.J. Xol. Hiol. (1992) 225> 661-678

A High Degree of Macronuclear Chromosome Polymorphism is Generated by Variable DNA Rearrangements in Paramecium primaurelia during Macronuclear Differentiation Franqois Caron Laboratoire de Ge’ne’tique Mole’culaire Ecole Normale Supe’rieure 46 rue d’Ulm 75230 Paris Cedex 05 France (Received

4 November

1991; accepted 12 February

1992)

DNA rearrangements in Paramecium lead to the formation of macronuclear chromosomes. the sizes of which range from 50 and 800 kb (1 kb is lo3 base-pairs). This process does not appear to be a simple size reduction of the micronuclear chromosomes by specific and reproducible DNA sequence elimination and chromosomal breakage followed by chromosomal amplification. On the contrary, this process generates a variety of different, but sequence-related, macronuclear chromosomes from a unique set of micronuclear chromosomes. This paper describes an attempt to understand the nature of the diversity of the macronuclear chromosomes and the mechanisms of their production. The structure of three macronuclear chromosomes, 480, 250 and 230 kb in size, have been determined utilizing chromosome-jumping and YAC-cloning techniques. The two smallest chromosomes correspond roughly to the two halves of the longest chromosome. The main contribution t’o the diversity arises from the chromosomal ends and is due to variable positions of the telomere addition sites and/or to variable rearrangements of DNA sequences. The 480 kb chromosome contains a region of variable length, which is likely to be due to a variable deletion, located at the position of telomerization seen in the two small chromosomes. A model of chromosomal breakage is proposed to rationalize this result where micronuclear DNA is first amplified, broken and degraded to various extent from the newly formed ends. which subsequently are either telomerized or religated. Potential implications of these processes for gene expression is discussed. Known phenotypes that. have a macronuclear determinism could be explained by this type of process. Keywords:

macronuclear

chromosomes; genome reorganization; alternative Paramecium primaurelia; jumping techniques

1. Introduction After each sexual process, ciliates rearrange one copy of their genome inside the macronucleus (the somatic nucleus). Another genomic copy of the zygotic nucleus is maintained unmodified as the micronucleus (the germinal nucleus). At the next sexual event, the micronucleus undergoes meiosis and one of the haploid products will contribute to the formation of the new zygotic nucleus, whereas the macronucleus will be destroyed and replaced by a new one issued from the new zygotic nucleus, and so on. Therefore, the genetic continuity is main-

rearrangement;

tained by the micronucleus (for a review, see Yao, 1989). In most ciliates, progress in understanding this genomic rearrangement has been made possible by the purification of micronuclear DNA from macronuclear DNA and the direct comparison of the macronuclear DNA sequences with their micronuclear precursors. At least three types of events occur in the DNA rearrangement process: chromosomal breakage, internal sequence elimination and DNA amplification. In Tetrahymena and in different species of hypotrichs, chromosomal breakage (Yao et al.. 1987. 1990; Raird & Klobutcher, 1989) and

662

k’. Caron

internal sequence elimination (Klobutcher et al.. 1984; Austerberry & Yao. 1987: Ribas-Aparicio et al.. 1987; Hunter ef al., 1989; ,Jahn et al., 1989: Godiska & Yao, 1990; Lorraine Tausta &, Klobutcher, 1990; Lorraine Tausta et al., 1991) occur rearrangement as reproducible genomic through the recognition of specific &-sequences that art as targets of the enzymatic machinery. In some cases, alternative internal DNA sequence elimination events have been reported, adding an extra level of complexity and variety (Cartinhour & Herrick, 1984; Howard & Blackburn, 1985; Herrick et al., 1987a,b: Austerberry & Yao, 1988; Kile et al., 1988). DNA is amplified both before and after chromosomal breakage and internal sequence elimination (Ammermann. 1965; Yokovama & Yao. 1982). The final degree of amplification varies from one species to another but,, apart from some genes that, are overamplified (for instance. Tetrahymencr rDNA). it appears fa.irly constant wit,hin one defined species. In the case of Paramecium, the absence of a reliable, efficient method for purification of micronuclear DNA has impeded the direct observation and study of the mechanisms of DNA rearrangement. Nevertheless, chromosomal breakage may be deduced from the comparison of the average size of macronuclear chromosomes (300 kb: 1 kb is lo3 base-pairs?) with the average size of micronuclear chromosomes (2000 to 3000 kb), calculated from the size of the genome and the number of chromosomes visualized at meiosis (Jones. 1956; Preer & Preer. 1979; Caron & Meyer, 1989; Phan et al., 1989). Some DNA amplification occurs before chromosomal breakage since, for a given macronuclear telomere, the positions of the telomere addition sites are numerous and variable (Baroin et al., 1987; Forney & Blackburn, 1988; Keller. A.-M. et al., unpublished results). The final level of amplification is about IOOO-fold and from the renaturation kinetics it appears to be fairly constant (McTavish & Sommerville, 1980). Alternative DNA rearrangements that involve DNA sequence elimination and/ or extra DNA sequence addition have been detected in Paramecium tetraurelia at the ends of the macronuclear chromosomes that contain the A surface antigen gene and they are, at present, the only of internal sequence elimination in evidence Paramecium (Forney & Blackburn. 1988). Since. during vegetative growth, transcription takes place exclusively in the macronucleus, it would not, be surprising to find that expression of some genes is modulated by these variable DNA rearrangements. This has been shown to be the case of the alleles of the gene encoding the G surface antigen in Paramecium primaurelia heterozygous cells (Keller, A.-M. et al., unpublished results).

Here, 1 have investigated the entire long-range structure of three macronuclear chromosomes of different sizes (230, 250 and 480 kb), t,wo of t’hem harbouring the G surface-antigen gene of strain 156 of Paramecium primaurelia at one end. By restriction mapping, I show that the 250 kb and the 230 kb chromosomes correspond roughly t,o the two halves of the 480 kb chromosome. Tn spite of the absence of any specific information about the mitronuclear DNA structure, restriction-mapping identities and genetic evidence strongly support, t’he fact that they are generated by an albernative D?iA rearrangement, which is likely to be an incompletr chromosomal breakage of micronuclear DNA. Thus, during macronuclear differentiat,ion. a Para,mrcium cell would build up a variety of macronuclear chromosomes by using different ways of brea(king its micronuclear chromosomes.

2. Materials and Methods (a)

&ruins

used and growth conditions

Caryonides are the cells issued from the first division after conjugation or autogamy. In these cells, only I event) of macronuclear differentiation is observed. A caryonidal clone consist,s of the descendants of a c’al’ytrnide. fa~ranzrcium primaure2ia strain 156 cells from a caryonidal clone were grown at 24°C in Scotch grass infusion inoculated the .dap before use with K/rDslr/ln pneumonine and supplemented with 1 ,~g If-ait~ostrrol~mi purchased from Merck (Sonneborn. 1970).

Total DNA was extracted either in solut’ion or as agarose inserts. The protocol used for t’he DNA preparation in solution was that published by Sambrook et al. (1989) with great care taken to avoid DT;A breakage: in particular the phenol step was performed by gently rotating the tubes in an almost horizontal position for 30 min. The average size of the DNA molecules prepared in this way was 250 to 300 kb and adequate for restriction analysis of DNA fragments up to 100 to 150 kb. All the libraries constructed for this work and all the restriction analyses were made from only 1 DKA preparation from a 12 1 culture of a caryonidal clone with a cell density of about lOOO/ml. For the analysis of longer fragments or for chromosomal DNA, agarose insert,s were made according to the method of Schwartz &. Cant,or (1984).

(c) (‘HEtr’ Contour-clamped electrophoresis with the same Chu et aE. (1986), used was 0.25 x ture was 12°C from 5 to 50 s.

(d) Jumping TAbbreviations used: kb, lo3 base-pairs; CHEF: Contour-clamped homogeneous electric field; X-gal, 5-bromo-4-chloro-3-indolyl-fi-o-galactoside; YAC, yeast artificial chromosome; bp, base-pair(s); Cbs. Chromosomal Breakage Sequence.

(i)

Modi$cation

electrophorrsis

homogeneous electric field ((‘H EF) was performed on a homemade apparatus electrode geometry as that published b> using the following conditions: the buffer TBE (Sambrook et al.. 1989), the temperaand different switching times used ranged

libraries

of pf,‘CIR

Since EcoRI was used as the frequent cutter in the construction of the jumping libraries. the unique site present in the polylinker of pUCl8 had to be removed.

Macronuclear

Chromosome Polymorphism

Various simple approaches could be used, but the one presented in Fig. 2(c) had the advantage of removing the EcoRI site of pUC18 by base deletion while preserving the KpnI site and the reading frame of the a-peptide coding part of the a-galactosidase gene. This last feature, although not indispensable, allowed the use of the colour test to measure the ratio of the recombinants with respect to the parent vector in the jumping library. pUC18 plasmid was first linearized using EcoRI and its cohesive ends were trimmed with S, nuclease to remove the protruding single-strand. The plasmid was then cut with Asp718 (an isoschizomer of KpnI with 5’, instead of 3’. protruding ends) to remove a few base-pairs (see Fig. 2) and the end corresponding to the Asp718 site was made blunt by filling with dNTPs using the Escherichia coli DNA polymerase Klenow enzyme. The resulting linear plasmid was circularized and ligated. The overall procedure removed 12 base-pairs in the polylinker of pUC18. Transformation of the bacterial strain TGl (Sambrook et al., 1989) gave rise to blue colonies in the presence of 5bromo-4-chloro-3-indolyl-fl-D-galactoside (X-gal). One. which contained a plasmid (pUC18E) with a KpnI site and no EcoRI site was selected and the presence of the expected deletion was checked by sequencing the polylinker. (ii) Construction of jumping libraries Macronuclear Paramecium DNA (01 to @5 pg) cut with Asp718 and 1 pg of the Asp718-linearized and dephosphorylated pUC18E plasmid (see Results) were ligated t)ogether overnight at 15°C in a total volume of 05 ml. XaCl (10 ~15 M) was added to obtain a final NaCl concentration of 100 mM for digestion with EcoRI. After complete digestion, DNA was extracted with phenol and precipitated with ethanol in the presence of 10 c(g tRNA as a carrier. The 2nd ligation was performed for 16 h at 15°C in a total volume of @5 ml. This final mixture was used directly for transformation (see Results). Different recA bacterial strains (DH5: Sambrook et al., 1989; SCSl; Stratagene) were used for transformation with equal success if their transformation efficiency was greater than lo8 transformants per fig of DNA. Transformation of TGl (Sambrook et al., 1989) with a small aliquot of the ligation mixture and measurement of the relative number of white colonies in the presence of X-gal gives the proportion of recombinant over wild-type clones. (e) YAC library (i)

Vector modijkation The unique KpnI site located in the middle of the His3 gene in pYAC3 (Burke et al., 1987) was deleted by digestion with Asp718 followed by brief trimming of the ends using nuclease BaZ31. The fact that the resulting His3 gene is no longer functional is of no importance, since its role in the plasmid is to act as a spacer between the 2 telomeric repeats. After recircularization and transformation of TGl, a recombinant plasmid devoid of the KpnI site was selected. This plasmid was then cut at the unique SnuBI site located inside the SUP4 intron and normally used as the cloning site. The linear plasmid was ligated in the presence of an 8 bp phosphorylated KpnI linker (Biolabs) and, after transformation, a recombinant that was linearized by KpnI was selected. To be sure that the recombinant plasmid contained only 1 KpnI linker, it was again digested to completion with KpnI, recircularized and used for transformation of TGl. The resulting plasmid was used for yeast spheroplast transformation: the insertion of the 8 bp linker affects neither the splicing

in Paramecium

663

of the tRNA intron, as checked by the white colour of the transformed colonies, nor the transformation efficiency (more than lo5 clones per pg of DNA). (ii) DNA preparation, and ligation. High molecular weight Paramecium DNA obtained in solution using the protocol described above was digested with Asp718, and run on a low melting temperature agarose gel (1 o/0 (w/v)) in a CHEF apparatus (switching time 10 s). The part of the gel corresponding to the marker (a bacteriophage 1 DNA ladder) was stained and exposed to ultraviolet light to locate the 60 to 100 kb region. This region was cut out from lanes loaded with the KpnI digest that had not been stained and exposed to ultraviolet light. After a 30 min equilibration with the agarase buffer (10 mM-[2-N-Morpholinolethanesulphonic acid (MES) (pH 65). 5 mM-EDTA. 150mM-NaCl, 05 mw-spermine, @5 mi%-spermidine), the gel bands were melted at 65°C and the agarose digested overnight at 40°C with agarase enzyme (Calbiochem) at a concentration of 20 to 50 units/ml. The presence of diamines protects DNA from breakage by shearing (Couto et al., 1989). After a gentle extraction with phenol and chloroform:octanol (24:l by volume), DNA was precipitated with 1 vol. isopropanol in the presence of 5pg tRNA as a carrier. The DNA pellet was washed twice with ethanol, resuspended in TE pH 8 (Sambrook et al., 1989) and the absence of detectable DNA breakage checked by CHEF electrophoresis. Approximately 1 to 2 pg DNA was recovered. The yeast artificial chromosome (YAC) plasmid vector was digested with Asp718 and BarnHI, dephosphorylated with bacterial alkaline phosphatase and the 2 arms purified on low melting temperature agarose using the agarase treatment described above. The ligation of Paramecium DNA with the 2 arms was carried out at 9°C for 3 days in the presence of bovine serum albumin (01 mg/ml: total volume 20 ~1). The remaining traces of spermine and spermidine do not inhibit T4 ligase. An equal volume of 2 x SCaCl (SCaCl: 1 M-sorbitol, 10 mM-CaCl,) was added just before spheroplast transformation. (iii) Yeast transformation The yeast strain SX4-6A (a adeZ-1 his3-532 trpl-289 ura3 inoscan’? a gift from B. Dujon) was used for transformation. It hm a high efficiency of transformation (more than lo5 transformants per pg supercoiled pYAC3 plasmid) and very reproducible kinetics of spheroplasting. The following method for spheroplasting and transformation was elaborated from the protocols of B. Dujon (unpublished results) and Burke et al. (1987). A 100 ml culture of exponentially growing yeast cells was harvested at a concentration of 10’ cells/ml (08 < A6,,,, < 1 for the strain SX4-6A), washed in 10 ml SE (1 M-sorbit)ol, 20 mM-EDTA pH 8), then incubated for 10 min in 10 ml SE plus 0.1 y0 (v/v) /?-mercaptoethanol. Cells were then washed in 10 ml SCE (1 M-sorbitol, 0.1 M-sodium citrate (pH 5.8). 20 mM-EDTA) and finally resuspended in 10 ml SCE. Digestion was performed at 37°C for 10 min with zymoliase (Seikagaku Kogyo Co. 1OOT) at a final concentration of 05 units/ml. The incubation time varies from one strain to another and must be checked for each new stock solution of zymoliase, but, once calibrated, it is valid for several months (the life-time of the stock solution). Following this step, the cells were always kept at 4 “C and great care taken in handling them (centrifugation at 300 g for 10 min or less, gentle resuspension). Spheroplasts were then washed twice in 10 ml SCaCl and finally resuspended in lml SCaCl. For the transformation.

664

F. Caron

@l ml of the spheroplast solution was mixed with 1~1 carrier DNA (sonicated sperm salmon DNA, 5 mg/ml) and the ligation mixture (see above) and then incubated at room temperature for 15 min followed by the addition of 1 ml PCT (20yb (w/v) polyethyleneglycol 8000 (Sigma), 10 mM-CaCl,, 10 mM-Tris.HCl (pH 7.4)), gentle mixing and incubation for 15 min at room temperature. After centrifugation at 150g for 10 min, spheroplasts were gently resuspended in 150ml SOS medium (25”/6 (v/v) YPD broth (1 y0 (w/v), yeast extract, 2% (w/v) peptone, 2% (w/v) glucose), l&i-sorbitol, @Ol M-CaCl,, 10 pg uracil/ml) and incubated for 30 min. The spheroplast suspension was spread on selective medium without uracil and incubated at 30°C for 2 to 3 days. A total of 10 red colonies were picked at random and shown to contain a YAC with a Paramecium DNA insert. About 1200 recombinants (red colonies) were obtained, among which only 700 were screened with the probe 4R. (f) Partial library of short Kpnl fragnwnta Paramecium DNA was digested to completion with Asp718 and ligated in the presence of an excess of pUCl8 plasmid linearized at the Asp718 site and dephosphorylated. After bacterial transformation, about 500 colonies were obtained. Screening with either probe 4L or 7L gave about 4 positive recombinants containing the expected KpnI fragment. (g) Partial EcoRi library in bacteriophaye lambdu EcoRI restriction fragments in the 15 to 22 kb range obtained from a partial digest of Paramecium DNA (Sambrook et al., 1989) were purified from low melting temperature agarose by agarase treatment. The 2 arms of EMBL4 were also purified in the same way. After ligation and encapsidation (Amersham kit), more than 120,000 recombinants were obtained and screened with the selected probe. Restriction mapping analysis of positive recombinants was done by the cos-mapping technique (Rackwitz et al., 1984). (h) Restriction analysis of s&-selected chromosomes In order to analyse each individual chromosome by restriction mapping, intact chromosomal DNA was first separated by CHEF electrophoresis in a 1% (w/v) low melting temperature agarose gel with a 50 s commutation time. In instances of short range restriction mapping (