Biochemical Society Transactions

Topological Aspects of DNA Function and Protein Folding

Topological similarity between the 2μm plasmid partitioning locus and the budding yeast centromere: evidence for a common evolutionary origin?

Makkuni Jayaram, Keng-Ming Chang, Chien-Hui Ma, Chu-Chun Huang, Yen-Ting Liu, Soumitra Sau

Abstract

The partitioning locus STB of the selfish plasmid, the 2μm circle, of Saccharomyces cerevisiae is essential for the propagation of this multi-copy extra-chromosomal DNA element with nearly chromosome-like stability. The functional competence of STB requires the plasmid-coded partitioning proteins Rep1 and Rep2 as well as host-coded proteins. Host factors that associate with STB in a Rep1- and Rep2-dependent manner also interact with centromeres, and play important roles in chromosome segregation. They include the cohesin complex and the centromere-specific histone H3 variant Cse4. The genetically defined point centromere of S. cerevisiae differs starkly from the much more widespread epigenetically specified regional centromeres of eukaryotes. The particularly small size of the S. cerevisiae centromere and the association of chromosome segregation factors with STB raise the possibility of an evolutionary link between these two partitioning loci. The unusual positive supercoiling harboured by the S. cerevisiae centromere and STB in vivo in their functional states, unveiled by recent experiments, bolsters the notion of their potential descent from an ancestral plasmid partitioning locus.

  • centromere
  • partitioning locus
  • 2μm plasmid
  • Saccharomyces cerevisiae

Introduction

The 2μm plasmid is an endogenous multi-copy extra-chromosomal DNA circle present in the nucleus of the budding yeast Saccharomyces cerevisiae [1,2]. The plasmid, which has an average copy number of 40–60 molecules per haploid cell, propagates itself with nearly the same stability as the chromosomes of its host. The loss rate of the plasmid is 10−4–10−5 per cell division, which is barely an order of magnitude higher than that of a chromosome. By contrast, artificial plasmids containing a functional replication origin (ORI) and a selectable marker (called ARS plasmids), when introduced into S. cerevisiae, are rapidly lost in a growing cell population, unless selection for the plasmid-borne marker is maintained. Even plasmids harbouring a copy of a chromosomal centromere (CEN) are much less stable than the 2μm plasmid under non-selective growth. The 2μm plasmid does not confer any obvious selective advantage on its host, and neither does it pose a significant metabolic burden at its steady-state copy number. S. cerevisiae strains cured of the plasmid [cir0] display no apparent defects under laboratory conditions. In fact, [cir0] strains have a slight advantage over isogenic [cir+] strains in a long-term competition growth assay [3].

The 2μm plasmid is considered to be a highly optimized selfish genome whose genetic organization is streamlined for long-term coexistence with its host genome (Figure 1A). The plasmid stability system, responsible for the equal or nearly equal segregation of replicated plasmid molecules to daughter cells, consists of the plasmid-coded proteins Rep1 and Rep2 plus the partitioning locus STB [1]. The partitioning system functions by overcoming the ‘mother bias’, which tends to favour the preferential retention of ARS plasmids in the mother cell compartment [46]. The plasmid also houses a copy number correction system, which restores copy number when it falls owing to rare missegregation events. This DNA amplification system comprises the Flp site-specific recombinase and its target sites (FRTs) present in inverted orientation within the plasmid genome [7,8]. Amplification is triggered by a Flp recombination event during bidirectional plasmid replication, causing one replication fork to be inverted with respect to the other. The resulting rolling circle replication spins out multiple tandem plasmid copies that may be resolved into monomeric plasmid units. This mechanism permits a rise in copy number without violating the cell cycle statute that each plasmid ORI fires once, and only once, during an S-phase. An intricate regulatory network (Figure 1A), with positive and negative elements, ensures rapid amplification response when required without incurring the risk of runaway increase in plasmid copy number [911]. Together, the partitioning system and amplification system account for the evolutionary success of the 2μm plasmid as a benign parasite genome.

Figure 1 The genetic organization of the 2μm plasmid of S. cerevisiae: regional and point centromeres

(A) The double-stranded circular plasmid is shaped as a dumb-bell to denote a long inverted repeat sequence that divides the genome into two unique regions. The Rep1 and Rep2 proteins, together with the STB locus, constitute the plasmid partitioning system. The Flp recombinase, along with the FRT sites, is responsible for plasmid copy number maintenance. The Raf1 protein is a positive regulator of amplification. ORI denotes the plasmid replication origin. The regulatory network, comprising the Rep1, Rep2 and Raf1 proteins, that controls plasmid gene expression is indicated. (B) The STB locus may be arbitrarily divided into two parts: the functional STB-proximal and the regulatory STB-distal. STB-proximal consists of a five copy tandem array of a 60 bp consensus element. A transcription terminator located within STB-distal prevents two plasmid transcripts from entering STB-proximal. (C) The ~75-kb-long regional centromere of fission yeast comprises a ~25-kb central region that includes a ~5-kb core flanked immediately by two large inverted repeats (imrs, innermost repeats). The distal borders of the centromere harbour smaller repeats (otrs, outer repeats). (D) The 125-bp-long point centromere of the budding yeast is organized into three well-defined centromere DNA elements (CDE I, II and III). A common identification mark of both types of centromeres is the presence of nucleosomes in which histone H3 is replaced by the H3 variant CENP-A (Cse4).

In the present article, we provide evidence for a striking similarity in the in vivo topology of STB chromatin and that of CEN chromatin [12,13]. In conjunction with other lines of circumstantial evidence, the shared DNA topology is consistent with the evolution of these two partitioning loci from a common ancestor.

The partitioning locus of the 2μm plasmid: an assembly site for host factors required for chromosome segregation

The STB locus is located between the 2μm plasmid ORI and the 3′-end of the Raf1 coding region, nestled between the recognition sites for PstI and AvaI restriction enzymes (Figure 1B). The locus can be arbitrarily divided into an origin-proximal region and an origin-distal region, STB-proximal and STB-distal respectively [14]. The STB-proximal consists of a 60 bp consensus element repeated five times in a tandem array and appears to harbour the plasmid partitioning potential. The STB-distal, which houses a transcription termination signal, is likely to be responsible for maintaining STB-proximal in its functional transcription-free state.

The association of Rep1 and Rep2 with STB is a prerequisite for equal plasmid segregation. In addition, a number of host factors with important functional roles at CEN (and in chromosome segregation) also associate with STB in a Rep1- and Rep2-dependent manner [1519]. Two such factors are the cohesin complex and the CEN-specific histone H3 variant Cse4 or CENP-A [13,16,18,20]. Whereas the cohesin complex is crucial for the one-to-one segregation of sister chromatids, Cse4 substitution for histone H3 is responsible for the unique nucleosome identity of CEN [21,22]. Whether the associations of these factors at STB are functionally relevant or merely relics of their possible shared evolutionary history can be debated. It is clear that both Cse4 and cohesin associate with STB in a highly substoichiometric fashion. Nevertheless, there is strong evidence for the functional coupling between the 2μm plasmid and chromosomes in segregation. Several conditions that cause chromosomes to missegregate also cause the plasmid to missegregate in tandem with the bulk of the chromosomes. One plausible model consistent with the available evidence is that plasmids segregate by physically tethering to chromosomes and hitchhiking on them [2,18].

Centromere organization among eukaryotes: Cse4 (CENP-A) is a common epigenetic mark of regional and point centromeres

Regional centromeres (Figure 1C) that typify the predominant majority of eukaryotic organisms are long but variable in length, are quite divergent in their DNA sequences, are epigenetically specified, and are almost always embedded within heterochromatin [23]. Components of the heterochromatin machinery and RNAi (RNA interference) system are important for their establishment. Chromosomal loci with no previous history of housing a centromere can transform into ‘neocentromeres’, and can be propagated through cell divisions. By contrast, the point centromere [24,25] (Figure 1D), exemplified by the centromeres in S. cerevisiae, is short (~125 bp), is genetically defined and is functional when transplanted into a non-native niche, for example, a plasmid. It consists of three DNA elements: CDE I (~10 bp), CDE II (~85 bp; AT-rich) and CDE III (~25 bp). The CBF3 complex, which interacts specifically with CEN DNA, provides the scaffold for organizing the kinetochore complex. The point centromere is restricted to the small fungal lineage of Saccharomycetaceae (budding yeasts). The origin of the point centromere, as well as its rarity, poses a serious conceptual challenge in centromere evolution [26].

A common epigenetic feature of the regional and point centromeres is that both house the specialized nucleosome in which histone H3 is replaced by CENP-A/Cse4.

Nucleosome organization at CEN and STB

Although there is no question that the Cse4-containing nucleosome is a unique identifier of the centromere, the composition of the CEN-specific nucleosome remains unsettled and controversial. Evidence has been presented for a hemisome containing a histone tetramer (H2A-H2B-Cse4-H4) or an octamer (H2A2-H2B2-Cse42-H42) or even a hexamer lacking H2A and H2B, but containing a histone chaperone (Cse42-H42-Scm32) [2731]. The spectre of a CEN nucleosome containing both H3 and Cse4 has also been raised [32]. The small size of the budding yeast point centromere is expected to limit its occupancy by a single Cse4-containing nucleosome. Although this indeed is the case [33], the notion of a single specialized nucleosome marking each chromosome CEN has been revised. Careful quantitative estimates suggest that additional Cse4-containing nucleosomes are likely to be associated with pericentric regions [34,35].

Cse4 has been detected at STB using chromatin immunoprecipitation assays [13,16]. Cse4–STB association, which requires the Rep1 and Rep2 proteins, occurs de novo during the G1–S transition stage of each cell cycle. This timing is the same as that for Cse4–CEN association [16,31,36]. Following recruitment, Cse4 stays put at CEN throughout the cell cycle and into the ensuing G1-phase of the following cell cycle [16,36]. By contrast, Cse4 exits STB during late telophase, presumably at the time of spindle disassembly [16]. Although the association of Cse4 at STB is consistent with the assembly of a Cse4-containing nucleosome, the substoichiometric nature of this association raises some uncertainty regarding this assertion. Even if STB harbours a Cse4-containing nucleosome, not all plasmid molecules possess such a nucleosome.

DNA topology within the CEN and STB chromatin

The Cse4-containing nucleosome has been shown to induce positive supercoiling at the budding yeast centromere [12]. The linking number difference (ΔLk) between the functional and non-functional states of CEN in vivo is estimated to be approximately 2 units. This would be consistent with a Cse4-containing nucleosome present at a functional CEN engendering a non-standard right-handed DNA writhe. When Cse4 is absent from a non-functional CEN, a standard H3-containing nucleosome with the normal left-handed DNA writhe would occupy it. As a result, the topological difference between CEN (Cse4) and CEN (H3) is given by ΔLk=+1−(−1)=+2 (Figure 2A). The topology of the in vitro assembled Cse4- or CENP-A-containing nucleosomes does not conform to a unique handedness. Depending on the conditions employed for assembly, these nucleosomes can adopt either a right-handed DNA wrap [12] or a left-handed DNA wrap [30,37]. Crystal structures of (CENP-A-H4)2 [37] and a CENP-A-containing histone octamer complexed with DNA [38] support the canonical left-handed DNA writhe for the CENP-A-containing nucleosome.

Figure 2 The topology of CEN and STB DNA

(A) Assuming that DNA is wound around a Cse4-containing nucleosome and an H3-containing (standard) nucleosome with opposite handedness, the replacement of one by the other would result in a net ΔLk change of 2 units. (B) The CENSTB reporter plasmid can be resolved into separate STB- and CEN-containing circles by R recombinase-mediated DNA exchange at the RRT sites. (C) The individual topoisomers of the STB and CEN circles are resolved by electrophoresis of DNA isolated from metaphase cells in the presence of chloroquine, followed by Southern blot analysis [39]. The ΔLk values are estimated from the centres of the topoisomer distribution tracings. The relaxed topoisomer (0) and the nicked circle (N) are not well resolved under the conditions employed. (D) Nocodazole treatment, which blocks Cse4–STB association, shifts the topology of the STB plasmid towards the more negative state.

The active form of the STB chromatin, analogous to CEN, harbours positive supercoiling [12,39]. The topology analyses have been performed using plasmids containing a copy each of CEN and STB. Such plasmids, when introduced into S. cerevisiae, maintain a copy number close to 1. Furthermore, their design is such that the CEN sequence, bordered by the target sites (RRTs) for the R site-specific recombinase, can be excised as a small circle (Figure 2B) by inducing the recombinase in G1-arrested cells. After releasing the cells from arrest, the topologies of the STB- and CEN-containing circles are assayed during the metaphase stage of the cell cycle. The CENSTB plasmid design permits qualitative and quantitative comparisons of the topologies of these loci under conditions that affect their functions differentially.

In the presence of both Rep1 and Rep2, the topoisomer distribution of the STB plasmid is shifted towards positive ΔLk compared with the distribution in the absence of Rep1 and Rep2 (Figure 2C). The magnitude of this shift, between functional and non-functional STB, is close to 2, the value reported for the ΔLk between functional and non-functional CEN [12]. Under these conditions, no change in the topology of the CEN circle is observed, as expected. The topology of STB, whose association with Cse4 is blocked by spindle disassembly in the presence of nocodazole, becomes more negative than that of the functional STB (Figure 2D). When the CEN and STB present on the reporter plasmid are individually or combinatorially inactivated, the corresponding topology distributions display the predicted shifts at least qualitatively (Figure 3A). The estimated ΔLk values for CEN inactivation by the ndc10 mutation from Figure 3(A) are somewhat smaller than those reported previously [12]. However, the topology shifts in the predicted directions are observed for CEN and STB after resolving them into separate circles by recombination (Figure 3B).

Figure 3 Topology changes in the CENSTB reporter plasmid upon inactivation of CEN or STB or both

Inactivation of STB and CEN are instituted by not supplying the Rep1 plus Rep2 proteins and by inactivating the Ndc10 protein respectively. (A) The topology analysis is performed on the CENSTB reporter plasmid without separating CEN and STB into separate circles. (B) Assay conditions are the same as in (A); however, topologies are probed after resolving the parent plasmid into the STB circle and the CEN circle. The + and − signs against Rep1, Rep2 indicate the functional and non-functional states of STB respectively. The + and − symbols against Ndc10 denote the functional and non-functional states of CEN respectively.

The sum of the topological assays (Figures 2 and 3) suggests that the in vivo topology of CEN and STB are quite similar, qualitatively and quantitatively. The occupancy of a Cse4-containing nucleosome with a positive DNA writhe at a functional STB and its replacement by an H3-containing nucleosome with a negative writhe would account for the observed ΔLk of nearly 2. This interpretation would provide a unified explanation for CEN and STB topologies conferred by a specialized nucleosome with an unusual DNA topology. However, there is a caveat that Cse4 is present at STB in less than stoichiometric amounts. Thus, whereas the topology of the functional STB may be dependent on Cse4, a Cse4-containing nucleosome may not be directly responsible for this topology. In the light of the biochemical and structural evidence that a Cse4-containing nucleosome does not preclude a left-handed DNA wrap around it, additional experimental verification is required to authenticate the nucleosome topology at CEN.

CEN and STB: evolutionarily linked partitioning loci?

As pointed out, the point centromere signifies an unexpected and quirky transition during centromere evolution [26]. The switch from the regional to point centromere appears to have occurred contemporaneously with the partial or near-complete loss of the machinery required for establishing heterochromatin in the budding yeast lineage [40]. Furthermore, plasmids similar in organization to the 2μm plasmid have been detected only among a limited number of fungal species belonging to this lineage [41]. These circumstantial pieces of evidence suggest the possibility that the loss of the capacity to establish a functional regional centromere was compensated for by the domestication of the plasmid partitioning locus as the central DNA element for directing chromosome segregation as well [26]. The present-day spindle-dependent chromosome segregation and chromosome-coupled plasmid segregation pathways may have been arrived at by divergent evolution from a common ancestral pathway. The conserved positive DNA topology at CEN and STB adds another tantalizing piece of evidence to a growing list that suggests a common origin for the partitioning locus of a benign parasite genome and that of the chromosomes of the host on which its survival depends.

Funding

The work presented was supported by the National Science Foundation [grant number MCB-1049925] and by the Robert F Welch Foundation [grant number F-1274].

Footnotes

  • Topological Aspects of DNA Function and Protein Folding: An Independent Meeting held at the Isaac Newton Institute for Mathematical Sciences, Cambridge, U.K., 3–7 September 2012, as part of the Isaac Newton Institute Programme Topological Dynamics in the Physical and Biological Sciences (16 July–21 December 2012). Organized and Edited by Andrew Bates (University of Liverpool, U.K.), Dorothy Buck (Imperial College London, U.K.), Sarah Harris (University of Leeds, U.K.), Andrzej Stasiak (University of Lausanne, Switzerland) and De Witt Sumners (Florida State University, U.S.A.).

References

View Abstract