Members of the Arabidopsis Actin Gene Family Are Widely Dispersed in the Genome

Corresponding author: R. B. Meagher, Department of Genetics, University of Georgia, Athens, GA 30602, meagher{at}bscr.uga.edu (E-mail).

Communicating editor: E. MEYEROWITZ

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Plant genomes are subjected to a variety of DNA turnover mechanisms that are thought to result in rapid expansion and presumable contraction of gene copy number. The evolutionary history of the 10 actin genes in Arabidopsis thaliana is well characterized and can be traced to the origin of vascular plant genomes. Knowledge about the genomic position of each actin gene may be the key to tracing landmark genomic duplication events that define plant families or genera and facilitate further mutant isolation. All 10 actin genes were mapped by following the segregation of cleaved amplified polymorphisms between two ecotypes and identifying actin gene locations among yeast artificial chromosomes. The Arabidopsis actin genes are widely dispersed on four different chromosomes (1, 2, 3, and 5). Even the members of three closely related and recently duplicated pairs of actin genes are unlinked. Several other cytoskeletal genes (profilins, tubulins) that might have evolved in concert with actins were also mapped, but showed few patterns consistent with that evoulutionary history. Thus, the events that gave rise to the actin gene family have been obscured either by the duplication of very small genic fragments or by extensive rearrangement of the genome.

OUR ability to map plant genomes has become increasingly sophisticated, and the genomes for which maps exist include an increasingly wider representation of the plant kingdom. As a result of this growing body of knowledge, duplications of various gene regions have been identified in a number of species (KOWALSKI et al. 1994 ; BENNETZEN and FREELING 1997 ). Some of these duplications may have been landmark events that led to the diversification of plant species. Gene duplications are largely regarded as necessary to remove functional genes from constraint (WALSH 1995 ) and allow new gene functions to evolve in concert with the macroevolution of organs and tissues (KIMURA 1991 ; MEAGHER 1995 ). Whether these duplications originally arose from endoredundancy of small chromosomal regions, from the nomadic duplication of sequences to a new site, from polyploidization of a whole genome, or from the fusion of two diverged genomes into an allopolyploid, the final result might look the same, with duplicated genomic regions becoming more highly rearranged over time. Because many of these duplications generated gene families early in land-plant evolution (MEAGHER et al. 1989 ; BELOSTOTSKY and MEAGHER 1993 ; HUANG et al. 1996 ) it is not surprising that even the small model genome of Arabidopsis contains a high frequency of duplicated gene coding regions (MCGRATH et al. 1993 ).

Plant gene family members produced by ancient duplications can potentially be used to monitor landmark genomic duplication events. The earliest duplication and divergence in the well-characterized Arabidopsis actin gene family generated the reproductive and vegetative classes from a common ancestral actin gene ~350–500 million years ago (mya) (MEAGHER and WILLIAMSON 1994 ; MCDOWELL et al. 1996 ). Similar ancient divergence times have been estimated for a few members of the rice, maize, soybean, and tobacco actin gene family. Due to the extreme conservation of the actin protein sequence and structure, the replacement nucleotide substitutions (RNS) within actin codons (i.e., substitutions that, when changed, cause a change in amino acid sequence) are estimated to have evolved slowly, at about 1% change per 50–100 my (HIGHTOWER and MEAGHER 1986 ; MCDOWELL et al. 1996 ) and were used to estimate the timing of these events. A gene tree, based on divergence of RNS and showing the relationship of the 10 Arabidopsis actins, is presented in Figure 1A. The most divergent of the functional actins, ACT1 and ACT2, differ by ~6.7% RNS or ~6% in amino acid sequence. ACT5 and ACT9 appear to be pseudogenes based on the quality and extent of sequence divergence and lack of detectable transcripts; thus their more extreme divergence is due to lack of functional constraint.

View larger version (25K): In this window In a new window Download PPT slide   Figure 1. —Structure of the Arabidopsis actin gene tree and a typical plant actin gene. (A) The Arabidopsis actin gene tree was prepared from a parsimony analysis of RNS among the eight functional actin genes and two pseudogenes similar to one of the trees presented in MCDOWELL et al. 1996 . The distant yeast actin gene gives a possible rooting for the tree. The scale of divergence can be judged from the vegetative actins (ACT2, ACT8, ACT7), which differ from the reproductive actins (ACT11, ACT1, ACT3, ACT4, and ACT12) by about 5–6.9% RNS, and which all differ from yeast actin (Sce) by about 17% RNS. Trees with similar topography are constructed by most tree-building methods. The validity of this tree structure has been discussed previously (MCDOWELL et al. 1996 ). The accession numbers for the various Arabidopsis actins are as follows: (ACT1) U39449, (ACT2) U41998, (ACT3) U39480, (ACT4) U27980, (ACT7) U27811, (ACT8) U42007, (ACT11) U27981, and (ACT12) U27982. (B) The plant actin gene structure presented holds for all 10 actins with the exception of ACT2, which lacks the intron after codon 20 and the two pseudogenes, which are highly diverged at their amino terminal ends. The location of the degenerate sense (PLACT1119S) and antisense (PLACT1A, 11A, and 284A) primers used in the PCR amplification (MATERIALS AND METHODS) of the various actin gene sequences are shown with their 3' ends indicated by the direction of the arrow.

Using similar arguments for the divergence of functional genes, each of the five subclasses of expressed actins is thought to have arisen by duplication from a common ancestral sequence 150 to 300 mya (MCDOWELL et al. 1996 ). The relatively recent duplications that generated closely related pairs of actin genes in three of the five Arabidopsis actin subclasses [ACT2 and ACT8 (vegetative), ACT1 and ACT3 (reproductive), ACT4 and ACT12 (reproductive)] are thought to have occurred 30 to 60 mya. The members of these gene pairs encode proteins with only one amino acid difference each (~0.3–0.6% RNS), not sufficient divergence to accurately date their ancestry. However, each of these three pairs of actins has diverged in 56–86% of the silent nucleotide substitution (SNS) sites within codons. At unselected or poorly selected sites, such as the SNS within codons (substitutions that do not cause an amino acid replacement), nuclear genes in most organisms evolve much more rapidly, about 1–2%/my (MEAGHER et al. 1989 ; WOLFE et al. 1989 ). Based on the SNS data, we dated these three duplications. It seems possible that all these genes were duplicated simultaneously during an ancient polyploidization event, perhaps about the time that Brassicacea arose. Thus, the evolution of the plant actin gene family and particularly that in Arabidopsis has been well characterized, making it an ideal subject for tracing genomic and subgenomic duplication events.

A functional analysis of the actin gene family members requires identifying actin mutants; this can be aided with knowledge of map position. Although Arabidopsis actin mutants have been identified using a sequence-based method that relies on screening for the junctions between actin gene sequence and an inserted foreign sequence in large insertional libraries (MCKINNEY et al. 1995 ), insertion libraries with complete coverage of the genome are not yet available. However, actin mutants may already exist in the vast collections of embryo lethal mutations (MEINKE 1991 ), and these could be tentatively identified by knowing actin gene map position. Futhermore, because transposons are thought to move more frequently to closely linked sites in the genome (KELLER et al. 1993 ), sets of mutant actin alleles could be constructed with knowledge of map position for the actins and adjacent mobile elements.

Therefore, all 10 actin genes were mapped in the Arabidopsis thaliana genome to allow this well-characterized gene family to be used to track genomic duplication events and to further the identification of mutants. Recombinant inbred lines descended from a hybrid between the Columbia and Landsberg ecotypes of A. thaliana were used to monitor the segregation of restriction site cleavage polymorphisms on specific actin gene fragments amplified by polymerase chain reaction [mapping of cleaved amplified polymorphic sites (CAPS)]. Actin gene-specific probes were used to identify the various actin genes on yeast artificial chromosomes (YACs), further defining their map positions. The relevance of these data to the functional analysis of actin and to studies of genome evolution are discussed.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

CAPS mapping:
Recombinant inbred (RI) lines of A. thaliana were generated from a cross between Landsberg erecta and Columbia by LISTER and DEAN 1993 . The 99 RI lines were obtained from the Arabidopsis Biological Resource Center (Columbus, OH, stock #CS1899). The individual lines were grown in 4-inch pots for 4 to 6 wk, and then leaf tissue was frozen in liquid nitrogen and stored at -70°. DNA was extracted from 1 g of tissue for each line as previously described in MCKINNEY et al. 1995 . Approximately 10 µg of DNA was purified per 1 g of tissue, and this was resuspended in 20 µl of water. One µl of DNA (0.5 µg) was used in a 50-µl PCR reaction.

The technique for CAPS mapping was described by KONIECZNY and AUSUBEL 1993 . The individual actin, profilin, or tubulin genes were amplified with PCR using either two unique oligonucleotides homologous to either the 5' or 3' UTR of the gene of interest or one unique and one degenerate oligonucleotide synthesized from within coding region sequence as described in Table 1. The PCR profile for amplification was as follows: initial 2 min incubation at 94°, and then 45 cycles of 1 min at 94°, 1 min at 52°, 1 to 2 min at 72°. One unit of Taq enzyme, 15 pmol of each-gene specific oligonucleotide, or 25 pmol of a degenerate oligonucleotide, and 1 mM dNTPs in Taq buffer (Promega Co., Madison, WI) were used in a 50-µl reaction. Each specific actin, tubulin, or profilin gene was amplified in both Columbia and Landsberg parent lines. The PCR products were then digested with a variety of 10 to 30 restriction enzymes that were known to cleave from two to six times within the specific gene sequence. The cleaved products were then resolved in adjacent lanes on a 2% agarose gel and the polymorphisms that were found between the parent strains were used to screen the population of RI lines. Each plant line was then scored as either having a segregation pattern characteristic of C (Columbia) or L (Landsberg) parents for that gene. In a few cases, the plant lines were scored H (heterozygous) for a specific gene. This data was then entered into the computer program Mapmaker for Macintosh V2.0 (LANDER et al. 1987 ) provided by SCOTT TINGEY (DuPont, Wilmington, DE) and mapped with respect to the LISTER and DEAN 1993 data set containing 67 markers on the five different chromosomes. As a function provided by the Mapmaker program, the relative likelihood ratios, and the negative log of a probability term, were determined for each marker being mapped relative to known flanking markers on that chromosome. The latest RI mapping data can be found at http://nasc.nott.ac.uk/new ri map.html.

YAC mapping:
The CIC (CREUSOT et al. 1995 ) and YUP (ECKER 1990 ) YAC libraries constructed from Arabidopsis Columbia genomic DNA were screened for the various actins and several profilins. The YAC library filters were generously supplied by CAROLINE DEAN's laboratory (John Innes Institute, Norwich, UK). Library filters were probed with ³²P-ATP-labeled 3' untranslated region (UTR) and/or 5' UTR gene-specific probes. These probes were generated for each gene by PCR amplifying their respective subclones with gene-specific oligonucleotides, as described in Table 2. Using the 3' UTR probes, for example, the sense primer originates at the stop codon and the antisense codon is located 200 to 300 base pairs downstream of the stop codon. The filters were then hybridized to the probe fragment using ³²P-labeled dATP incorporated with random primer. The filters were prehybridized at 60° overnight in 20% formamide, 6x sodium dodecylsulfate (SDS), 5x Denhardt's (MANIATIS et al. 1989 ), 0.5% SDS, 25 mM sodium phosphate pH 6.5, 50 µg ml^-1 tRNA, and 0.1% gelatin. The filter was then hybridized overnight with the gene-specific probe under the same conditions as the prehybridization. The filters were washed twice for 10 min at 50° in 1x SSC, 0.2% SSC, and then exposed to X-ray film for 3 to 7 days. As a control, two filters were probed with purified ACT2 and ACT4 cDNAs, respectively, which hybridize to all plant actin containing YACs in the library (not shown). Similarly, filters were hybridized to a general profilin probe. The profilin probe was generated from a floral cDNA library using the two degenerate profilin oligonucleotides described in Table 1 to PCR amplify the majority of profilin coding sequences. These general probes were labeled and treated under the same conditions as the gene-specific probes. All the signals identified with the general actin probes could be accounted for with clones that were also identified with gene-specific probes and further characterized.

Yeast strains containing YAC clones of the various actins were then ordered from the Arabidopsis Biological Resource Center. Genomic DNA was purified from these strains (AUSUBEL et al. 1989 ). The DNA was then digested with EcoRI, HindIII, BglII, and XbaI restriction enzymes, then treated as described in MCDOWELL et al. 1996 , and the Southern blots of these samples were probed with the appropriate 5' UTR actin gene-specific probe. The YACs of interest were then located on one of the five Arabidopsis chromosomes using the following resources: Arabidopsis thaliana Genome Center at the University of Pennsylvania (http://cbil.humgen.upeenn.edu/~atgc/atgcup.html); chromosome 1, Dr. J. ECKER (personal communication); chromosome 2, Dr. H. GOODMAN (ZACHGO et al. 1996 ); chromosome 3, Dr. D. BOUCHEZ (personal communication); chromosome 4, Dr. C. DEAN (SCHMIDT et al. 1995 ); and chromosome 5, Dr. C. DEAN (SCHMIDT et al. 1997 ).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Mapping actin gene polymorphisms using recombinant inbred lines:
Cleaved amplified polymorphisms (CAPS) between the Columbia (CC) and Landsberg (LL) ecotypes of A. thaliana were identified for nine of the 10 actin genes family members (KONIECZNY and AUSUBEL 1993 ). A portion of each actin gene was amplified using one PCR primer specific for the actin gene of interest and one opposing degenerate actin primer (Table 1). Because the two pseudogenes ACT5 and ACT9 were known to be closely physically linked (MCDOWELL et al. 1996 ), only ACT9 was mapped genetically. The 550- to 2280-bp PCR products from the various genes (Table 1) were cleaved with a variety of restriction enzymes. The restriction enzymes screened were selected based on the presence of their recognition sequences in the known Columbia actin sequences at intervals of ~200–900 bp. The polymorphisms used in actin gene mapping are shown in Figure 2A. The segregation of the CAPS markers was then analyzed in a well-characterized population of 99 recombinant inbred plant lines descended from a CL hybrid (LISTER and DEAN 1993 ). Well-supported map positions were determined for all nine genes, as summarized on the left side of Table 3 and illustrated in Figure 3. The relative likelihood ratios determined on either side of the actin markers were all less than -3.0, suggesting that their ordering relative to flanking markers is unambiguous and their map positions are reliable.

View larger version (90K): In this window In a new window Download PPT slide   Figure 2. —Identifying CAPS markers and YAC containing clones for actin gene mapping. (A) The cleaved amplified polymorphisms (CAPs) for the eight functional actins and one actin pseudogene between Columbia (CC) and Landsberg (LL) lines are shown. The size of the original PCR product and the size of the diagnostic fragments are given in Figure 1, and the mapping data are summarized in Table 3 and Figure 3. The AluI-digested pBR322 DNA size standards are 908, 659, 657, 521, 403, 281, 257, 220, and 100 bp. (B) The yeast artificial chromosomes (YACs) containing various Arabidopsis actins were identified by a Southern blot of the yeast colony imprint of an 8 x 12 grid of colonies. The filter was hybridized with a 3' gene-specific probe for ACT8 (Table 2). Seven YACs containing ACT8 are identified on this blot (see Table 3). (C) Each of the 96 positions in the YAC grid contain two imprints of 12 different YAC clones. One of the patterns observed in the grid in B is interpreted as YAC #1.

View larger version (41K): In this window In a new window Download PPT slide   Figure 3. —Chromosomal map positions of the various Arabidopsis actins. Locations of the various Arabidopsis actins, profilins, and ß-tubulins are shown on a linkage map as determined by CAPS mapping and confirmed by YAC mapping. Those cytoskeletal genes shown in bold were mapped in this study. Those genes that were mapped previously are underlined. The majority of CAPs markers shown were derived from LISTER and DEAN 1993 and enhanced by markers reported by JARVIS et al. 1994 . The tubulin genes TUB1, TUB5, and TUB6 were mapped previously using RFLPs (MCGRATH et al. 1993 ), and their map positions have not been confirmed by CAPS mapping. The profilin genes PRF1, PRF2, and PRF4 (HUANG et al. 1996 ) were identified in an independent study as PFN1, PFN2, and PFN3, respectively (CHRISTENSEN et al. 1996 ).

The actin genes are widely dispersed (Figure 3A) on four (chromosomes 1, 2, 3, 5) of the five different chromosomes in A. thaliana. ACT8 maps to chromosome 1. The two presumptive pseudogenes, ACT5 and ACT9, were separated by only 1.1 cM from the functional gene, ACT1, on chromosome 2. The location of ACT1 on the map agrees well with that obtained from RFLP mapping data (CHANG et al. 1988 ; NAIRN et al. 1988 ). Four genes (ACT11, ACT2, ACT12, and ACT3) are spread out along chromosome 3, spaced by 6.5 cM (ACT11-ACT2), 35.0 cM (ACT2-ACT12), and 14.6 cM (ACT12-ACT3). ACT4 and ACT7 both mapped to chromosome 5, but they are too far apart to be genetically linked. None of the three pairs of actin genes that encode closely related actin proteins (ACT1 and ACT3; ACT2 and ACT8; ACT4 and ACT12) are on the same chromosomes.

Mapping on YACs:
In order to more precisely map the various actin genes, YACs containing each gene were identified. Actin gene-specific probes were prepared using the primers described in Table 2 to PCR amplify the 5' or 3' UTR of each gene. Nylon filters with the DNA imprint from ~900 different YACs were hybridized to gene-specific probes for each of the nine actins (see MATERIALS AND METHODS). For example, a filter hybridized at high stringency to an ACT8 gene-specific probe is shown in Figure 2B. The autoradiograph reveals seven YACs containing the ACT8 gene. The duplication of each of 12 YACs in one of the 96 5-by-5 grids helps to confirm that the signal is due to hybridization, as interpreted for one ACT8 hybridizing YAC in Figure 2C. Many of the actin containing YACs identified are arranged in contiguous overlapping sets (contigs) on maps of the various Arabidopsis chromosomes as is the case for six of the seven ACT8 hybridizing YACs (Table 3; D. BOUCHEZ and J. R. ECKER, unpublished data). YACs were identified similarly for the other seven functional actins. DNA was prepared from the yeast strains containing the various actin gene-containing YACs and digested with restriction endonucleases to produce a diagnostic pattern with each gene (MCDOWELL et al. 1996 ). These digests were separated by agarose gel electrophoresis, blotted to a filter, and probed with the respective actin sequence. In every case, the YACs identified contained actin gene fragments of the expected size and no others.

The YACs identified and confirmed by Southern blotting for the eight functional actins are summarized on the right side of Table 3. No YACs were identified for the two pseudogenes, ACT5 and ACT9, in spite of repeated attempts to probe filters with 5'- and 3'-specific probes. Hybridization of these filters with a general actin probe (see MATERIALS AND METHODS) identified the same YACs as the gene-specific probes and no others. The map positions of 25 of the 27 YACs have been determined by other groups (see MATERIALS AND METHODS). There is an excellent agreement between the location of these 25 YACs containing actin genes and the map position of that gene determined by mapping CAPS on the recombinant inbred lines. The only YAC identified as containing ACT3 had not been mapped previously but can now be tentatively assigned to chromosome 3 (Table 3) based on our CAPs mapping data.

Linkage of cytoskeletal gene families:
It has been proposed that preceding the macroevolution of novel organs, tissues, and cytoskeletal structures, large sets of genes were duplicated. Once expressed in a newly evolved organellar or cellular environment, these genes were modified through mutation and selection to fill more specialized tasks (KIMURA 1991 ; MEAGHER 1995 ). For example, Arabidopsis profilin gene family members PRF1, PRF2, and PRF3 are expressed primarily in vegetative tissues, whereas PRF4 and PFN4 are most strongly expressed in reproductive tissues such as pollen (CHRISTENSEN et al. 1996 ; HUANG et al. 1996 ). These two ancient classes of profilins may have evolved in concert with the correspondingly expressed classes of actins during the evolution of vascular plant tissues and organs. Profilin is involved in G-actin sequestration and F-actin polymerization with more than 20 amino acid side chains contacting residues in actin. Thus, a pollen-specific profilin gene might be expected to coevolve with and perhaps be genetically linked to a pollen-specific actin gene. Using the same CAPs mapping protocol described above for actin, we mapped the vegetative PRF1 gene to chromosome 2, quite distant from any of the vegetative actins (ACT2, ACT7, and ACT8), and mapped the pollen-specific profilin gene, PRF4, to chromosome 4, where no other actins mapped.

It is somewhat suprising that PRF2, a vegetative profilin is closely and physically linked to the pollen-specific profilin, PRF4 (CHRISTENSEN et al. 1996 ). From CAPs mapping we determined that PRF1, a vegetative profilin, was also genetically linked to the pollen-specific gene, PFN4, with no crossovers between them in the 99 RI lines examined (Table 3). Our physical mapping of PRF1 and PFN4 placed them on the same YAC clones. So in each case a vegetatively expressed profilin gene is closely physically paired with a pollen-specific gene. The two vegetative genes are very close sequence homologs as are the two pollen-specific genes (CHRISTENSEN et al. 1996 ; HUANG et al. 1996 ), and thus it is reasonable to propose that these two linked sets of profilins share a recent common ancestry by gene duplication. A more detailed sequence analysis among these four genes revealed that the sequence homologs in each pair differ by 60 to 80% in SNS. Based on the standard SNS rate of 1–2%/MY, the two pairs of homologs should not have had a common ancestor for at least 30 my. This degree of divergence is quite similar to that predicted for the three closely related actin pairs.

Although the tubulins encode a fundamental component of a separate cytoskeletal system from actin, microtubules are involved in many of the same cellular processes as actin. Several Arabidopsis ß-tubulin genes with predominantly vegetative- (TUB1, TUB5, TUB6, and TUB8) or floral-specific (TUB2, TUB3, TUB7, and TUB9) expression patterns have been identified (SNUSTAD et al. 1992 ). Like actin, tubulin gene expression must have been linked into the macroevolution of vascular plant organs and tissues (MEAGHER 1995 ). Seven of the nine ß-tubulin genes were mapped previously by monitoring the segregation of RFLPs for several genes simultaneously (MCGRATH et al. 1993 ). In order to better integrate the ß-tubulin map positions with the actin gene map developed herein, we performed CAPs mapping on three of these tubulin gene loci (TUB2/3, TUB4, and TUB9) and one unmapped tubulin gene (TUB7). The results are shown in Figure 3. The vegetative genes TUB6, TUB8, and ACT7 mapped near each other (22–24 cM) on chromosome 5. Similarly, among the floral-specific genes, TUB7 mapped near ACT1 (6.9 cM) and TUB2/3 mapped close to ACT4 (3 cM). These data suggest some possible common origin for these loci. However, a random distribution of the eight ß-tubulin loci and nine actin loci would have spaced them an average of 30–40 cM apart. Thus, it is difficult to say that these few correlations in the map positions of similarly expressed genes are not coincidental, just falling within a random distribution of loci.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The 10 Arabidopsis actin genes were genetically mapped using CAPs markers on a previously defined population of recombinant inbred lines. Sufficient polymorphism was available between the Columbia and Landsberg lines such that at least one CAPs marker could be found for each gene within the coding sequence or within ~1000 bp flanking it. The actin map positions were confirmed for seven of the eight functional actins by physically mapping each onto previously characterized YACs. The YAC libraries were complete enough that at least one YAC was identified for each of the eight functional actins, although no YAC was found to hybridize to the two linked actin pseudogenes.

The eight functional actin genes were widely dispersed on four of the five Arabidopsis chromosomes, in spite of their common ancestry through gene duplication. The two potential pseudogenes, ACT5 and ACT9, were genetically closely linked to ACT1 (1.1 cM). However, these pseudogenes are not closely related to ACT1, and they are separated from ACT1 by crossovers in at least four of the 99 recombinant inbred lines. ACT1 was alone on the one YAC identified (it did not contain ACT5 and ACT9). We know that the earliest divergence in the actin gene family, which split the vegetative and reproductive classes of genes from a common ancestor, occurred early in land-plant evolution (Figure 1). Thus, it was not surprising to find that the vegetative actins (ACT2, ACT7, and ACT8) were not closely linked to the reproductive actins (ACT1, ACT3, ACT4, ACT11, and ACT12).

It would also be expected that during the hundreds of millions of years since these two classes of genes had a common ancestry the actin genes have been rearranged numerous times. However, the three duplications that separated the three closely related pairs of actins (ACT1/ACT3, ACT2/ACT8, and ACT4/ACT12) from three ancestral sequences are predicted to have occurred only 30–60 mya, about the time that the Brassicaseae are first detected in the fossil record (CRONQUIST 1981 ). Thus, it seems possible that remnants of these duplications might still remain in the genome, and this well-characterized set of actin genes might be used to date these duplication events. GAPB is adjacent to ACT8, and ACT2 is flanked on either side by the GAPA and GAPC loci (SHIH et al. 1991 ). GAPC is closer to ACT11 and is a highly divergent member of the GAP gene family, perhaps tracing its origin to the origin of eukaryotic plant cells (SHIH et al. 1988 ). This leaves GAPA and GAPB as possible candidates for a gene duplicated on a large DNA fragment along with ACT2 and ACT8 in some common ancestor. GAPA is 18.9 cM from ACT2, and there are no other known duplicated genes in this region in common with the ACT8-GAPB region.

MCGRATH et al. 1993 identified a large number of randomly isolated cDNA probes that hybridized to more than one gene in the Arabidopsis genome and presumably represent closely related members of gene families. RFLP data were used to map these loci and revealed several potentially duplicated genomic regions. Two of these markers, TUB4 (ß-tubulin gene 4) and 213A, map in regions near ACT4 on chromosome 5, and two sequence homologs of these, TUB2/3 (ß-tubulin genes 2 and 3 are within a few kb of each other) and 213C, map in regions near ACT12 on chromosome 3. Thus, the regions surrounding ACT4 and ACT12 had the potential to represent the product of a large recent genomic duplication. Our CAPs mapping data confirmed that TUB4 was linked, but not closely (20 cM), to ACT4 on chromosome 5. However, TUB2/3 mapped immediately adjacent to ACT4 (3 cM), inconsistent with previous RFLP data that had placed it on chromosome 3 in the region of ACT12 (MCGRATH et al. 1993 ). TUB9 is more closely related in sequence to TUB4 than are TUB2 or TUB3 and could represent a more recent duplication from a common ancestor with TUB4. Thus, it seemed possible that TUB9 was the missing tubulin gene originally mapped adjacent to ACT12 on chromosome 3. The location we determined by CAPs mapping for TUB9 on chromosome 4 was consistent with the previous data (MCGRATH et al. 1993 ). Thus, the supposition that ACT12 and ACT4 are on large blocks of related DNA sequence remaining from the original duplication event was not supported by the mapping of adjacent tubulin markers.

We have mapped two loci in which representatives of the vegetative and reproductive profilin gene classes are closely physically linked, and these two loci share a recent common ancestry. In identifying the CAPs markers for these two regions we found a surprising difference in the frequency of polymorphisms between the Columbia and Landsberg ecotypes of Arabidopsis. The PRF2/PRF4 pair on chromosome 4 had only a few linked polymorphisms, as expected (one polymorphism out of 25 different four-base restriction enzymes tested). The other pair of profilins, PRF1 and PFN4, had at least a 10 times higher degree of polymorphism (11 out of 12 enzymes and 14 out of 16 enzymes tested, respectively). No other loci examined in this study showed anywhere near this degree of polymorphism between the two Arabidopsis ecotypes. The PRF1 and PFN4 genes appear to encode relatively normal plant profilin protein sequences typical of their two profilin classes, and both are strongly expressed at the RNA level. Thus, there is nothing to suggest that this highly polymorphic pair is not under selection. One or both of these genomic regions may be part of an isochore undergoing rapid sequence divergence (MONTERO et al. 1990 ; BERNARDI 1993 ).

In petunia, the local duplication of actin gene subfamilies at distant loci in the genome resulted in a family with more than 100 members (MCLEAN et al. 1988 ; MCLEAN et al. 1990 ). Apparently endoredundancy is common for genes in petunia (RICK 1943 , RICK 1971 ). Something quite distinct appears to have happened to the actin gene family through the recent ancestry of Arabidopsis. Even the most closely related and recent actin duplications in Arabidopsis are unlinked, and remains of the original duplication were not obvious from this study. These actin mapping data are consistent with rapid shuffling and reorganization of small genic regions suggested from recent large-scale mapping (LAGERCRANTZ and LYDIATE 1996 ) and sequencing (TREMOUSAYGUE et al. 1997 ) projects in Arabidopsis. There are a few explanations for the present dispersed locations of the three pairs of closely related actins. First, the various actin genes originally may have been duplicated during a genomic polyploidization event or may have resulted from subchromosomal duplications. The present dispersed map positions would then have resulted from numerous subsequent rearrangements of the genome. Second, the duplications may have resulted directly in the translocation of the duplicate copy to a new locus in the genome as proposed in the "nomad" model for duplications in plants (PICHERSKY 1990 ). In particular, the more recent duplications that gave rise to three pairs of closely related actins must have been affected by one of these two mechanisms in the Brassica ancestry of Arabidopsis.

Only rearrangements or nomadic duplications that preserved actin genes with required functions would have been expected to have survived subsequent evolution and DNA turnover (PREISLER and THOMPSON 1981 ). Although it is easy to imagine mechanisms of large genomic duplication leading to a viable organism with extra gene copies (CAVENER 1987 ; DICKINSON 1988 ), it is more difficult to propose mechanisms by which massive reductions in the amount of DNA and numerous rearrangements could leave Arabidopsis with such a small genome and leave each functional actin gene (let alone the thousands of other required genes) on their own small island of DNA. Thus, the latter mechanism of nomadic duplications seems the more likely scenario for the dispersal of the actin, tubulin, and profilin genes discussed herein. Much more detailed mapping of the actin gene family and other gene family regions will be required to trace the origin of these DNA turnover events in Arabidopsis. Understanding the details of genome evolution in the small genome of Arabidopsis will greatly facilitate our interpretation of these processes in complex genomes of crop plants.

We have previously isolated insertion mutants in Arabidopsis actin genes ACT2 and ACT4, using a sequence-based screening approach (MCKINNEY et al. 1995 ). Recent research has shown that in spite of the existence of a closely related copy of each of these genes, loss of either gene is deleterious to the survival of Arabidopsis grown in a population (E. C. MCKINNEY and R. B. MEAGHER, unpublished results; M. A. ASMUSSEN, L. U. GILLILAND and R. B. MEAGHER, unpublished results). Because information demonstrating a requirement for each actin family member is invaluable to our dissecting actin function, many more actin alleles are needed. Knowing the chromosomal map positions for each of the functional actin genes, we may be able to identify actin mutants from existing stocks of embryo lethal mutants (MEINKE 1991 ) or to construct insertional mutants in actins using closely linked transposons. Active insertion element systems in Arabidopsis have been constructed in several laboratories (DEAN et al. 1990 ; BANCROFT et al. 1992 ; FELDMANN et al. 1994 ; OSBORNE et al. 1995 ). One the first available libraries with 10 well-mapped Ds elements (SMITH et al. 1996 ) contains one transposon within 2 cM of ACT4 and others within 10–15 cM of ACT7 and ACT8. Clearly, the functional genetic analysis of complex gene families is greatly enhanced by the diversity of genetic resources and techniques available in Arabidopsis.

	ACKNOWLEDGMENTS

This project was made possible by the generosity of several laboratories. Special thanks are extended to CAROLINE DEAN and KATRINA LOVE, at the John Innes Institute, Norwich, England, who provided the YAC library filters; to DAVID BOUCHEZ and his laboratory in Laboratoire de Biologie Cellulaire, Institute Nationale de la Recherche Agronomique, Versaille, France; and JOE ECKER and CHRISTOPHER KIM and colleagues at Penn State University, College Station, Pennsylvania, for their unpublished mapping data on numerous YACs and BACs. GARY KOCHERT in the Botany Department here at the University of Georgia has offered many useful insights and was particularly helpful with the use of Mapmaker.

Manuscript received January 7, 1998; Accepted for publication February 20, 1998.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

AUSUBEL, F. M., R. BRENT, R. E. KINGSTON, D. D. MOORE, J. G. SEIDMAN et al., 1989 Current Protocols in Molecular Biology. Greene Publishing Associates and Wiley-Interscience, New York.

BANCROFT, I., A. M. BHATT, C. SJODIN, S. SCOFIELD, and J. D. G. JONES et al., 1992  Development of an efficient two-element transposon tagging system in Arabidopsis thaliana.. Mol. Gen. Genet. 233:449-461[Medline].

BELOSTOTSKY, D. and R. B. MEAGHER, 1993  Differential organ-specific expression of three poly(A) binding protein genes from Arabidopsis.. Proc. Natl. Acad. Sci. USA 90:6686-6690[Abstract/Free Full Text].

BENNETZEN, J. L. and M. FREELING, 1997  The unified grass genome: synery in synteny. Genome Res. 7:301-306[Free Full Text].

BERNARDI, G., 1993  The vertebrate genome: isochores and evolution. Mol. Biol. Evol. 10:186-204[Abstract].

CAVENER, D. R., 1987  Combinatorial control of structural genes in Drosophila: solutions that work for the animal. BioEssays 7:103-107[Medline].

CHANG, C., J. L. BOWMAN, A. W. DEJOHN, E. S. LANDER, and E. M. MEYEROWITZ, 1988  Restriction fragment length polymorphism linkage map for Arabidopsis thaliana.. Proc. Natl. Acad. Sci. USA 85:6856-6860[Abstract/Free Full Text].

CHRISTENSEN, H. E. M., S. RAMACHANDRAN, C.-T. TAN, U. SURANA, and C.-H. DONG et al., 1996  Arabidopsis profilins are functionally similar to yeast profilins: identification of a vascular bundle-specific profilin and a pollen-specific profilin. Plant J. 10:269-279[Medline].

CREUSOT, F., E. FOUILLOUX, M. DRON, J. LAFLEURIEL, and G. PICARD et al., 1995  The CIC library: a large insert YAC library for genome mapping in Arabidopsis thaliana.. Plant J. 8:763-770[Medline].

CRONQUIST, A., 1981 An Integrated System of Classification of Flowering Plants. Columbia Unversity Press, New York.

DEAN, C., C. SJODIN, E. LAWSON, C. LISTER and I. BANCROFT, 1990 Development of an efficient transposon tagging system in Arabidopsis. Abstract, p. 8, of the Fourth International Conference on Arabidopsis Research.

DICKINSON, W. J., 1988  On the architecture of regulatory systems: evolutionary insights and implications. BioEssays 8:204-208[Medline].

ECKER, J. R., 1990  PFGE and YAC analysis of the Arabidopsis genome. Methods 1:186-194.

FELDMANN, K. A., R. L. MALMBERG and C. DEAN, 1994 Mutagenesis in Arabidopsis, pp. 137–172 in Arabidopsis, edited by E. M. MEYEROWITZ and C. R. SOMERVILLE. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

HIGHTOWER, R. C. and R. B. MEAGHER, 1986  The molecular evolution of actin. Genetics 114:315-332[Abstract/Free Full Text].

HUANG, S., C. CHAMBLISS, and R. B. MEAGHER, 1996  The Arabidopsis profilin gene family: evidence for an ancient split between constitutive and pollen specific genes. Plant Physiol. 111:115-126[Abstract].

JARVIS, P., C. LISTER, V. SZABO, and C. DEAN, 1994  Integration of CAPS markers into the RFLP map generated using recombinant inbred lines of Arabidopsis thaliana. Plant Mol. Biol. 24:685-687[Medline].

KELLER, J., E. LIM, and H. DOONER, 1993  Preferential transposition of Ac to linked sites in Arabidopsis.. Theor. Appl. Genet. 86:585-588.

KIMURA, M., 1991  Recent development of the neutral theory viewed from the Wrightian tradition of theoretical population genetics. Proc. Natl. Acad. Sci. USA 88:5969-5973[Abstract/Free Full Text].

KONIECZNY, A. and F. M. AUSUBEL, 1993  A procedure for mapping Arabidopsis mutations using co-dominant ecotype-specific PCR-based markers. Plant J. 4:403-410[Medline].

KOWALSKI, S. P., T.-H. LAN, K. FELDMANN, and A. H. PATERSON, 1994  Comparative mapping of Arabidopsis thaliana and Brassica oleracea chromosomes reveals islands of conserved organization. Genetics 138:499-510[Abstract].

LAGERCRANTZ, U. and D. J. LYDIATE, 1996  Comparative genome mapping in Brassica. Genetics 144:1903-1910[Abstract].

LANDER, E. S., P. GREEN, J. ABRAHMASON, A. BARLOW, and M. DALY et al., 1987  MAPMAKER: an interactive computer package for constructing primary genetic linkage maps of experimental and natural populations. Genomics 1:174-181[Medline].

LISTER, C. and C. DEAN, 1993  Recombinant inbred lines for mapping RFLP and phenotypic markers in Arabidopsis thaliana.. Plant J. 4:745-750.

MANIATIS, T., E. F. FRITSCH and J. SAMBROOK, 1989 Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

MCDOWELL, J. M., S. HUANG, E. C. MCKINNEY, Y.-Q. AN, and R. B. MEAGHER, 1996  Arabidopsis thaliana contains ten actin genes encoding six ancient protein subclasses. Genetics 142:587-602[Abstract].

MCGRATH, J. M., M. M. JANCSO, and E. PICHERSKY, 1993  Duplicate sequences with a similarity to expressed genes in the genome of Arabidopsis thaliana.. Theor. Appl. Genet. 86:880-888.

MCKINNEY, E. C., N. ALI, A. TRAUT, K. A. FELDMANN, and D. A. BELOSTOTSKY et al., 1995  Sequence based identification of T-DNA insertion mutations in Arabidopsis: actin mutants act2-1 and act4-1.. Plant J. 8:613-622[Medline].

MCLEAN, M., W. V. BAIRD, A. G. M. GERATS, and R. B. MEAGHER, 1988  Determination of copy number and linkage relationships among five actin gene subfamilies in Petunia hybrida.. Plant Mol. Biol. 11:663-672.

MCLEAN, M., A. G. M. GERATS, W. V. BAIRD, and R. B. MEAGHER, 1990  Six actin gene subfamilies map to five chromosomes of Petunia hybrida.. J. Hered. 81:341-346.

MEAGHER, R. B., 1995 The impact of historical contingency on gene phylogeny: plant actin diversity, pp. 195–215 in Evolutionary Biology, edited by M. K. HECHT, R. J. MACINTYRE and M. T. CLEGG. Plenum Press, New York.

MEAGHER, R. B., and R. E. WILLIAMSON, 1994 The Plant Cytoskeleton, pp. 1049–1084 in Arabidopsis, edited by E. MEYEROWITZ and C. SOMERVILLE. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

MEAGHER, R. B., S. BERRY-LOWE, and K. RICE, 1989  Molecular evolution of the small subunit of ribulose bisphosphate carboxylase: nucleotide substitution and gene conversion. Genetics 123:845-863[Abstract/Free Full Text].

MEINKE, D. W., 1991  Embryonic mutants of Arabidopsis thaliana.. Dev. Genet. 12:382-392.

MONTERO, L. M., J. SALINAS, G. MATASSI, and G. BERNARDI, 1990  Gene distribution and isochore organization in the nuclear genome of plants. Nucleic Acids Res. 18:1859-1867[Abstract/Free Full Text].

NAIRN, C. J., L. WINESETT, and R. J. FERL, 1988  Nucleotide sequence of an actin gene from Arabidopsis thaliana.. Gene 65:247-257[Medline].

OSBORNE, B. I., U. WIRTZ, and B. BAKER, 1995  A system of insertional mutagenesis and chromosomal rearrangement using the Ds transposon and Cre-lox. Plant J. 7:687-701[Medline].

PICHERSKY, E., 1990  Nomad DNA—a model for movement and duplication of DNA sequences in plant genomes. Plant Mol. Biol. 15:437-448[Medline].

PREISLER, R. S. and W. F. THOMPSON, 1981  Evolutionary sequence divergence within repeated DNA families of higher plant genomes. J. Mol. Evol. 17:78-84[Medline].

RICK, C. M., 1943  Cyto-genetic consequences of x-ray treatment of pollen in petunia. Bot. Gaz. 104:528-540.

RICK, C. M., 1971 Some cytogenetic features of the genome in diploid plant species. The Stadler Genetics Symposium, Columbia, MO. 1 and 2: 154–175.

SCHMIDT, R., J. WEST, K. LOVE, Z. LENEHAN, and C. LISTER et al., 1995  Physical map and organization of Arabidopsis thaliana chromosome 4.. Science 270:480-483[Abstract/Free Full Text].

SCHMIDT, R., K. LOVE, J. WEST, Z. LENEHAN, and C. DEAN, 1997  Description of 31 YAC contigs spanning the majority of Arabidopsis thaliana chromosome 5.. Plant J. 11:563-572[Medline].

SHIH, M., P. HEINRICH, and H. M. GOODMAN, 1988  Intron existence predated the divergence of eukaryotes and prokaryotes. Science 242:1164-1166[Abstract/Free Full Text].

SHIH, M.-C., P. HEINRICH, and H. M. GOODMAN, 1991  Cloning and chromosomal mapping of nuclear genes encoding chloroplast and cylosolic glyceraldehyde-3-phosphate-dehydrogenase from Arabidopsis thaliana.. Gene 104:133-138[Medline].

SMITH, D., Y. YANAI, Y.-G. LIU, S. ISHIGURO, and K. OKADA et al., 1996  Characterization and mapping of Ds-GUS-T-DNA lines for targeted insertional mutagenesis. Plant J. 10:721-732[Medline].

SNUSTAD, D. P., N. A. HAAS, S. D. KOPCZAK, and C. D. SILFLOW, 1992  The small genome of Arabidopsis thaliana contains at least nine expressed ß-tubulin genes. Plant Cell 4:549-556[Abstract/Free Full Text].

TREMOUSAYGUE, D., C. BARDET, P. DABOS, F. REGAD, and F. PELESE et al., 1997  Genome DNA sequencing around the EF-1 multigene locus of Arabidopsis thaliana indicates a high gene density and a shuffling of noncoding regions. Genome Res. 7:198-209[Abstract/Free Full Text].

WALSH, J. B., 1995  How often do duplicated genes evolve new functions. Genetics 139:421-428[Abstract].

WOLFE, K. H., P. M. SHARP, and W.-H. LI, 1989  Rates of synonymous substitution in plant nuclear genes. J. Mol. Evol. 29:208-211.

ZACHGO, E. A., M. L. WANG, J. DEWDNEY, D. BOUCHEZ, and C. CAMILLERI et al., 1996  A physical map of chromosome 2 of Arabidopsis thaliana.. Genome Res. 6:19-25[Abstract/Free Full Text].

This article has been cited by other articles:

				C. Ringli, N. Baumberger, and B. Keller The Arabidopsis Root Hair Mutants der2-der9 are Affected at Different Stages of Root Hair Development Plant Cell Physiol., July 1, 2005; 46(7): 1046 - 1053. [Abstract] [Full Text] [PDF]

				L. U. Gilliland, M. K. Kandasamy, L. C. Pawloski, and R. B. Meagher Both Vegetative and Reproductive Actin Isovariants Complement the Stunted Root Hair Phenotype of the Arabidopsis act2-1 Mutation Plant Physiology, December 1, 2002; 130(4): 2199 - 2209. [Abstract] [Full Text] [PDF]

				E. C. McKinney, M. K. Kandasamy, and R. B. Meagher Arabidopsis Contains Ancient Classes of Differentially Expressed Actin-Related Protein Genes Plant Physiology, March 1, 2002; 128(3): 997 - 1007. [Abstract] [Full Text] [PDF]

				V. Laporte and D. Charlesworth Non-Sex-Linked, Nuclear Cleaved Amplified Polymorphic Sequences in Silene latifolia J. Hered., July 1, 2001; 92(4): 357 - 359. [Abstract] [Full Text] [PDF]

				R. B. Meagher, E. C. McKinney, and M. K. Kandasamy Isovariant Dynamics Expand and Buffer the Responses of Complex Systems: The Diverse Plant Actin Gene Family PLANT CELL, June 1, 1999; 11(6): 995 - 1006. [Full Text]