).
Lettres* Research University and by grants ANR-08-BLAN-0233 ‘ParaDice’ and
29. Aronica, L. et al. Study of an RNA helicase implicates small RNA-noncoding RNA
ANR-12-BSV6-0017 ‘INFERNO’ to E.M., L.S. and S.D., an ‘Equipe FRM’ grant to E.M.,
interactions in programmed DNA elimination in Tetrahymena. Genes Dev. 22,
grants ANR-2010-BLAN-1603 ‘GENOMAC’ and CNRS ATIP-Avenir to S.D., and
2228–2241 (2008).
National Science Foundation grant MCB-1050161 to M. Lynch (JFG). D.P.S. was
30. Mochizuki, K. & Gorovsky, M. A. Conjugation-specific small RNAs in Tetrahymena
supported by Ph.D. fellowships from the Erasmus Mundus program and from the Ligue
have predicted properties of scan (scn) RNAs involved in genome rearrangement.
Nationale Contre le Cancer. M.L.-A. was supported by Ph.D. fellowships from the
Genes Dev. 18, 2068–2073 (2004).
Ministère de l’Enseignement Supérieur et de la Recherche and from the Fondation de
31. Schoeberl, U. E., Kurth, H. M., Noto, T. & Mochizuki, K. Biased transcription and
la Recherche Médicale. A.P. was supported by grant RFBR 13-04-01683a. Some
selective degradation of small RNAs shape the pattern of DNA elimination in
strains used in this study are maintained at the Centre of Core Facilities ‘Culture
Tetrahymena. Genes Dev. 26, 1729–1742 (2012).
Collection of Microorganisms’ in St Petersburg State University. The sequencing of the
32. Catania, F., Wurmser, F., Potekhin, A. A., Przybos, E. & Lynch, M. Genetic diversity in
mtBO and mtCO MAC genomes benefited from the facilities and expertise of the
the Paramecium aurelia species complex. Mol. Biol. Evol. 26, 421–431 (2009).
high-throughput sequencing platform of IMAGIF (Centre de Recherche de Gif, http://
33. Hall, M. S. & Katz, L. A. On the nature of species: insights from Paramecium and
www.imagif.cnrs.fr). The mtBO and mtCO transcriptomes were sequenced at the
other ciliates. Genetica 139, 677–684 (2011).
Genomic Paris Centre - IBENS platform, member of ‘France Gènomique’
34. Phadke, S. S. & Zufall, R. A. Rapid diversification of mating systems in ciliates. Biol. J.
(ANR10-INBS-09-08). This study was carried out in the context of the CNRS-supported
Linn. Soc. 98, 187–197 (2009).
European Research Group ‘Paramecium Genome Dynamics and Evolution’ and the
35. Haggard, B. W. Interspecies crosses in Paramecium aurelia (syngen 4 by syngen 8).
European COST Action BM1102.
J. Protozool. 21, 152–159 (1974).
36. Catania, F., McGrath, C. L., Doak, T. G. & Lynch, M. Spliced DNA sequences in the Author Contributions D.P.S. did almost all of the experimental work presented here
Paramecium germline: their properties and evolutionary potential. Genome Biol. and contributed to the design of experiments. B.S. characterized mRNAs and
Evol. 5, 1200–1211 (2013). contributed to silencing experiments and northern blot analyses. G.G. contributed to
37. Xu, X., Kumakura, M., Kaku, E. & Takahashi, M. Odd mating-type substances may gene sequencing, plasmid construction, PCR analyses and cell line maintenance. J.-F.G.
work as precursor molecules of even mating-type substances in Paramecium did the microarray analysis, and A.A.-F. the confocal analysis of mtA–GFP fusions. A.A.,
caudatum. J. Eukaryot. Microbiol. 48, 683–689 (2001). K.L. and J.-M.A. carried out the deep sequencing of small RNAs, and C.B. that of the
38. Cervantes, M. D. et al. Selecting one of several mating types through gene segment mtBO and mtCO transcriptomes; O.A. and L.S. did the bioinformatic analyses. K.B.,
joining and deletion in Tetrahymena thermophila. PLoS Biol. 11, e1001518 (2013). M.L.-A., V.T. and S.D. showed the role of scnRNA pathway genes in mtA promoter
39. Bourc’his, D. & Voinnet, O. A small-RNA perspective on gametogenesis, excision. S.B. did the mtA promoter dsRNA feeding experiment. A.P. contributed to the
fertilization, and early zygotic development. Science 330, 617–622 (2010). analysis of the mtAO mutant and provided P. octaurelia and septaurelia strains. M.P.
40. Malone, C. D. & Hannon, G. J. Small RNAs as guardians of the genome. Cell 136, contributed to the analysis of the mtBO mutant and prepared samples from the
656–668 (2009). P. octaurelia cross, which was carried out by E.P. E.M. conceived the study and wrote
41. Schoeberl, U. E. & Mochizuki, K. Keeping the soma free of transposons: the paper.
programmed DNA elimination in ciliates. J. Biol. Chem. 286, 37045–37052 (2011).
42. Siomi, M. C., Sato, K., Pezic, D. & Aravin, A. A. PIWI-interacting small RNAs: the Author Information Microarray data have been deposited at the Gene Expression
vanguard of genome defence. Nature Rev. Mol. Cell Biol. 12, 246–258 (2011). Omnibus database49 under accession number GSE43436. RNA-seq data
43. Khurana, J. S. et al. Adaptation to P element transposon invasion in Drosophila (transcriptomes of mtBO and mtCO mutants) have been deposited in the European
melanogaster. Cell 147, 1551–1563 (2011). Nucleotide Archive (EBI) under accession number ERP002291. Small RNA sequences
44. Castel, S. E. & Martienssen, R. A. RNA interference in the nucleus: roles for small have been deposited at the EBI under accession number ERP001812. The mtA, mtB
RNAs in transcription, epigenetics and beyond. Nature Rev. Genet. 14, 100–112 and mtC sequences of all strains and species studied have been deposited at GenBank
(2013). under accession codes KJ748544–KJ748569. Reprints and permissions information
45. Daxinger, L. & Whitelaw, E. Understanding transgenerational epigenetic is available at www.nature.com/reprints. The authors declare no competing financial
inheritance via the gametes in mammals. Nature Rev. Genet. 13, 153–162 (2012). interests. Readers are welcome to comment on the online version of the paper.
46. Luteijn, M. J. & Ketting, R. F. PIWI-interacting RNAs: from generation to Correspondence and requests for materials should be addressed to
transgenerational epigenetics. Nature Rev. Genet. 14, 523–534 (2013). E.M. (emeyer@biologie.ens.fr).
4 5 2 | N AT U R E | VO L 5 0 9 | 2 2 M AY 2 0 1 4
ARTICLE RESEARCH
METHODS Confocal analysis of mtA–GFP localization. Cells were fixed in 2% paraformalde-

Paramecium strains and cultivation. Unless otherwise stated, all experiments hyde in PBS buffer and rinsed in the same buffer supplemented with 3% BSA. All sub-
were carried out with the entirely homozygous strain 51 of P. tetraurelia. The mutant sequent steps were performed in this buffer: cells were incubated with a commercial
strains mtAO, mtBO and mtCO, as well as strain 32 of P. tetraurelia, were from the stock anti-GFP rabbit serum (Molecular Probes A6455, Invitrogen) at a 1/1,000 dilution,
collection of the Centre de Génétique Moléculaire in Gif-sur-Yvette, France. The mtAO rinsed twice, incubated with an Alexa 568-coupled anti-rabbit antibody (Molecular
mutant was independently ordered from ATCC (ATCC number 30762). Strains 138 Probes goat anti-rabbit IgG(H1L) A11011, Invitrogen), rinsed in 1 mg ml21 Hoechst
and GFg-1 of P. octaurelia were from the stock collection of the Institute of Systematics 33258, mounted in Citifluor (Citifluor Ltd London), and observed under a Leica Con-
and Evolution of Animals, Krakow, Poland. Strain 227 from P. septaurelia was from focal SP8 microscope with 405, 488 and 552 laser line excitations for blue, GFP and
the stock collection of the Centre of Core Facilities ‘Cultivation of Microorganisms’ in DIC, and red detection, respectively. Images were merged using ImageJ and Adobe
St Petersburg State University, Russia. Strain 38 of P. septaurelia was ordered from Photoshop.
ATCC (ATCC number 30575). Cultivation and autogamy were carried out at 27 uC Deep sequencing of scnRNAs. Small-RNA libraries were previously constructed
(unless otherwise stated) as described50,51. using 24-nt adaptors12. The corresponding PCR products were extended with 6
DNA and RNA extraction, Southern and northern blots. DNA and RNA samples degenerate nucleotides at each end and ligated to Illumina adaptors essentially
were typically extracted from 200- to 400-ml cultures of exponentially growing cells following Illumina’s recommendations (TruSeq DNA sample prep kit protocol)
at ,1,000 cells ml21 or of autogamous cells at 2,000–4,000 cells ml21 as previously before sequencing on the GAIIx system (100-nt single reads) (Illumina, USA).
described5. Small-scale DNA samples were prepared from #1,000 cells using the Whole-genome transcriptome profiling. Eight Illumina sequencing libraries were
NucleoSpin Tissue kit (Macherey-Nagel). The TRIzol (Invitrogen) RNA extraction constructed (two biological replicates for mtBO, two for mtCO, and two wild-type con-
procedure was adapted for small-scale cultures (,20 ml) of individual transformed trols for each) using the Illumina TruSeq RNA Sample Preparation kit. Approxi-
clones during sexual reactivity. Electrophoresis and blotting were carried out accord- mately 30 million 101-nt read pairs were obtained for each library with the Illumina
ing to standard procedures. For small-RNA northern blots, small RNAs were en- HiSeq 1000 sequencing system. Reads were processed to remove adapters and trimmed
riched by PEG8000 precipitation, and the equivalent of 75 mg of total RNA for each for quality before mapping to the strain 51 genome3,62 using BWA software63. The frag-
time point was run on a 15% polyacrylamide-urea gel, transferred on Hybond NX ments that mapped to each annotated gene62 were counted only if both reads in the
membranes, and chemically crosslinked52. pair mapped to the same gene and only if the mapping for both reads was unique in
Microarray expression data. Expression data were obtained from single-channel the genome. Fragments were only counted if neither read in the pair had more than
NimbleGen microarrays covering all 39,642 annotated genes, with six different 50-mer two mismatches. Fragment counts per gene were obtained from the BWA output
probes per gene. Raw signals were processed using the standard RMA method53. This using Samtools64 and a custom Perl script. Surface antigen genes were excluded from
includes a first step of background subtraction for each array, followed by between- the analysis as the cell cultures were not controlled for serotype and surface antigen
array normalization which was carried out using the normalizeBetweenArrays func- mRNA can account for 3% of total cellular mRNA. Differential gene expression was
tion from the limma package54. The latter step adjusts signals so that expression values determined from fragment counts using the DESeq package65. P values were cor-
have similar distributions in the two arrays considered in the analysis. The expression rected for multiple testing using the Benjamini–Hochberg procedure.
level of each gene was taken as the median signal from the six probes. The microarray
platform has been described in more detail elsewhere55. 50. Beisson, J. et al. Mass culture of Paramecium tetraurelia. Cold Spring Harb. Protoc.
Alignment and prediction programs. DNA and protein sequences were aligned 2010, http://dx.doi.org/10.1101/pdb.prot5362 (2010).
using the MUSCLE software on the phylogeny.fr website56. Prediction of protein locali- 51. Beisson, J. et al. Maintaining clonal Paramecium tetraurelia cell lines of controlled
age through daily reisolation. Cold Spring Harb. Protoc. 2010, http://dx.doi.org/
zation and transmembrane protein topology used the PSORT II57 and PolyPhobius58 10.1101/pdb.prot5361 (2010).
servers. 52. Pall, G. S. & Hamilton, A. J. Improved northern blot method for enhanced detection
Mating-type tests. Testers were prepared from cell lines of known mating types by of small RNA. Nature Protocols 3, 1077–1084 (2008).
re-feeding ,1,000 autogamous cells in tubes with 4 ml of 0.2 X WGP medium bacter- 53. Irizarry, R. A. et al. Exploration, normalization, and summaries of high density
ized with Klebsiella pneumoniae (light medium) and incubating overnight at 27 uC. oligonucleotide array probe level data. Biostatistics 4, 249–264 (2003).
The next day, tubes were re-fed with 8 ml of light medium and again incubated over- 54. Smyth, G. K. & Speed, T. Normalization of cDNA microarray data. Methods 31,
265–273 (2003).
night at 27 uC. The following day, reactive cells concentrated near the top of the tube
55. Arnaiz, O. et al. Gene expression in a paleopolyploid: a transcriptome resource for
were collected (,1.5 ml per tube), checked by mixing aliquots with reactive cells of the ciliate Paramecium tetraurelia. BMC Genom. 11, 547 (2010).
the complementary type, and used in mating type tests. Mass post-autogamous pro- 56. Dereeper, A. et al. Phylogeny.fr: robust phylogenetic analysis for the non-specialist.
genies to be tested were made reactive in the same way. To test individual clones, Nucleic Acids Res. 36, W465–W469 (2008).
single karyonides (out of autogamy or conjugation) were isolated in 250 ml of light 57. Nakai, K. & Horton, P. PSORT: a program for detecting sorting signals in proteins
medium and incubated until starved. They were then re-fed with 250 ml and tested and predicting their subcellular localization. Trends Biochem. Sci. 24, 34–35
the next day. (1999).
58. Kall, L., Krogh, A. & Sonnhammer, E. L. An HMM posterior decoder for sequence
RNAi-mediated gene silencing by dsRNA feeding and microinjections. Procedures feature prediction that includes homology information. Bioinformatics 21
for RNAi and DNA microinjection into the MAC, as well as inserts used to silence (suppl. 1), 251–257 (2005).
scnRNA-pathway genes and PGM and the procedure to silence genes during auto- 59. Beisson, J. et al. DNA microinjection into the macronucleus of Paramecium. Cold
gamy, have been described5,11,12,27,59,60. To silence genes during sexual reactivity, cells Spring Harb. Protoc. 2010, http://dx.doi.org/10.1101/pdb.prot5364 (2010).
were made reactive as described under ‘Mating-type tests’, except that the second re- 60. Beisson, J. et al. Silencing specific Paramecium tetraurelia genes by feeding double-
feeding with 8 ml was done with dsRNA feeding medium instead of normal WGP stranded RNA. Cold Spring Harb. Protoc. 2010, http://dx.doi.org/10.1101/
pdb.prot5363 (2010).
medium.
61. Marker, S., Le Mouel, A., Meyer, E. & Simon, M. Distinct RNA-dependent RNA
Plasmid constructs and probes. The plasmids used in this study are described in polymerases are required for RNAi triggered by double-stranded RNA versus
the relevant figures or Extended Data. The mtA-GFP fusion transgene used in Fig. 2d–f truncated transgenes in Paramecium tetraurelia. Nucleic Acids Res. 38, 4092–4107
is a modified version of pmtA-ES (Extended Data Fig. 1) in which the EGFP coding (2010).
sequence, flanked by Gly-Ser-Gly-Gly and Gly-Gly linkers, was inserted in place of 62. Aury, J. M. et al. Global trends of whole-genome duplications revealed by the ciliate
mtA residue Asn 1002. The pmtA227 transgene used in Fig. 5c is the P. septaurelia Paramecium tetraurelia. Nature 444, 171–178 (2006).
equivalent of pmtA-ES, with complete flanking intergenic regions (without the down- 63. Li, H. & Durbin, R. Fast and accurate short read alignment with Burrows-Wheeler
transform. Bioinformatics 25, 1754–1760 (2009).
stream unannotated gene). Complete sequences of all plasmids are available on request.
64. Li, H. et al. The sequence alignment/map format and SAMtools. Bioinformatics 25,
The mtA probe used to hybridize northern blots covered the whole coding sequence, 2078–2079 (2009).
except on the left panel of Fig. 1c where it spanned positions 415–1010 of the corrected 65. Anders, S. & Huber, W. Differential expression analysis for sequence count data.
gene model (PTETG5300016001, see ParameciumDB). The P. septaurelia probe in Genome Biol. 11, R106 (2010).
Fig. 5b and Extended Data Fig. 8d covered region 78–2245 of the mtA38 gene model. 66. Brygoo, Y., Sonneborn, T. M., Keller, A. M., Dippell, R. V. & Schneller, M. V. Genetic
For other genes, probes covered the indicated regions of each gene model: mtAL1, analysis of mating type differentiation in Paramecium tetraurelia. II. Role of the
3053–3599 of GSPATG00025159001; mtAL2, 3328–3790 of GSPATG00002922001; micronuclei in mating-type determination. Genetics 94, 951–959 (1980).
67. Brygoo, Y. Genetic analysis of mating-type differentiation in Paramecium
PTIWI09, 50–439 of GSPATG00020796001; NOWA2, 2513–3330 of GSPATG00016 tetraurelia. Genetics 87, 633–653 (1977).
668001; PGM, 1540–2390 of GSPATG00016627001; PTIWI10, 4–264 of GSPATG 68. Kapusta, A. et al. Highly precise and developmentally programmed genome
00009468001. Oligonucleotide probes for scaffold-22 endogenous siRNAs (Extended assembly in Paramecium requires ligase IV-dependent end joining. PLoS Genet. 7,
Data Fig. 5c) have been described61. e1002049 (2011).

RESEARCH ARTICLE
Extended Data Figure 1 | The 195-bp segment contains the mtA promoter
and ensures maternal inheritance of its own retention. a, Plasmids used for
microinjection into the MAC of O cells. The thick grey lines on either side of
inserts represent vector sequences. Plasmid pmtA-E contains the E MAC
version with entire intergenic regions; pmtA-ES lacks the downstream
unannotated gene. pmtA-IES lacks intergenic sequences upstream of the
195-bp segment. pIES contains only the 195-bp segment. The PCRs used for
testing transformation are shown. b, Transformation of O cells with pmtA-E or
pmtA-ES. PCR1 amplifies the promoter-containing form from either plasmid
and the promoter-excised form from the endogenous mtA gene; relative
amounts provide an indication of plasmid copy numbers. The mating types of
injected clones and of their mass post-autogamous progeny are indicated. C1
and C2, uninjected controls; pE, control PCR on pmtA-E. c, Transformation of
O cells with pmtA-IES. The plasmid-specific PCR2 (1,533 bp) identifies
transformed clones. The O type of clones 9 and 10 could be due to transgene
silencing. d, Transformation of O cells with pIES. Injected clones were tested
with a duplex PCR: PCR4 amplifies a 450-bp product from the plasmid and
PCR3 a 1,035-bp product from the endogenous mtA gene, hardly detectable in
high-copy transformants. C1 and C2, uninjected controls; pD, control duplex
PCR on an equimolar mix of pIES and pmtA-E. PCR1 revealed that the 195-bp
segment was retained in the mass progeny of transformed clones; the selfer
phenotype (S) of the mass progeny from clone 4 probably reflects heterogeneity
among individual post-autogamous clones.

ARTICLE RESEARCH
Extended Data Figure 2 | Genetic analysis of mutations affecting mating- cells. E determination of the mutant parent is revealed by E determination and
type expression. a, Mendelian segregation of a pair of alleles. In both expression of the derived F1, which has received the wild-type allele. In this
conjugation and autogamy, the two MICs undergo meiosis, but only one cross all cells are determined for E, and the expressed mating type simply
haploid product is retained in each cell; an additional mitosis then produces two depends on genotype. e, mtA promoter retention in the cross shown in
identical gametic nuclei. During conjugation of genetically different cells (m/m c, illustrated here for a cross of the mtAO mutant with wild-type E cells. Top
and 1/1), reciprocal exchange of one gametic nucleus, followed by karyogamy, panel: PCR analysis (PCR5) of two sets of 4 mtAO/mtA1 F1 karyonides
therefore results in genetically identical zygotic nuclei in the two exconjugants. (after fertilization, two new MACs develop in each cell from copies of the
The drawing shows the heterozygous MICs and MACs that develop from zygotic nucleus, and then segregate to daughter cells, called karyonides, at the
copies of the zygotic nuclei in each of the F1 cells. During autogamy, the two first cellular division; each pair of conjugants thus gives rise to 4 F1 karyonidal
identical gametic nuclei fuse together, resulting in entirely homozygous clones). Each set contained 2 O clones that had excised the mtA promoter, and
zygotes; post-autogamous F2 progeny of heterozygotes have a 50% probability probably derived from the mutant O parent, and two E clones that retained it,
to keep each of the parental alleles. b, Maternal (cytoplasmic) inheritance of and probably derived from the wild-type E parent (mating types are indicated
mating types. Coloured Os and Es around the cells indicate the mating type below each lane). Bottom panels: after autogamy of two mtAO/mtA1 F1
expressed by each vegetative clone; the MACs are coloured to symbolize their heterozygotes of mating type E (clone 4 in each set), 12 F2 homozygous progeny
mating-type determination states (O, blue; E, orange). Because little cytoplasm were isolated for each, grown and tested for mating-type expression and for
is exchanged during conjugation, the parental MACs independently condition mtA promoter retention using PCR6 (Supplementary Table 6), which amplifies
zygotic MACs for the same mating types, resulting in cytoplasmic inheritance: products of 665 bp and 470 bp from the promoter-containing and promoter-
the F1 derived from the O parent is determined for O in 94% of cases, and the F1 excised versions, respectively. All clones retained the promoter, although a
derived from the E parent is E in 98% of cases66. The frequency of mating-type fraction of them (14 of 23) expressed mating type O, as expected for mtAO
reversal is much lower at autogamy: ,1/50,000 in the O-to-E direction, and homozygotes. ND, not determined. f, mtA promoter retention in F1
1/3,000 in the E-to-O direction67. c, Cross of O-determined expression mutants heterozygotes from the cross shown in d, illustrated by a typical set of 4 mtBO/
mtAO, mtBO and mtCO to wild-type E cells. F2 mutant homozygotes formed in mtB1 F1 karyonides from the cross of an E-determined, O-expressing mtBO
the E cytoplasmic lineage express mating type O as a result of the mutation, but homozygote (as produced in c) to wild-type E cells. PCR5 showed that the mtA
remain determined for E (orange MAC). d, Cross of O-expressing, promoter was now retained in F1 karyonides from both parents; all were of
E-determined mutants mtAO, mtBO and mtCO (as produced in c) to wild-type E mating type E.

RESEARCH ARTICLE
Extended Data Figure 3 | Complementation of the mtBO and mtCO Clones containing detectable plasmid amounts expressed mating type E,
phenotypes with the wild-type alleles of GSPATG00026812001 and indicating that the GFP fusion protein is functional; the selfing phenotype (S) of
GSPATG00009074001, respectively. a, Structure of the MIC and MAC clone 11 may be due to some cells having lost the plasmid. PCR1 (bottom panel)
versions of the mtB gene, and of the GFP fusion transgene used for confirmed that all clones retained the mtA promoter. c, Structure of the
complementation. The coding sequence (open arrow) is shown with the MIC and MAC versions of the mtC gene, and of the plasmid used for
complete upstream and downstream intergenic regions. The MIC version complementation. The coding sequence (open arrow) is shown with the
contains two IESs (black boxes). Plasmid pmtB51-NGFP contains the MAC complete upstream and downstream intergenic regions. The MIC version
version with complete intergenic regions, and the EGFP coding sequence was contains one IES (black box). Plasmid pmtC51 contains the MAC version with
fused at the 59 end of the mtB coding sequence. Thick grey lines on either side 349 bp and 98 bp of upstream and downstream intergenic sequences,
represent plasmid vector sequences. b, PCR analysis and mating types of respectively. The plasmid-specific PCR8 amplifies a 419-bp product
E-determined mtBO mutant clones transformed with pmtB51-NGFP. PCR7 (Supplementary Table 6). d, PCR analysis and mating types of E-determined
(top panel) amplifies products of 1,148 bp from the plasmid, and of 419 bp from mtCO mutant clones transformed with pmtC51. PCR8 (top panel) shows that all
the endogenous mtB gene (Supplementary Table 6). The relative abundance of positive clones expressed mating type E. C1, C2 and C3, uninjected control
the two products gives an indication of plasmid copy number in each clone. C1 clones. PCR1 (bottom panel) confirmed that all clones retained the mtA
and C2, uninjected control clones. Mating types are indicated below each lane. promoter.

ARTICLE RESEARCH
Extended Data Figure 4 | Deep sequencing of scnRNAs from an early version of the mtA gene region. Coding sequences of the corrected mtA gene
conjugation time point (early meiosis). A total of 39,041,474 small-RNA model (PTETG5300016001, after correction of assembly indels and
sequences of 25 nucleotides in length were obtained by Illumina sequencing of re-annotation) and of the short gene downstream (PTETG5300018001) are
four libraries previously constructed from gel-purified molecules migrating at shown as red boxes interrupted by introns and IESs. The yellow box represents
23, 24, 25, or 26 nucleotides (from the 7hND7 total RNA sample12). the 195-bp segment of the mtA promoter that is excised in O cells; the 4 IESs are
a, 25-nucleotide reads were first mapped (no mismatch allowed) on possible shown as green boxes. 25-nucleotide reads mapping on the top or bottom
contaminant sequences (genomes of bacteria commonly found in cultures, strands of the region are colour-coded to indicate the number of times each one
P. tetraurelia mitochondrial genome, P. tetraurelia rDNA and other non- was sequenced: blue, one read; green, 2–9 reads; red, $10 reads.
coding RNAs). The remainder was then mapped on known MIC sequences (the d, Compositional profiles (nucleotide frequency on the left, and deviation from
‘MAC1IES’ genome, and 9 individual copies of the Sardine transposon3). To randomness on the right) of reads mapping to the MAC1IES genome, to the
determine whether any of the remaining reads could correspond to IES excision MAC genome, or to ‘PGM contigs’. For each set, logos are shown for all reads
junctions or to spliced transcripts, they were then mapped on the MAC genome (left), or for the non-redundant subset only (right). For the ‘MAC1IES’
and on the genome-wide set of spliced transcripts. Very few hits were found and and ‘PGM contigs’ sets, the logos computed from all reads clearly show the
these did not show the characteristic 59-UNG signature of scnRNAs (see 59-UNG signature typical of scnRNA guide strands, as is the case for the major
d), suggesting that these molecules represent longer forms of endogenous subset of reads starting with U (59U), while the minor subset of sequences
siRNAs and/or could be mapped because of IES or intron annotation errors. Of not starting with U (59A/C/G), which may represent the steady-state amount of
the remaining unmapped reads (,49%), close to one-half could be mapped on passenger strands, shows the complementary signature CNA at positions
‘PGM contigs’, a ,25-Mb preliminary assembly of MIC-specific sequences that 21–23. Deviation from randomness is greater when computed from all reads
are not collinear to MAC chromosomes3 and are thus likely to represent bona than when computed from the non-redundant subset only, indicating that
fide scnRNAs. b, Statistics about the coverage of the ‘MAC1IES’ genome molecules with the signature are intrinsically produced in higher amounts, or
(17,786,284 reads). The average coverage is similar for exons, introns and are more stable. The small set of reads that mapped only to the MAC genome
intergenic regions, but is ,2-fold higher for IES sequences. This may mean (putative IES excision junctions) does not show a clear 59-UNG signature,
either that scnRNAs are initially produced in higher amounts from IESs, or that suggesting that most of those are not scnRNAs. The same is true of the small
active degradation of scnRNAs homologous to MAC sequences is already number of reads mapping to exon–exon junctions (not shown).
under way at this early stage. c, Mapping of 25-nucleotide reads on the MIC

RESEARCH ARTICLE
Extended Data Figure 5 | Northern blot analysis of mtA-promoter RNAs. The top panels show hybridization with the 195-bp mtA-promoter
scnRNAs during autogamy of O or E cells. Mass cultures were allowed to probe, revealing accumulation of 25-nucleotide scnRNAs slightly later than
starve, and RNA samples were extracted during exponential growth (Exp) and expression of the meiosis-specific genes PTIWI09 and NOWA2. As a control,
then at different times (0–20 h) after the appearance of the first meiotic cells. the same blots were rehybridized with an oligonucleotide probe specific for a
Cells become committed for autogamy at a fixed point of the cell cycle, so that cluster of 23-nucleotide endogenous siRNAs on scaffold 2212,61, which are
the best synchrony that can theoretically be achieved is the duration of one abundantly produced at all stages of the life cycle (bottom panels).
cell cycle; in these experiments, the time between the first and the last cells to Quantification of the mtA-promoter scnRNA signal and normalization with
begin meiosis was ,12 h. a, Proportions of cells in different cytological stages at the siRNA signal did not reveal any significant difference in their amount or
each time point, as determined by DAPI staining. Veg, vegetative cells; Mei, timing between the two mating types (not shown). Previous studies showed
meiosis (crescent stage, meiosis I, meiosis II); Frg, cells with fragmented old that the double-strand breaks that initiate IES excision in the new MACs start
MAC but new MACs not yet clearly visible; Dev, cells with two clearly visible being detectable before the maximum of expression of the putative
developing new MACs; Kar, cells after the karyonidal division, with only one endonuclease PGM5,68 (no later than 10 h in these time courses). A PCR analysis
developing new MAC. b, Northern blot analysis of mRNAs for early (PTIWI09, of post-autogamous DNA samples confirmed that the mtA promoter was fully
NOWA2), middle (PGM), or late (PTIWI10) genes. The same blots were excised in mating type O, and fully maintained in mating type E (not shown).
hybridized successively with the 4 probes. c, Northern blot analysis of small

ARTICLE RESEARCH
Extended Data Figure 6 | Phylogenetic tree of different strains from most here are corrected in red. Strain V5-13 was probably mis-assigned to
P. aurelia species based on sequence polymorphisms in three nuclear genes. P. septaurelia through the same error as for strain GFg-1 (see Extended Data
This figure is modified with permission from figure 2 of Catania et al. Fig. 7). Species showing maternal inheritance or random determination of
Genetic diversity in the Paramecium aurelia species complex Mol. Biol. Evol. mating types are highlighted in red and blue, respectively.
2009, 26, 421–431 (ref. 32). Wrong species assignment of some strains studied

RESEARCH ARTICLE
Extended Data Figure 7 | Strain GFg-1 belongs to the same species as strain some F2 clones obtained by autogamy of these F1 heterozygotes. The parental
138, that is, P. octaurelia. GFg-1 is among a set of strains that were originally origin of each F1 clone was ascertained by a sequence polymorphism in the
assigned to P. septaurelia on the basis of conjugation tests with strain 38, a mitochondrial COXII gene (PCR9, Supplementary Table 6). PCR amplification
reference strain for that species. However, the stock of strain 38 used in these and sequencing of a segment of the mtA gene encompassing the promoter
tests was not 38, but instead some P. octaurelia strain, as shown by the (PCR10, Supplementary Table 6) showed that the two F1 clones deriving from
comparison of mtA sequences with those from the original strain 38 obtained the GFg-1 parent were heterozygotes, and that the mtA promoter was precisely
from ATCC (ATCC number 30575) and those from strain 138, a reference excised from both alleles (see Supplementary Data 1a). ND, not determined.
strain for P. octaurelia. a, Scheme of the cross GFg-1 (O) 3 138 (E). The mating Analysis of 6 viable F2 clones obtained by autogamy of the F1 clones deriving
types of parents were determined by cross-agglutination with P. tetraurelia from the 138 parent showed that 4 of them were homozygous for the GFg-1
tester lines. The green and blue arrows beside each cell represent the mtA gene, allele, whereas the other two were homozygous for the 138 allele; the mtA
colour-coded to indicate the GFg-1 and 138 alleles; the box at the 59 end promoter was retained in all cases. The evidence for successful genetic exchange
symbolizes retention of the mtA promoter in the MAC genome on the E side of between strains GFg-1 and 138 and for viable recombinant F2 progeny
the cross. b, Molecular characterization of two pairs of F1 heterozygotes, and of demonstrates that these strains belong to the same species; that is, P. octaurelia.

ARTICLE RESEARCH
Extended Data Figure 8 | Genetic and molecular analysis of the cross clones. Given the requirements of IES excision in P. tetraurelia and in sibling
between strains 227 and 38 of P. septaurelia. a, Maternal inheritance of species36, this can be predicted to make the same deletions impossible in the
mating types in the wild-type strain 227. b, Strain 38 is genetically restricted to mtB38 allele, which would explain the constitutive E determination effect. The
O expression, but constitutively determined for E. The mutation identified in full-length mtB38 pseudogene in the MAC of 38 cells would indeed be expected,
strain 38 is of particular interest because it affects both the expression and the after a cross to 227, to protect the highly similar zygotic mtB227 allele against
inheritance of mating types, which suggests that it lies in a gene that controls coding-sequence deletions in the derived F1, through titration of homologous
mating-type determination through an alternative rearrangement. Indeed, scnRNAs. d, Molecular analysis of the 38 3 227 cross. To verify this maternal
known P. tetraurelia mutations fall in two distinct categories. mtAO, mtBO and effect, we crossed an E-expressing 227 clone (pmtB51-transformed clone T,
mtCO prevent expression of type E but have no effect on the rearrangement that same as in Fig. 5b; C, uninjected control) with strain 38, and F1 heterozygotes
determines mating types or on its maternal inheritance, whereas mtFE lies in a were tested for mating types and for mtA expression by northern blotting. As
trans-acting factor required for a subset of rearrangements during MAC expected, F1 heterozygotes deriving from the 38 parent (1b and 2b, as
development but has no effect on the expression of mating types during sexual determined by sequencing of a mitochondrial polymorphism) were E and
reactivity. The only type of mutation that can be envisioned to affect both expressed mtA, indicating that the incoming mtB227 allele had been rearranged
expression and determination/inheritance would be a mutation preventing into a functional, full-length form in the MAC. After autogamy of F1 clone 2b,
expression of a functional protein required for E expression, and at the same 24 independent F2 homozygotes were isolated and tested for mating types.
time preventing in cis (by destroying a potential Pgm cleavage site) the Consistent with the Mendelian segregation of mtB alleles, 12 were O and 12
rearrangement that normally inactivates this gene in the MAC of wild-type were E (Supplementary Table 5); northern blot analysis of 3 clones of each type
O cells. The mtXIII allele of strain 38 restricts cells to O expression in a showed that only E clones expressed mtA. e, All F2 clones maintained the full-
recessive manner, but also has a maternal effect that enforces constitutive E length mtB gene in the MAC. PCR11 (Supplementary Table 6) amplifies an
determination in sexual progeny. Notably, elegant experiments showed the 888-bp fragment from the MAC version of mtB38, and an 893-bp fragment
latter effect to be dominant20: the sexual progeny of a cell carrying at least one from the full-length MAC version of mtB227. C1 and C2, control PCR11 on two
mtXIII allele can never be determined for O or transmit O determination. O clones of strain 227, showing the 806-bp and 744-bp fragments resulting
c, Sequencing of the mtB38 allele revealed features that may account for both from the two alternative coding-sequence deletions. f, Mating types co-
effects. A frameshift mutation makes it a pseudogene, explaining the genetic segregate with mtB alleles among F2 homozygotes. Digestion of the PCR
restriction to O expression. In addition, a 6-bp deletion removes one of the IES- products with AluI distinguishes the 38 and 227 alleles. mtA and mtC alleles
like boundaries used in the mtB227 allele for coding-sequence deletions in O segregated independently (Supplementary Table 5).

RESEARCH ARTICLE
Extended Data Figure 9 | Transformation of P. septaurelia 227 O cells with E-expressing clones, as indicated below each lane. This was the case of the 4
the full-length mtB227 gene induces heritable E expression. a, Plasmid high-copy clones; other clones were presumably O, although it was impossible
pmtB227, containing the full-length E MAC form of mtB227, was microinjected to make sure that the cells were sexually reactive (indicated by ‘-’). Injected
into the MACs of O cells of strain 227, and 12 injected clones were tested by clones were then taken through autogamy, and the mass progenies were again
PCR11. b, This PCR amplifies an 893-bp product from the transgene, and tested for mating types. In this case, the mass progenies from clones 9 and 11,
products of 744 and 806 bp from the two internally deleted MAC forms of the which proved to be pure E, were used as E testers to carry out a full test for some
endogenous mtB227. C1 and C2, control uninjected clones. Clones 4, 8, 9 and 11 of the other progenies. Of the other two transformed clones expressing E, clone
contained high copy numbers of the transgene, so that the PCR products from 4 gave rise to a selfing progeny (S), probably a mix of O and E clones, whereas
the endogenous gene are not detectable. Because only O tester lines were clone 8 gave rise to a pure O progeny.
available, the mating types of injected clones could be ascertained only for

ARTICLE RESEARCH
Extended Data Figure 10 | A general model for mating-type determination P. tetraurelia and P. octaurelia, mating type O is determined during MAC
in P. aurelia species. In the three species examined, mating type E depends on development by excision of the mtA promoter as an IES, preventing expression
expression of the mtA protein during sexual reactivity. mtA transcription in of the gene. In P. septaurelia, mating type O is determined by the excision of
turn requires the mtB and mtC gene products (the requirement for mtC in segments of the mtB coding sequence as IESs, which similarly prevents mtA
P. septaurelia, and for both genes in P. octaurelia, remains to be verified). In expression.

Epigenetic Mating-Type Inheritance

Uploaded by

Copyright:

Available Formats

Epigenetic Mating-Type Inheritance

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Epigenetic Mating-Type Inheritance

Uploaded by

Copyright:

Available Formats

ARTICLE doi:10.

Genome-defence small RNAs exapted for

determined by the rearrangement of one of the genes required for E

mtA rearrangements determine mating type

a in mtA transcription and E expression in both cases (Fig. 3b and Extended

METHODS Confocal analysis of mtA–GFP localization. Cells were fixed in 2% paraformalde-

©2014 Macmillan Publishers Limited. All rights reserved

©2014 Macmillan Publishers Limited. All rights reserved

©2014 Macmillan Publishers Limited. All rights reserved

©2014 Macmillan Publishers Limited. All rights reserved

©2014 Macmillan Publishers Limited. All rights reserved

©2014 Macmillan Publishers Limited. All rights reserved

©2014 Macmillan Publishers Limited. All rights reserved

©2014 Macmillan Publishers Limited. All rights reserved

©2014 Macmillan Publishers Limited. All rights reserved

©2014 Macmillan Publishers Limited. All rights reserved

©2014 Macmillan Publishers Limited. All rights reserved

You might also like