Skip Navigation

This Article
Right arrow Abstract Freely available
Right arrow Print PDF (151K) Freely available
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (22)
Right arrowRequest Permissions
Right arrow Commercial Re-use Guidelines
for Open Access NAR Content
Google Scholar
Right arrow Articles by Prescott, D. M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Prescott, D. M.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

Nucleic Acids Research Pages 1243-1250  

The evolutionary scrambling and developmental unscrambling of germline genes in hypotrichous ciliates
Introduction
IESs And MDSs
Scrambled Micronuclear Genes
   The actin I gene
   The scrambled [alpha]TP gene
   A model of the origin of non-random scrambling
   The DNA pol [alpha] gene
Origin Of Random Scrambling
Conversion Of A Micronucleus Into Amacronucleus
Acknowledgements
References

The evolutionary scrambling and developmental unscrambling of germline genes in hypotrichous ciliates

The evolutionary scrambling and developmental unscrambling of germline genes in hypotrichous ciliates

David M. Prescott*

Department of Molecular, Cellular and Developmental Biology, University of Colorado, Boulder, CO 80309-0347, USA

Received July 17, 1998; Revised and Accepted November 20, 1998

ABSTRACT

Genes in the germline (micronuclear) genome of hypotrichous ciliates are interrupted by multiple, short, non-coding, AT-rich sequences called internal eliminated segments, or IESs. During conversion of a micronucleus to a somatic nucleus (macronucleus) after cell mating, all IESs are excised from the germline genes and the gene segments, called macronuclear-destined segments, or MDSs, are spliced. Excision of the ~150 000 IESs from a haploid germline genome in Oxytricha nova requires ~150 000 recombinant events. In three of 10 genes the MDSs are scrambled. During macronuclear development the MDSs are unscrambled, possibly by folding of the DNA to allow MDSs to ligate in the correct order. The nine MDSs in the actin I gene of O.nova are scrambled in the random order, 3-4-6-5-7-9-2-1-8, and MDS 2 is inverted. The 14 MDSs in the [alpha]TP gene of O.nova and Stylonychia mytilus are scrambled in the non-random order, 1-3-5-7-9-11-2-4-6-8-10-12-13-14. The 45 MDSs in the DNA pol [alpha] gene are non-randomly scrambled into an odd/even series, with an inversion of one-third of the gene. Additional IESs have been inserted into these three genes during evolution of Oxytricha trifallax, slightly modifying scrambling patterns. The non-random scrambled patterns in the [alpha]TP and DNA pol [alpha] genes are explained by multiple, simultaneous IES insertions. The randomly scrambled pattern in the actin I gene may arise from an initially non-randomly scrambled pattern by recombination among multiple IESs.Alternatively, IESs inserted sporadically (individually) in a non-scrambled configuration might subsequently recombine, converting a non-scrambled gene into a randomly scrambled one. IESs shift along a DNA molecule, most likely as a result of mutations at MDS/IES junctions. Shifting of IESs has the effect of ‘transferring’ nucleotides from one MDS to another, but does not change the overall sequence of nucleotides in the combined MDSs. In addition to shifting in position, IESs accumulate mutations at a high rate and increase and decrease in length within a species and during speciation. The phenomena of IESs and of MDS scrambling represent remarkable flexibility of the hypotrich genome, possibly reflecting a process of MDS shuffling that facilitates the evolution of genes.

INTRODUCTION

The DNA in hypotrichous ciliates has undergone extraordinary modifications in organization and behavior during evolution. The story of these DNA modifications is far from complete and what they mean for the evolution and genetic operation of ciliates is still not clear. However, at the very least, observations thus far attest to a remarkable flexibility in the organization and processing of DNA in the evolution and function of hypotrich genomes.

IESs AND MDSs

The germline genes in the micronucleus of hypotrichous ciliates are interrupted by multiple, non-coding sequences called internal eliminated segments, or IESs (1). In the example shown in Figure 1a, the germline gene that encodes [beta] telomere-binding protein ([beta]TP) in Oxytricha trifallax is split into seven parts by six IESs, one in the 5[prime]-non-translated leader region and five in the open reading frame (ORF). The gene sections separated by IESs are called macronuclear-destined segments, or MDSs. IESs resemble introns in the sense that IESs are non-coding and interrupt genes, but in other fundamental respects IESs and introns are very different phenomena. Unlike introns, which are removed from RNA transcripts, IES sequences are spliced out of DNA during conversion of a micronucleus to a macronucleus after cell mating. After IES excision, the genes are individually released from their chromosomal locations as separate, short DNA molecules and the chromosomal spacer DNA between genes is degraded. Each of the macronuclear DNA molecules consists of a 5[prime]-non-translated leader segment, an ORF and a 3[prime]-non-translated trailer (Fig. 1b). These gene-sized molecules are capped at their ends with 4.5 copies (usually) of the telomeric repeat sequence 3[prime]-G4T4-5[prime].


Figure 1. (a) Diagram of the micronuclear gene encoding [beta]TP in O.trifallax. Six IESs (lines) separate seven gene segments (clear boxes labeled MDS 1-7). The black box is an intron. (b) Diagram of the macronuclear DNA molecule encoding [beta]TP. The vertical lines in the clear box indicate where MDSs have been spliced after IES removal. The striped boxes at the ends are telomeres, with 3[prime] 16 base overhangs. (c) IES 3 (lower case letters) and its junctions with MDSs 3 and 4 (upper case letters) in the micronuclear [beta]TP gene of O.trifallax. The 4 bp repeat sequence (ACTC) is underlined. From Prescott and DuBois (7).

Splitting of micronuclear genes by IESs has no consequence for gene expression because micronuclear genes are transcriptionally silent. During macronuclear development the new somatic genes are activated. Although removal of IESs, excision of genes from chromosomes and activation of genes are closely related temporally, there is no evidence to suggest that these events are related mechanistically. Some progress has been made in understanding the molecular mechanism of IES excision (2,3) and gene excision/telomere addition (4-6), but nothing is known about the molecular mechanism that maintains the silence of the micronuclear genome.

IESs are randomly distributed in micronuclear genes, occurring in the 5[prime] leader, the ORF, and/or the 3[prime] trailer. IESs are generally short; >95% are <100 bp (8). The shortest IESs detected so far in hypotrichs are 5 bp in length (9), while the longest is 595 bp (8). IESs are AT-rich (75-100% AT), unique sequences with no consistent distinguishing structural characteristics, e.g. palindromes or consensus sequences. Pairs of 2 to 7 bp repeat sequences that are part of the MDSs (Fig. 1c) are present at the junctions of an IES with its adjacent MDSs. These repeats probably help direct MDS splicing. IESs are excised during macronuclear development in Euplotes crassus by staggered cuts at the two IES/MDS junctions such that the IES and one copy of the repeat pair are removed (2). The other copy of the repeat remains at the splice site between the two adjacent MDSs.

SCRAMBLED MICRONUCLEAR GENES

The structures of 10 different micronuclear genes and their macronuclear counterparts have been completely characterized in various species of Oxytricha and Stylonychia (1,9,7,-17). In three of these genes, encoding actin I, [alpha] telomere-binding protein ([alpha]TP) and DNA polymerase [alpha] (DNA pol [alpha]), multiple MDSs have been rearranged into a scrambled order during evolution.

The actin I gene

The actin I gene in Oxytricha nova (Fig. 2f) consists of eight IESs and nine MDSs in the order 3-4-6-5-7-9-2-1-8, with MDS 2 inverted (15). MDSs 3 and 4 are not scrambled relative to each other, but the other seven MDSs are scrambled in largely random disorder. How did these MDSs become scrambled? One possibility is the relatively simple model in Figure 2. The first step consists of insertion of eight IESs simultaneously or separately in time into the actin I gene (Fig. 2b). This creates nine MDSs in the non-scrambled order 1-2-3-4-5-6-7-8-9. Subsequently, four hypothetical intragenic recombinations among the IESs shown in Figure 2 create the contemporary scrambled state of the actin I germline gene. These postulated scrambling events are not necessarily intragenic as pictured in the model in Figure 2; both alleles in the micronucleus may have participated.


Figure 2. A model describing the origin of scrambling of the actin I gene in O.nova. (a) The micronuclear gene prior to invasion by IESs. (b) Insertion of eight IESs, creating nine non-scrambled MDSs. (c) Inversion and translocation of MDS 2 to the 5[prime]-end of the gene. (d) Translocation of MDSs 2 and 1 to the 3[prime]-end of the gene. (e) Translocation of MDS 8 to the 3[prime]-end of the gene. (f) Translocation of MDS 6 to a position between MDSs 4 and 5, producing the contemporary arrangement of MDSs in the actin I gene. MDSs are clear blocks and IESs are lines between MDSs.

Scrambled genes continue to undergo evolutionary change. This is affirmed by comparing the structures of particular genes among different species. The structural arrangements of MDSs and IESs in the complete actin I gene in two species, O.nova (15) and O.trifallax (11), and in part of the actin I gene in Oxytricha sp.? (Aspen) (8), are similar, but there are several major differences (Fig. 3). Because of rapid accumulation of mutations, corresponding IESs from species to species share no sequence identity and differ considerably in length. In addition, positions of IESs within the actin I gene are shifted by one to many base pairs from species to species. Shifting of IES positions is explained by a mutational model (11) in which nucleotides at MDS/IES junctions are ‘shifted’ one or two at a time between an MDS and an IES. This phenomenon is illustrated in Figure 4a with the junctions between a hypothetical IES and its flanking MDSs. A mutation at the right junction shortens the IES by one base and a single mutation at the left junction lengthens the IES by two bases and shortens the MDS by two bases. As the junctions shift the sequence of the pair of repeats changes, but the coding sequence generated by splicing the MDSs during macronuclear development remains the same. This mutational model of IES shifting is applied in Figure 4b to the junction of MDS 4 withIES 2 and the junction of IES 3 with MDS 5 in the scrambled actin I gene of O.nova, O.trifallax and Oxytricha sp.? (Aspen) (Fig. 4c). The shifting of IES/MDS junctions has major structural con-sequences for the micronuclear actin I gene. The sequences of the repeats are progressively changed, IESs shift positions along the DNA molecule and MDSs are lengthened or shortened. In Figure 4b MDS 4 is shortened at its right end and MDS 5 is lengthened at its left end, yet the coding sequence of the gene does not change. The actin I gene in O.trifallax also differs by the presence of an additional 276 bp IES (IES 6) (Fig. 3). This IES was likely inserted into the original MDS 9 sometime after the original scrambling event, creating MDSs 9 and 10 in O.trifallax. Thus, the actin I gene likely became scrambled in a common ancestor of O.nova, O.trifallax and Oxytricha sp.? (Aspen) and alterations in gene structure have continued in the derivative species.


Figure 3. The scrambled micronuclear actin I gene in O.nova, O.trifallax and Oxytricha sp.? (Aspen). The three genes are scrambled in similar, largely random patterns. MDS 2 is inverted ([larr]) in all three species. The O.trifallax gene contains an additional IES and MDS. MDSs are clear blocks. IESs are lines between MDSs. TAS, telomere addition site. From Prescott and DuBois (7).


Figure 4. (a) A model for mutational shifting of MDS/IES junctions. MDS sequences are in upper case letters and the IES is in lower case letters. Repeat sequences are underlined and mutations are indicated with small arrows. A single base change in the IES a->C at the right junction lengthens the MDS and lengthens the repeat sequence from 7 to 8 nt. A second mutation of C->a at the left junction shortens the MDS and shortens the repeat sequence from 8 to 6 nt, shifting the MDS/IES junction by 2 bp towards the 5[prime]-end of the gene. Modified from Prescott and DuBois (7). (b) Actual change in sequences at the MDS 4/IES 2 and IES 3/MDS 5 junctions in O.nova, O.trifallax and Oxytricha sp.? (Aspen). A hypothetical series of mutations in the actin I gene of O.nova shows how a single mutation subtracts two bases from MDS 4 and adds them to the left end of IES 2 in O.trifallax. At the right end of IES 3 in O.nova two mutations lengthen MDS 5 by two bases and shorten the IES by two bases in O.trifallax. Additional hypothetical mutations show how the sequences at junctions in O.trifallax could progress to the junction sequences in Oxytricha sp.? (Aspen). The sequences of the repeat pairs progressively change and the lengths of the MDSs change as the junctions shift. An overall effect is that coding sequence is ‘transferred’ from MDS 4 to MDS 5, a process that does not change the coding sequence. Removal of the IES during macronuclear development yields the same sequence in all three species. All of the mutational changes could have occurred in the opposite direction and in a different order among the organisms. The O.nova data are from Greslin et al. (25), the O.trifallax data are from DuBois and Prescott (11) and the Oxytricha sp.? (Aspen) data are from Prescott and DuBois (7).

The scrambled [alpha]TP gene

Unlike the actin I gene, the scrambled pattern of the gene encoding [alpha]TP is non-random. The micronuclear [alpha]TP gene consists of 14 MDSs in O.nova (13) and in Stylonychia mytilus (9) in the non-random order 1-3-5-7-9-11-2-4-6-8-10-12-13-14 (Fig. 5a). The gene in O.trifallax contains 17 MDSs in the order 1-3-5-7-10-12-2-4-6-8-9-11-13-14-15-16-17. (An intron interrupts MDS 2 between codons 35 and 36 in all three species.) From the similarity of the scrambled patterns, it is evident that the [alpha]TP gene became scrambled in an ancestor of the three species in the pattern currently shared by O.nova and S.mytilus. The pattern in O.trifallax is somewhat different and very likely resulted from insertions of IESs 10, 14 and 15 after divergence of O.trifallax from O.nova and S.mytilus. IES 10 in O.trifallax has been inserted into the original MDS 8, dividing it into MDSs 8 and 9. As a result, MDS and IES numberings change in O.trifallax for all MDSs and IESs that numerically follow MDS 8 andIES 10. IESs 14 and 15 were likely inserted into the original MDS 13, creating MDSs 14-16 in O.trifallax. Unscrambling and ligation of MDSs produces three very similar macronuclear genes (Fig. 5b).


Figure 5. (a) The non-randomly scrambled micronuclear [alpha]TP genes in three hypotrichs. MDSs are clear boxes and numbered above each gene. IESs are lines connecting MDSs and numbered below each gene. Introns are black boxes. (b) Macronuclear versions of the three genes shown in (a) after excision of IESs and reordering and splicing of MDSs during macronuclear development. Vertical lines mark MDS splice sites. Telomeric sequences are striped boxes with 3[prime] overhangs at the ends of genes. Data for O.nova are from Mitcham et al. (13). Data for O.trifallax and S.mytilus are from Prescott et al. (8).

Unlike the actin I gene, none of the MDSs in [alpha]TP are inverted and most of the MDSs in the [alpha]TP gene are distributed in a strikingly non-random, odd/even pattern. In O.nova and S.mytilus the exceptions are the non-scrambled MDSs 12-14 and in O.trifallax the exceptions are the non-scrambled (relative to one another) MDSs 8 and 9 and non-scrambled MDSs 12-15. Like the actin I gene, IESs that correspond from species to species are completely different in sequence, differ considerably in length and have shifted positions along the DNA. As a consequence, the lengths of MDSs, and therefore the distribution of coding sequence among the MDSs, are changed, but these MDS changes do not alter the coding sequence. This is illustrated in Figure 4 with MDSs 4 and 5 in the actin I gene in three species of Oxytricha. Although MDSs 4 and 5 differ in length from species to species, removal of the IES and one copy of the repeat during macronuclear development yields the same sequence in all three species.

A model of the origin of non-random scrambling

The odd/even pattern of MDSs in the [alpha]TP gene implies that scrambling most likely took place in a single event in an ancestor common to the three species. A model of this event is shown in Figure 6, beginning with an [alpha]TP gene that lacks IESs. Alignment of the gene with a closed or open loop of AT-rich, non-coding DNA (most of the germline genome is made up of such sequence), followed by multiple, simultaneous recombinations between the [alpha]TP gene and the AT-rich loop (Fig. 6c), creates an odd/even series of scrambled MDSs 1-12. In the model the pairs of repeats are created essentially as described previously (8) for a non-scrambled gene with IESs. First, MDSs are formed by staggered cuts in the [alpha]TP gene in Figure 6a, as illustrated for MDSs 2-4 in Figure 6b. Pieces of the AT-rich loop immediately ligate to the single-stranded tails. Filling the resulting single-stranded gaps creates the pairs of repeat sequences (R1/R1, R2/R2, etc.) in the scrambled gene in Figure 6b.

IESs between the non-scrambled MDSs 12-14 (Fig. 6c) probably arose by separate, unrelated insertions of AT-rich DNA in the ancestor of the three species either before or after scrambling of MDSs 1-12. After divergence of O.trifallax from O.nova and S.mytilus, a single IES was inserted into the original MDS 8 of O.trifallax, dividing it into MDSs 8 and 9, slightly disrupting the odd/even MDS pattern (Figs 5a and 6d). Two more IESs were inserted in the O.trifallax gene into the original MDS 13 to create a total of five non-scrambled MDSs in the end of the O.trifallax version. These observations on the structural arrangements of the [alpha]TP gene in the three species strongly support two primary conclusions: (i) simultaneous invasion of a gene by multiple IESs in a regularized series of recombinations can create a non-random pattern of MDSs; (ii) single IES insertions can occur in previously established MDSs. Thus, IES insertions can be concerted (inserted as a group at one time as one event) or sporadic (inserted individually at different times). This distinction implies two, somewhat different mechanisms for IES invasion.


Figure 6. A hypothesis to account for the origin of the non-randomly scrambled structure of the micronuclear [alpha]TP genes in O.nova, S.mytilus and O.trifallax. (a) A loop of AT-rich DNA is aligned with an [alpha]TP gene that contains no IESs. (b) Staggered cuts are made in the [alpha]TP gene, followed by recombination of sections of AT-rich DNA with the single-stranded overhangs resulting from staggered cuts. Fill-in of single-strand gaps to create repeat pairs (R1, R2, etc.) (c) Insertion of multiple IESs at staggered cuts in the [alpha]TP gene (as shown in b) creates 11 IESs separating MDSs 1-12. MDSs 13 and 14 are created by separate insertions. (d) The regular odd/even pattern of MDSs in O.nova and S.mytilus evolve into the O.trifallax pattern by insertion of three additional IESs, dividing MDS 8 into MDSs 8 and 9 and dividing MDS 13 into MDSs 14-16. MDS 14 in O.nova/S.mytilus becomes MDS 17 in O.trifallax. From Prescott et al. (8).

The DNA pol [alpha] gene

The gene encoding DNA pol [alpha] represents a third pattern of scrambling, one that is more complex than those in the actin I or the [alpha]TP genes. The DNA pol [alpha] gene consists of 45 MDSs and 44 IESs in O.nova (13) and 51 MDSs and 50 IESs in O.trifallax (Fig. 7; 18). Almost all of the MDSs are arranged in a non-random, odd/even pattern that is disrupted by inversion of approximately one-third of the gene with respect to the other two-thirds in both species. Like the actin I and [alpha]TP genes, a few MDSs (e.g. MDSs 26 and 27, MDSs 44 and 45 and MDSs 1-3 in the DNA pol [alpha] gene of O.nova) are not scrambled relative to one another. Essentially the same model for the origin of non-random scrambling of MDSs in the [alpha]TP gene in Figure 6 also accounts for the non-random, odd/even pattern of MDSs in the DNA pol [alpha] gene. The final step in scrambling in O. nova is recombination between the IESs flanking MDSs 1 and 4 to create the large inversion. The second inversion point is in DNA flanking one or the other end of the gene. As in the [alpha]TP and actin I genes, the non-scrambled MDSs probably originated by insertion of individual IESs in unrelated events (sporadic insertions) after gene scrambling, e.g. splitting of an original MDS 26 in the DNA pol [alpha] gene in O.nova (Fig. 7) into non-scrambled MDSs 26 and 27 by insertion of a new IES.


Figure 7. Diagrams of the scrambled DNA pol [alpha] genes in O.trifallax and O.nova. The genes continue on two lines. MDSs are clear boxes and IESs are lines between MDSs. MDSs 32 and 51 of the O.trifallax gene are drawn as open-ended boxes because the sequences at the 5[prime]- and 3[prime]-ends have not been determined. MDSs 29, 31, 33, 35, 37, 39, 41 and 43 in the O.nova gene and MDSs 34, 36, 38, 40, 42, 44, 46 and 48 are not present with the main body of the gene. An inversion is present in the same location within the gene in both species. The second inversion point is not known but must be present in the DNA that flanks one or the other end of the micronuclear gene. A suggested location in the DNA flanking MDS 43 in O.nova is shown in Figure 10. From Hoffman and Prescott (12).

The six additional MDSs in the DNA pol [alpha] gene of O.trifallax are most parsimoniously explained by insertion of the six IESs after evolutionary divergence between O.nova and O.trifallax. Four new MDSs were created in O.trifallax, after the initial, non-random scrambling of the DNA pol [alpha] gene, by a mechanism that preserved the odd/even pattern of MDSs in that region of the gene. In a model of this mechanism (Fig. 8; 12), the original MDS 6, which is still present as MDS 6 in the DNA pol [alpha] gene of O.nova, was split into MDSs 6a-6c in O.trifallax by the insertion of two new IESs. Because of the inversion in the gene between MDSs 1 and 4, the molecule can be folded into a hairpin such that the MDSs in each branch of the hairpin face in the same direction with respect to transcription (arrows in Fig. 8). Recombination between the two new IESs and the IES between original MDSs 5 and 7 translocates new MDS 6b to a position between MDSs 5 and 7. In this way, two IESs and two MDSs are added to the O.trifallax gene while maintaining the regular odd/even scrambled pattern of MDSs. However, the additions change the numbering of MDSs and IESs (Figs 7 and 8). By the same scheme, two more scrambled MDSs were apparently created in O.trifallax without disrupting the odd/even MDS pattern via insertion of two IESs into original MDS 10 of O.nova, creating MDSs 10a-10c. Recombination then inserted MDS 10b between MDSs 9 and 11. Finally, an IES appears to have been inserted into MDS 3 to create non-scrambled MDSs 3 and 4 and an IES inserted into MDS 45 of O.nova resulted in MDSs 50 and 51 in O.trifallax. These changes increase the total number of IESs and MDSs in O.trifallax by six of each.


Figure 8. A model for the introduction of additional non-randomly scrambled MDSs into the DNA pol [alpha] gene in O. trifallax. An inversion is present between MDSs 1 and 4, so the molecule is drawn as a hairpin to put MDSs in the same right-left direction with respect to transcription. Arrows indicate the 5[prime]->3[prime] direction of the coding strand. (A) Two new IESs are inserted into original MDS 6 (carets). (B) MDS 6 has been divided into MDSs 6a-6c. (C) Two recombinations between the two new IESs and the IES between MDSs 5 and 7 inserts MDS 6b between MDSs 5 and 7. (D) MDS 6a becomes MDS 6, MDS 6c becomes MDS 8, and MDS 6b becomes MDS 7, maintaining the odd/even pattern of MDSs of the O.trifallax gene. From Hoffman and Prescott (12).

A prerequisite for the model in Figure 8 is folding of the DNA pol [alpha] gene into a hairpin, as shown in Figure 10. Such folding might be prescribed by alignment of the homologous sequences in the pairs of repeats at the ends of numerically consecutive MDSs. This places the newly created MDSs in position for recombination that transfers a new MDS (MDS 6b in Fig. 8) to the opposite branch of the hairpin.

Comparison of DNA pol [alpha] gene structure between O.nova and O.trifallax demonstrates that this gene has undergone the same type of evolutionary changes that occurred in the actin I and [alpha]TP genes. Corresponding IESs are completely different in sequence, differ substantially in length and have shifted along the DNA molecule. IES shifting has changed the lengths of corresponding MDSs, in some cases by >50 bp. Last, six more IESs have been inserted into the O.trifallax gene. Thus, there is a consistent pattern of evolutionary change in all three scrambled genes.

Oxytricha trifallax contains one, three and six more IESs in the actin I, [alpha]TP and DNA pol [alpha] genes, respectively, than does O.nova. Similarly, the non-scrambled genes encoding [beta]TP and histone H4 (ORF only) contain six and one IESs, respectively, in O.trifallax compared with three and none in O.nova. The greater number of IESs in five out of the five genes characterized so far implies that O.trifallax (82 IESs) has a greater propensity for insertion of IESs than does O.nova (68 IESs).

Scrambling of the DNA pol [alpha] gene presents another complexity. MDSs 29-31-33-35-37-39-41-43, encoding 230 bp of the 4479 bp in the ORF in O.nova, became separated from the main body of the gene after scrambling had taken place. These missing MDSs are not present in the first several kilobases flanking either end of the gene and their chromosomal location remains unknown. This same portion of the gene, consisting of 193 bp in MDSs 34-36-38-40-42-44, is also missing from the main body of the gene in O.trifallax. Translocation of this segment must have occurred in an ancestor common to O.nova and O.trifallax after the scrambling event that created the odd/even pattern. A large inversion could have displaced the missing MDSs more than several kilobases into the DNA flanking either end of the gene. Most of these missing MDSs in O.trifallax have been identified by PCR as a group in the genome, but their chromosomal location relative to the bulk of the gene is still unknown. The difference in the amount of translocated coding sequence (230 versus 193 bp) can be accounted for by IES shifting subsequent to scrambling. IES shifting has had the effect of ‘transferring’ coding sequence from the translocated MDSs to MDSs in the main body of the gene or vice versa, as illustrated by the ‘transfer’ of coding sequence between MDS 4 and 5 in the actin I gene in Figure 4.

ORIGIN OF RANDOM SCRAMBLING

IESs may have been originally inserted into the actin I gene in sporadic fashion, without producing scrambling. This must have happened in the non-scrambled [beta]TP, C2 and R1 genes of O.nova, the [beta]TP gene of O.trifallax and S.mytilus and the MA83s-A gene of Oxytricha fallax. Random scrambling of the actin I gene may have subsequently been produced by intragenic IES recombination as already described in Figure 2. Alternatively, the actin I gene may have been originally scrambled in a non-random pattern by the concerted mechanism illustrated for the [alpha]TP gene in Figure 6. Subsequently, the gene evolved to a random pattern as described in Figure 9. By a series of three transpositions, inversion of MDS 2 and insertion of an additional IES, an original odd/even pattern of 1-3-5-7-2-4-6-8 could have been converted to the current pattern of 3-4-6-5-7-9-2-1-8 in the actin I gene. Analysis of the micronuclear actin I gene in other hypotrich species might provide a test of this model of how the scrambled state of this gene in O.nova evolved.


Figure 9. Possible origin of the largely random, scrambled arrangement of MDSs in the actin I gene of O.nova. (a) The gene was originally scrambled in an odd/even pattern by the mechanism shown in Figure 6. (b) Recombination 1 between IES 3 and flanking DNA displaces MDSs 7 and 2 to the right of MDS 8. (c) Recombination 2 inverts MDS 2 and inserts it between MDSs 8 and 7. (d) Recombination 3 inserts MDS 1 between MDSs 2 and 7. (e) Insertion of a new IES in MDS 3 splitsMDS 3 into MDS 3 and a new MDS 4. Old MDSs 5, 4, 6, 8 and 7 are now renumbered as 6, 5, 7, 9 and 8 to yield the current structure of the actin I gene in O.nova.


Figure 10. Folding of the DNA pol [alpha] gene of O. nova to align pairs of repeat sequences that flank IESs. MDSs are heavy lines and IESs are thin lines. A repeat pair is present wherever MDSs in the two branches of the hairpin overlap at their ends. Recombination between each pair of repeats splices MDSs in the unscrambled order, removes the IESs and removes one copy of the repeat in each pair. From Hoffman and Prescott (12). An inversion point is present between MDSs 1 and 4. The other end of the inversion is shown in DNA flanking MDS 43, but theoretically it could be in flanking DNA at the other end of the gene.

CONVERSION OF A MICRONUCLEUS INTO AMACRONUCLEUS

Extrapolation from the number of IESs in seven micronuclear genes in O.nova gives an estimate of ~150 000 IESs/haploid genome. Removal of these IESs and assembly of the macronuclear genome during development from micronuclear DNA requires four types of processes (19). (i) Removal of the IESs via ~150 000 recombination events. Since IES excision occurs during the polytene chromosome stage of macronuclear development, the number of recombinational excisions of IESs is much greater than 150 000, potentially as great as ~107. (ii) In at least some hypotrichs, one or more large families of transposon-like elements are also excised and destroyed during the polytene chromosome stage (20-22). (iii) In association with IES excision, scrambled MDSs are unscrambled and ligated. (iv) Spacer DNA between genes in micronuclear chromosomes is selectively destroyed, releasing each gene as an individual, short DNA molecule to which telomeric sequences are added.

Little is known about how these DNA processing events are directed. IES removal requires that the excision mechanism must be able to identify the two ends of the IES. The mechanism must be able to cut the DNA at the MDS/IES junctions and splice the MDSs. The 2-7 bp repeats flanking IESs that separate non-scrambled MDSs (Fig. 1) contain inadequate information for IES recognition. No other discernible, consistent structural features of IESs or MDS/IES junctions in hypotrichs have been observed. In contrast, repeats that flank IESs between scrambled MDSs are longer (average 11 bp) than those between non-scrambled MDSs (average 4 bp). Theoretically, the longer repeats are sufficient to guide folding of the DNA molecule directed by alignment of repeat pairs [as proposed for unscrambling of the actin I, [alpha]TP, and DNA pol [alpha] (Fig. 10) genes]. In this case, homologous recombination between the two aligned repeats in a pair eliminates the IES plus one copy of the repeat and splices MDSs in the orthodox order. In addition to solving the unscrambling problem, the folding model may also eliminate the need for specific recognition of IESs. IES removal simply becomes a by-product of MDS splicing. However, the stiffness of the DNA duplex must be reduced, perhaps by proteins, because most of the IESs and many of the MDSs are too short to fold without assistance. The folding model may not apply in non-scrambled situations because the repeats are too short to guide folding unambiguously. This is illustrated by the 292 bp IES 6 in the [beta]TP gene of O.trifallax (Fig. 1a). IES 6 is flanked by the 3 bp repeat AGT, a trinucleotide occurring five times within the IES, and at 5, 81 and 91 bp downstream of the IES and 44 bp upstream. Additional information of unknown form and source is required to direct the excision mechanism to the correct trinucleotides. Last, the difference in repeat lengths between scrambled and non-scrambled MDSs suggests that the two configurations of MDSs and IESs might be processed by fundamentally different mechanisms.

An alternative to the folding model is suggested by observations on IES removal in Paramecium. In this case the old macronucleus appears to transmit sequence information to the developing macronucleus, perhaps effecting IES excisions by a template mechanism (23). In hypotrichs, DNA molecules transmitted from the old macronucleus to the developing macronucleus could solve three related requirements in DNA processing: IES excision, unscrambling of MDSs and excision of genes from chromosomes (although no template mechanism for these three processing events is easily visualized).

The presence of IESs and scrambled MDSs in hypotrich genes raises many questions. Did IESs originate as transposons that then degenerated to their current simple structure, as proposed for the IESs in Euplotes crassus and in Paramecium (24)?

However, contemporary IESs in Oxytricha and Stylonychia species bear virtually no resemblance to any known transposons, which means that IES excision is different from known mechanisms of transposon excision. IESs are markedly different from transposons in structure and they also accumulate mutations at a high rate, completely changing in sequence over time. IESs change in size, and shifting along DNA alters the sequences of the short, flanking repeats. Thus, the IES excision mechanism is vastly different from what is known about transposon excision.

Hypotrichous ciliates are very successful, ubiquitous eukaryotes, thriving in virtually every body of water on Earth. Whether the splitting of germline genes into MDSs by IESs and whether MDS scrambling are innovations that somehow contribute in any way to this success is not obvious. The presence of IESs is innocuous since germline genes are genetically silent and because all IESs are removed when a macronucleus is formed. MDS scrambling, which serves no discernible purpose, is made possible by the presence of IESs. Scrambling is tolerable because its resolution is an integral part of IES removal. At the very least, the phenomena of IESs and of MDS scrambling demonstrate a remarkable flexibility in the organization and behavior of hypotrich DNA sequences. MDS scrambling may, in fact, be only a surface reflection of a more profound process in which MDS shuffling facilitates and accelerates the evolution of genes.

ACKNOWLEDGEMENTS

This work was supported by NIGMS Research Grant GM56161 to D.M.P. The author is grateful to Gayle Prescott for preparation of the manuscript and Jason Prescott for preparation of figures.

REFERENCES

1. Klobutcher,L.A., Jahn,C.L. and Prescott,D.M. (1984) Cell, 36, 1045-1055. MEDLINE Abstract

2. Klobutcher,L.A., Turner,L.R. and LaPlante,J. (1993) Genes Dev., 7, 84-94. MEDLINE Abstract

3. Klobutcher,L.A., Gygax,S., Podoloff,J., Vermeesch,J.R., Price,C.M., Tebeau,C.M. and Jahn,C.L. (1998) Nucleic Acids Res., 26, 4230-4240. MEDLINE Abstract

4. Roth,M. and Prescott,D.M. (1985) Cell, 41, 411-417. MEDLINE Abstract

5. Klobutcher,L.A., Huff,M.E. and Goyne,G.E. (1988) Nucleic Acids Res., 16, 251-264. MEDLINE Abstract

6. Vermeesh,J.R. and Price,C.M. (1994) Mol. Cell. Biol., 14, 554-566. MEDLINE Abstract

7. Prescott,D.M. and DuBois,M.L. (1996) J. Eukaryot. Microbiol., 43, 432-441. MEDLINE Abstract

8. Prescott,J.D., DuBois,M.L. and Prescott,D.M. (1998) Chromosoma, 107, 293-303.DuBois,M.L. and Prescott,D.M. (1997) Mol. Cell. Biol., 17, 326-337. MEDLINE Abstract

10. Bierbaum,P., Dönhoff,T. and Klein,A. (1991) Mol. Microbiol., 5, 1567-1575. MEDLINE Abstract

11. DuBois,M. and Prescott,D.M. (1995) Proc. Natl Acad. Sci. USA, 92, 3888-3892. MEDLINE Abstract

12. Hoffman,D.C. and Prescott,D.M. (1997) Nucleic Acids Res., 25, 1883-1889. MEDLINE Abstract

13. Mitcham,J.L., Lynn,A.J. and Prescott,D.M. (1992) Genes Dev., 6, 788-800. MEDLINE Abstract

14. Mitcham,J.L., Prescott,D.M. and Miller,M.K. (1994) J. Eukaryot. Microbiol., 41, 478-480. MEDLINE Abstract

15. Prescott,D.M. and Greslin,A.F. (1992) Dev. Genet., 13, 66-74. MEDLINE Abstract

16. Ribas-Aparicio,R.M., Sparkowski,J.J., Proulx,A.E., Mitchell,J.D. and Klobutcher,L.A. (1987) Genes Dev., 1, 323-336. MEDLINE Abstract

17. Williams,K.R. and Herrick,G. (1991) Nucleic Acids Res., 19, 4717-4724. MEDLINE Abstract

18. Hoffman,D.C. and Prescott,D.M. (1996) Nucleic Acids Res., 24, 3337-3340. MEDLINE Abstract

19. Prescott,D.M. (1994) Microbiol. Rev., 58, 233-267. MEDLINE Abstract

20. Jahn,C.L., Krikau,M.F. and Shyman,S. (1989) Cell, 59, 1009-1018. MEDLINE Abstract

21. Krikau,M.F. and Jahn,C.L. (1991) Mol. Cell. Biol., 11, 4751-4759. MEDLINE Abstract

22. Tausta,S.L., Turner,L.R., Buckley,L.K. and Klobutcher,L.A. (1991) Nucleic Acids Res., 19, 3229-3236. MEDLINE Abstract

23. Duharcourt,S., Keller,A.-M. and Meyer,E. (1998) Mol. Cell. Biol., 18, 7075-7085. MEDLINE Abstract

24. Klobutcher,L.A. and Herrick,G. (1997) Prog. Nucleic Acid Res. Mol. Biol., 56, 1-62. MEDLINE Abstract

25. Greslin,A.F., Prescott,D.M., Oka,Y., Loukin,S.H. and Chappell,J.C. (1989) Proc. Natl Acad. Sci. USA, 86, 6264-6268. MEDLINE Abstract

*Tel: +1 303 492 8381; Fax: +1 303 492 7744; Email: prescotd@spot.colorado.edu

This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 11 Feb 1999
Copyright©Oxford University Press, 1999.
Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 Plant CellHome page
S. Henikoff
Rapid Changes in Plant Genomes
PLANT CELL, November 1, 2005; 17(11): 2852 - 2855.
[Full Text] [PDF]

Home page
 Proc. Natl. Acad. Sci. USAHome page
A. Maor-Shoshani, V. Ben-Ari, and Z. Livneh
Lesion bypass DNA polymerases replicate across non-DNA segments
PNAS, December 9, 2003; 100(25): 14760 - 14765.
[Abstract] [Full Text] [PDF]

Home page
 GeneticsHome page
D. H. Ardell, C. A. Lozupone, and L. F. Landweber
Polymorphism, Recombination and Alternative Unscrambling in the DNA Polymerase {alpha} Gene of the Ciliate Stylonychia lemnae (Alveolata; class Spirotrichea)
Genetics, December 1, 2003; 165(4): 1761 - 1777.
[Abstract] [Full Text] [PDF]

Home page
 Proc. Natl. Acad. Sci. USAHome page
L. F. Landweber, T.-C. Kuo, and E. A. Curtis
Evolution and assembly of an extremely scrambled gene
PNAS, March 28, 2000; 97(7): 3298 - 3303.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Print PDF (151K) Freely available
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (22)
Right arrowRequest Permissions
Right arrow Commercial Re-use Guidelines
for Open Access NAR Content
Google Scholar
Right arrow Articles by Prescott, D. M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Prescott, D. M.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?