Recombination

Abraham B Korol, University of Haifa, Haifa, Israel

r 2013 Elsevier Inc. All rights reserved.

Glossary one segment affects the probability of exchange in an

Chiasmata w-Like configurations in meiosis. adjacent segment) and chromatid interference (which
Crossing-over Complex interaction of homologs within determines the probabilities of chromatid configurations in
bivalents at pachytene resulting in reciprocal exchange of successive pairs of chiasmata, i.e., the proportions of 2-, 3-,
genetic material between nonsister chromatids. and 4-strand double exchanges).
Gene conversion A process of nonreciprocal Synaptonemal complex (SC) A three-part proteinaceous
(unidirectional) transfer of genetic information. structure that affects the number and distribution of
Genetic interference Interdependence of crossover events crossovers and converts crossovers into functional
that includes chiasma interference (when an exchange in chiasmata.

Types of Recombination corresponding alleles is 1 : 1, although a departure from in-

dependent assortment of unlinked genes, termed quasilinkage
There are various ways to classify different types of re- or pseudolinkage, was observed in both plants and animals,
combination. No matter how logical a classification system especially in interspecific hybrids (Korol et al., 1994). Pseudo-
may seem, the simplest way is to compare alternative aspects linkage is widespread in fish species, presumably reflecting the
(modes) of the general phenomenon of recombination. These genome paleopolyploid nature (Danzmann et al., 2008).
modes may be based on types of participating cells (meiotic or
mitotic and male or female), cell compartments (nuclear or
Mitotic Recombination
mitochondrial), chromosomes and DNA sequences (hom-
ologous or nonhomologous), and types of exchanges (re- Exchange of genetic information can also occur during mitotic
ciprocal or nonreciprocal). cell division, but it does so with a frequency of a few orders of
magnitude lower than that of meiotic ones. The major ex-
clusion from this rule is mitotic recombination resulting in
Crossing-Over and Chiasmata
high variation of somatic cells of the vertebrate immune sys-
Two types of cell division are known for eukaryotic organisms: tem. Mitotic recombination may occur in the germline cells,
mitotic and meiotic. Recombination takes place in both. In for example, in Drosophila males where meiotic recombination
cells that enter meiosis, the chromosomes have already is normally suppressed. Some genetic factors increase the rate
undergone DNA replication so that each chromosome consists of mitotic recombination manifold. For example, homo-
of two sister chromatids connected by a shared centromere. zygosity for an inserted mobile element may result in high
During the meiotic prophase, pairing of homologs brings premeiotic recombination rates in Drosophila males compar-
about bivalent formation; each bivalent combines four chro- able to that in female meiosis (Sved et al., 1991). Mitotic
matids. Complex interaction of homologs within bivalents at crossing-over between the centromere and any locus results in
pachytene results in reciprocal exchange of genetic material two daughter cells homozygous for the entire segment distal
between nonsister chromatids, a process referred to as cross- to the site of exchange (Figure 2). This may have a strong
ing-over (Figure 1). Observation of corresponding meiotic negative impact if the initial cell was heterozygous for some
stages (diplotene–diakinesis) reveals characteristic w-like detrimental alleles (see Applied Aspects).
configurations called chiasmata. These configurations are
interpreted as cytological manifestation of crossing-over.
Nonreciprocal Exchange
Chiasmata were observed and interpreted as a result of an
unknown process of rejoining genetic material at the end of Normally, meiotic and mitotic crossovers are based on re-
the nineteenth century, a decade before the discovery of partial ciprocal physical exchange of genetic material between the
linkage between genes in plants and two decades before parental chromosomes. Studies of recombination in fungi
Morgan’s studies on crossing-over on Drosophila. (Sordaria, Neurospora, and Ascobolus) have revealed a process of
nonreciprocal (unidirectional) transfer of genetic information
called gene conversion that accompanies reciprocal exchange
Nonhomologous Chromosome Segregation
(crossing-over). Gene conversion is based on DNA repair of
After crossing-over occurs, chromosomes of each bivalent seg- mismatched strands along the ‘‘recombination tract.’’ The
regate independently relative to chromosomes of other biva- demonstration of this process was due to the possibility of
lents in two consequent meiotic divisions. These two processes, classifying marker segregation (spore color) of ordered spores
crossing-over and recombination of nonhomologs, are charac- in asci resulting from individual meiotic cells in hybrids be-
teristic of all sexual eukaryotes. With independent segregation, tween colored and wild-type strains. In addition to the ex-
the ratio of parental and nonparental combinations of pected regular 4 : 4 (colored : noncolored) octads, aberrant

Encyclopedia of Biodiversity, Volume 6 http://dx.doi.org/10.1016/B978-0-12-384719-5.00120-9 353

354 Recombination

12 34 12 34

2−3

2−4

(a) F1 P1 P2 R1 R2 (b) F1 P1 R1 R2 R3
Figure 1 Crossing-over at the four-chromatid stage: (a) Single exchange resulting in two parental (P1 and P2) and two recombinant (R1 and
R2) chromosomes. (b) Double exchange involving nonsister chromatids, 2 and 3 at the crossover point in the short arm and 2–4 in the long arm
– the result is one nonrecombinant (P1), two single-recombinant (R1 and R2), and one double-recombinant (R2) chromosomes. Note that if the
same two nonsister chromatids participate in the exchange at both positions (e.g., 1–3 and 1–3) then two parental and two double-recombinant
products will appear. Meiotic configurations involving all four chromatids in the double exchange (e.g., 1–4 at the first site and 2–3 at the
second) will result in four single-recombination products.

M m M m M M m m
(a) F1 (b) (c) (d)

Figure 2 Segmental homozygosity resulting from mitotic crossing-over: (a) A diploid heterozygote for a marker locus (M/m); (b) after DNA
duplication each homolog is represented by two sister chromatids (the position of mitotic crossing-over between nonsister chromatids is marked
by  ); (c) the structure of the homologs after crossing-over; and (d) the daughter cells resulting from mitosis are homozygous for the entire
segment distal to the point of exchange.

ratios 2 : 6, 6 : 2, 3 : 5, and 5 : 3 are also observed, reflecting recombination. In particular, numerous families of repeated
nonreciprocal exchange (see Genetic Control and Mechanisms elements that are spread in eukaryotic genomes may partici-
of Recombination). The availability of sequence data for genes pate in ectopic recombination. Retrotransposons can be es-
of several higher eukaryotic organisms (wild barley, maize, pecially active, either from nonhomologous chromosomes or
and Drosophila) allowed estimating population genetic par- from different sites of the same chromosome (giving rise to
ameters characterizing the rate of reciprocal (crossing-over) inter- and intrachromosomal ectopic recombination), re-
and nonreciprocal (gene conversion) recombination events, as sulting in chromosome rearrangements. In fact, the frequency
well as the contribution of recombination and mutation to of such events is very low, and it is unclear which mechanisms
sequence variation at the population level. Both in outcrossing reduce this danger. One possibility may be a strong bias of
and self-fertilizing species, the contribution of gene con- recombination interaction in favor of gene conversion if such
version to total recombination may be twice as much as ectopic contacts occur. Ectopic recombination may also occur
compared to crossing-over. Recombination, in total, contrib- between such partners as chromosome and a plasmid, such as
utes to sequence diversity no less than the mutation process a 2-mm plasmid in yeast harboring the gene for site-specific
(Morrell et al., 2006). recombinase Flp and the tumor-inducing T-DNA segment of
the Ti plasmid of Agrobacterium tumefacience interacting with
the host plant chromosome.
Ectopic Recombination
Crossing-over occurring between homologous sequences of
homologous chromosomes generates new genetic variation
Illegitimate Recombination and Horizontal Transfer
without changing genome structure. However, crossing-over
may also involve homologous sequences of nonhomologous Unlike homologous and site-specific ectopic recombination,
parts of the genome. This process is called ectopic illegitimate recombination is a process of joining DNA
 Recombination 355

molecules that have only a small sequence homology. Illegit- from the observed proportions of recombinant and non-
imate recombination is common in nature, especially in recombinant individuals coincide with map distances only for
interactions between retroelements and host genome DNA short segments when multiple exchanges can be ignored.
and in some host–parasite systems. One of the most common Therefore, transformation of experimentally estimated rf val-
examples is the previously mentioned interaction between ues into map distances (x) is needed. Relations of the form
bacteria, such as A. tumefasciens, and their host plants medi- rf¼ rf(x) are referred to as mapping functions.
ated by the bacterial plasmid transformation (see Ectopic Re-
combination). Although this process can be considered as a
Genetic Interference and Mapping Functions
‘‘physiological’’ rather than a genetic transfer, it may also have
evolutionary consequences as one of the mediators of hori- Multiple exchanges are relatively rare among eukaryotes. These
zontal gene transfer. Illegitimate recombination is involved in vary between one and three for the majority of higher or-
many situations of foreign DNA integration in direct trans- ganisms, with five or six rarely occurring, although values up
formation experiments aimed at obtaining genetically engin- to 12 have been reported (Korol et al., 1994). In Drosophila,
eered (transgenic) organisms. the frequency of tetrads with four or more crossovers is ex-
tremely low (usually much less than 1%).

Extranuclear Recombination, Exchanges between Cell Interference: Its Measurement and Basic Properties
Compartments Consider two adjacent marked segments m1–m2 and m2–m3
Recombination was proven to affect the organization of mito- with recombination rates rf1 and rf2. If exchanges in adjacent
chondrial DNA. Mitochondrial recombination was detected in segments occur independently, the probability of a double
Drosophila, mussels, and various vertebrate species including exchange within m1–m3 can be determined as rf12 ¼ rf1rf2.
fishes, birds, and mammals (Tsaousis et al., 2005). Convincing Under the same assumption, recombination frequency rf3
evidence of widespread recombination in mitochondrial gen- between loci m1 and m3 is given by
omes has been obtained for plants. In fact, a specific substrate rf3 ¼ rf1 ð1  rf2 Þ þ rf2 ð1  rf1 Þ ¼ rf1 þ rf2  2rf12
for frequent mitochondrial recombination was discovered and
sequenced in a few plant species. Especially interesting from the Much evidence indicates that the observed frequency of
evolutionary point of view are putative DNA exchanges double crossovers usually differs from the expected one, a
between different cell compartments: nuclear–mitochondrial, phenomenon termed as genetic interference. The degree of
nuclear–chloroplast, and mitochondrial–chloroplast. Many interference is measured by the coefficient of coincidence (c):
such sequences demonstrating ‘‘chloroplast - mitochondrial’’
transfer were obtained for cereals. The existence of mitochon- rf12 ðobservedÞ rf12 ðobservedÞ
c¼ ¼
drial recombination limits the validity of evolutionary re- rf12 ðexpectedÞ rf1 rf2
constructions based on a certain part rather than on the entire
mitochondrial genome. The existence of interference was shown both genetically
and cytologically. It is ubiquitous in eukaryotes. To date, only
in a few species (Shizosaccharomyces pombe, Aspergillus nidulans,
Tetrahymena thermophila, and possibly Ustilago maydis) has re-
Basic Features of Reciprocal Homologous
combination been found to proceed with no interference
Recombination (Crossing-Over)
(Loidl and Scherthan, 2004, and references therein). If the
occurrence of crossing-over in one segment reduces the
Recombination Rate and Map Distance
probability of exchange in an adjacent segment (co1), then it
The average number of crossovers in a given segment repre- is positive interference; when one exchange increases the fre-
sents its genetic length. Correspondingly, genetic distance be- quency of another (c41), it is negative interference. The
tween two loci can be defined as the average number of nematode Caenorhabditis elegans is an example of extreme
recombination events occurring in the segment. Therefore, for positive interference: a single crossover formed per chromo-
loci flanking a unit length segment one exchange per meiosis some suppresses formation of additional exchanges.
is expected on average, resulting in 50% recombinant gametes. Two types of interference have been distinguished: chiasma
This notion of genetic distance serves as a basis for gene interference, in which an exchange occurring in a segment
mapping. Two fundamental facts make genetic mapping affects the probability of another one occurring in an adjacent
possible: (1) linear organization of genetic material in segment, and chromatid interference, which determines the
chromosomes and (2) increasing recombination rate between pattern of chromatid configurations in successive pairs of
loci with the increase in their physical distance. Map distance chiasmata in a bivalent (i.e., the proportions of two-, three-,
is of additive nature: For any subdivision of a segment, the and four-strand double exchanges) (Figure 1). In the absence
segment length is the sum of lengths of its subintervals. of chromatid interference, the ratio should be 1 : 2 : 1. Evi-
A limitation of map distance as a measure is that it cannot be dence of chromatid interference is scarce. Chiasma inter-
estimated directly: an odd number of exchanges in a segment ference determines the pattern of distribution between
marked by two loci leads to the appearance, in the progeny, of successive exchanges along a chromosome. The amount of
the same recombinant genotypes as those resulting from one interference within a segment depends on its distance from
exchange, whereas an even number results in no observable the centromere. Near the centromere, interference between
recombinants. Recombination frequencies (rfs) estimated exchanges in the adjacent segments of the same chromosome
356 Recombination

arm is maximal. It is generally believed that no interference The Effects of Sex

exists across the centromere, although many exclusions
By the Haldane–Huxley rule, if crossing-over is strongly sup-
are known.
pressed or lacking in meiosis of one of the sexes, then that sex
is invariably the heterogametic one. The effect of sex on rf has
Mapping Functions
been established in a variety of animal species, with this rule
Positive interference (co1) is a widespread phenomenon. It
holding in all cases of alternative (all-or-none type) sex dif-
was found that in many cases the approximation c(rf) ¼ 2rf
ferences. Provided crossing-over occurs normally during the
(Kosambi interference) fits the data well (i.e., cE0 at small
formation of both male and female gametes, then rf in the
distances and cE1 at large distances). The corresponding
heterogametic sex may be either lower or higher than that in
mapping function
the homogametic one, with the sign and magnitude of sex
rf ¼ 0:5tanhð2xÞ or x ¼ 0:25 lnðð1 þ 2rf Þ=ð1  2rf ÞÞ difference in rf being segment specific. Sex differences in re-
combination have also been revealed by cytological analysis.
appeared to be a good approximation for Drosophila, rice, Higher female recombination was observed in many species.
mouse, and other eukaryotes. The formula for rf values in In humans, the total length of female maps is nearly twice as
adjacent intervals, corresponding to Kosambi’s function, takes long as that from male maps; in some regions (mainly sub-
the following form: telomeric ones), the opposite tendency is observed. Most of
the examined genomic regions in maize exhibit higher male
rf1 þ rf2
rf3 ¼ recombination or no difference; the extent of the difference (if
1 þ 4rf1 rf2
it exists) may strongly depend on the genotype. A substantially
Situations with no interference can be represented by the higher crossing-over rate in male compared to female meiosis
mapping function proposed by J. Haldane: has been found in Arabidopsis (Korol et al., 1994). This is also

true for some animals exhibiting a higher degree of linkage
rf ðxÞ ¼ 0:5 1 þ e2x and/or more localized pattern of exchange distribution in the
homogametic sex than in the heterogametic one, such as
Some results indicate that negative interference (i.e., an
marsupials (Trivers, 1988).
excess of double exchanges over the level expected with the
The mechanisms responsible for differences in rf between
assumption of independence), may also occur. Such data were
male and female meiosis are unknown. In most animals, sex is
observed for Drosophila and some plants, fungi, and viruses.
determined by a certain chromosome. Thus, it can be sug-
One should take into account that values of c41 can also
gested that the difference between male and female meiosis is
result from marker scoring errors and data pooling, owing to
controlled by genes located on this chromosome. Alter-
heterogeneity of individual rf values.
natively, sex chromosome can canalize the development of
general physiological (biochemical) sex differences (e.g., in
Distribution of Crossovers within Eukaryotic Chromosomes the hormonal status) in a toggle-like manner, with specific
recombination features of male and female meiosis being a
The efficiency of crossing-over strongly depends on between-
secondary effect of the main regulator. The latter suggestion is
and within-chromosome patterns of exchange distribution in
supported by data from the fish Oryzias latipes in which the
the nucleus. Here, two different (albeit related) problems are
heterogametic sex is represented by XY males (Yamamoto,
generally distinguished: (1) crossover distribution in relation
1961). When males XY were transformed into functional XY
to major components of the chromosome, that is, the cen-
females by a hormonal treatment, rf increased by a factor of
tromere, telomeres, heterochromatin (intensively stained
5 compared to the level in normal males. In humans, it was
cytological segments of chromosomes – gene-poor regions),
found that the distribution of sex difference in rf is related to
and euchromatin (faintly stained gene-rich regions), and (2)
genomic imprinting manifested in region-specific differential
relative positions of crossovers with respect to one another.
modification (methylation) of parental alleles. We speculated
These questions were thoroughly studied based on scoring
that gametic selection may be an important contributing fac-
recombination events between multiple markers and com-
tor to the evolution of sex differences in recombination (Korol
plemented by cytological observations. A wide range of spatial
et al., 1994). Lenormand and Dutheil (2005) supported this
variation in exchange frequency was revealed – from long
hypothesis by a detailed analysis of available evidence on
stretches of heterochromatin in which crossing-over is almost
many species.
blocked to short segments containing recombination hot-
spots (see Recombination Signals, Hotspots of Recombination,
and Transcription). Heterochromatic blocks affect the distri-
The Effects of Environmental Conditions
bution of crossovers in the entire genome. Recombination
non uniformity of the chromosomes was first established in If diversity is essential to population survival, it would be
Drosophila and then confirmed in a wide range of organisms. highly adaptive for organisms to increase genetic variation in
Thus, in the B genome of wheat the physical length of DNA challenging (stressful) environments and decrease it under
per exchange varies more than 150-fold along the chromo- optimal conditions (Korol et al., 1994). Such regulation
some (Lukaszewski and Curtis, 1993). In the centromeric re- should increase the chances of population survival without
gion of the human X chromosome, the recombination rate is dissipation of the accumulated hidden genetic variation in
at least eight-fold lower than the average rate of female re- favorable times. An abundance of evidence on environmental
combination on the X chromosome. modulation was found for the main sources of variation:
 Recombination 357

mutation, chromosomal rearrangements, transposition of attack caused increased somatic recombination in leaves of the
mobile elements, and recombination. The environment can treated plants and of several subsequent generations (Molinier
affect recombination in two ways. Firstly, it can affect re- et al., 2006). These findings were interpreted by the authors as
combination indirectly when changes in the frequency of al- a manifestation of the epigenetic effect. Using ten different
leles at recombination controlling loci depend, through a physical and chemical stress treatments, Pecinka et al. (2009)
feedback mechanism, on changes in the frequency of the se- tried to repeat these results on several lines of Arabidopsis by
lected (fitness-related) loci (see Evolution of Recombination). scoring for somatic recombination in the treated (S0) and
Secondly, it can affect recombination directly when in re- subsequent generations of selfing (S1 and S2). The results
sponse to environmental factors, the organism modifies were negative: despite highly significant induction observed in
(within genetically determined limits) the value of rf, that is, S0, only small and nonconsistent effects were observed in S1
when not rf but the norm of reaction with respect to rf is and S2, suggesting that the transgeneration effect in Arabidopsis
genetically determined. The fact of direct influence of the en- is not a rule and occurs only under very special conditions.
vironment on recombination has long been known in genet- Obviously, a similar question of reproducibility should also be
ics. As early as 1917, H Plough observed in Drosophila a addressed for stress memory related to meiotic effects.
U-shaped dependence of rf on temperature: Recombination
was minimal at optimum temperature and increased with the Feedback Regulation of Genetic Variation
departure of temperature from the optimum. Similar data If stress may be recombinogenic, it is natural to assume that
indicating an increase in recombination level with deterior- the degree of the effect should depend on the intensity of the
ating environment have been reported in many works. Ex- stress, that is, the degree to which an individual is stressed. In
treme conditions also tend to reduce interference. other words, induced changes in recombination in mala-
dapted genotypes should be higher than in well adapted ones
Ontogenetic Memory and Transgeneration Effect (Zhuchenko and Korol, 1983). Experimental evidence for
A common belief is that environmental factors may affect re- the last principle has been obtained in the author’s lab for
combination only when the treatment occurs during a rela- Drosophila and tomato (reviewed by Korol et al., 1994). In
tively short period between premeiotic S-period and prophase Drosophila, the increase in male recombination induced by
of the first meiotic division. Therefore, in many organisms the heat treatment was negatively correlated with flies’ resistance
sensitive stage may comprise a very small part of the life cycle. to high temperatures. In tests with tomato varieties and their
Does it mean that environmental stresses on earlier stages of F1 hybrids, low temperature did not increase the chiasma
the development cannot affect recombination? Some limited frequency in cold-resistant varieties and in their F1 hybrids,
evidence available for a few organisms indicate that just the whereas the reverse was true for cold-sensitive genotypes. By
opposite may be true. Neel (1941) studied the effect of severe contrast, elevated temperatures caused an increase in re-
starvation of Drosophila larvae on meiotic recombination. An combination only in heat-sensitive genotypes. Populations
increased rate of recombination in chromosome 3 was ob- may benefit from such a feedback mechanism by generating
served during the whole lifetime of the females, representing additional variation to meet new, challenging conditions
some memory of starvation-induced changes in germline cells while preserving previously evolved adaptive gene combin-
during many mitotic cell divisions until meiosis. Zhuchenko ations, when there is no need for additional variation, thereby
et al. (1983) showed that temperature stress applied to reducing the recombination load. In the last decade, the
Drosophila F1 hybrids at the embryonic stages can affect the concept of plastic- and fitness-associated recombination has
rate of meiotic recombination during the entire lifetime of F1 become the subject of theoretical studies (Hadany and Beker,
females, implying that stress-induced changes survived many 2003). Theoretical models and growing evidence indicate that
divisions of germline cells until meiosis. In S. pombe, similar strategies may also be relevant to other systems af-
X-irradiation caused an increase in the recombination rate fecting genetic variation: transposition of mobile genetic
which persisted for 8–10 cell generations after irradiation elements, DNA repair, mutation rate, and sexual versus asexual
(Takeda et al., 2008). Limited data indicate that the memory of reproduction (e.g., Lupu et al., 2006; Agrawal and Wang, 2008;
stress-induced changes may extend over generations. In ex- Schoustra et al., 2010).
periments of Landner (1970) with Neurospora, temperature
treatments applied as early as 3–4 days before fertilization,
affected the rate of meiotic recombination, that is, in the next
Genetic Control and Mechanisms of Recombination
generation. A similar transgeneration effect was also found in
the aforementioned study of Zhuchenko et al. (1983) with
Selection for Changed Recombination, Genetic Modifiers of
Drosophila: changes in meiotic recombination rate were re-
Recombination
corded in F1 females whose wild-type mothers were subjected
to heat treatment of 36 1C for 12 h. Selection for altered recombination proved an important tool
Somatic recombination is considered a general plant re- in analyzing genetic control of recombination. Such experi-
sponse to environmental stresses like UV, high temperature, ments have been conducted on dozens of organisms,
herbicides and other factors, but the usual expectation is that including Drosophila, flour beetle Tribolium castaneum, grass-
stress-induced stimulation should disappear together with hopper Schistocerca gregaria, silkworm Bombix mori, lima bean,
the inducing factor (Pecinka et al., 2009). Experiments with and fungi Neurospora and Schizophyllum (Korol et al., 1994).
Arabidopsis have shown that this may not be the case: treat- These experiments complemented the numerous findings on
ment with UV light or with a chemical mimicking pathogen the existence of a considerable amount of genetic variation in
358 Recombination

rf in natural and laboratory populations (see Intraspecific (e.g., from a prokaryote to eukaryote). However, this does not
Variation for the Rate of Recombination) available for selec- exclude high interspecific differences in many important fea-
tion to act on. The effectiveness of directional selection for tures of recombination (Korol et al., 1994).
altered rf has been shown to depend on the segment under Comparison of recombination in meiotic mutants and in
study, the size, structure, and origin of the start population; the corresponding wild-type genotypes provides an interesting
breeding system; the estimation procedure; and the intensity conclusion. It appears that genomic distribution of exchanges
and duration of selection. Genetic analysis of the accumulated in normal meiosis is more restricted, less random, than in
differences between selected lines enabled the question of the meiosis altered by mutations, indicating that these restrictions
genetic basis of variation in recombination to be addressed. In are largely a result of evolutionary adjustment of recombin-
as few as 10 generations, divergent selection for rf in the p–Y ation. Meiotic mutants tend to lessen the restrictions, and
region of silkworm chromosome 2 succeeded to produce lines regions normally excluded from exchanges become involved
with rf¼ 37–39% and 5–7%, starting from 25.6%. The ob- in crossing-over. Simultaneously, a reduction of interference is
tained evidence suggests that recombination (frequency and also observed despite a general tightening of linkage (Bhagat
distribution) in eukaryotic genomes is under complex control et al., 2004). In meiotic mutants, the correspondence between
of polymorphic modifiers with small to moderate effect. The physical and recombination length of chromosome segments
rate of recombination in a given region may depend on both becomes closer. In many cases, mei mutations referred to as
linked modifiers and genes located on other chromosomes. hyper-rec can significantly enhance crossing-over (and not
only in specific regions) or increase the rate of multiple ex-
changes, intragenic recombination, and/or mitotic crossing-
over manyfold.
Meiotic Mutants as an Analytical Tool in Recombination
The meiotic system manifests a close connection between
Studies: Overlapping of DNA Repair, Recombination, and
homologous conjugation, crossing-over, and regulation of
Segregation Systems
chromosome segregation. This association is invariable for
The first discovered point mutation (in 1922) affecting re- normal meiosis and served as a basis for the view that regu-
combination was Drosophila gene c(3)G, which almost com- lation of chromosome segregation is one of the main func-
pletely blocks crossing-over in females. During subsequent tions of crossing-over. Meiotic mutants were isolated where
years, many other genes with strong effect on different meiotic the two phenomena, segregation and recombination, are not
steps and recombination were isolated in eukaryotes (mainly coupled. For instance, crossing-over is normally almost com-
in fungi, Drosophila, maize, and recently Arabidopsis and the pletely suppressed in chromosome 4 of Drosophila melanoga-
nematode Caenorhabditis elegans). The availability of diverse ster. Some mei-mutants with reduced recombination and/or
mutants affecting the coordinated steps of meiosis and re- altered exchange distribution within large chromosomes ex-
combination allowed a detailed cytogenetic characterization hibit crossing-over within chromosome 4 and disturbed seg-
of meiosis (reviewed by Heyer et al., 2010). These include regation of this chromosome (Sandler and Szauter, 1978).
commitment to meiosis, regulation of major steps in
chromosome conjugation and formation of the synaptonemal
complex (SC) (see Ultrastructure of Recombination: SC and
Ultrastructure of Recombination: SC and RNs
Recombination Nodules (RN)), DNA breaks formation and
strand exchange, chiasma maintenance, and chromosome Electron microscopy reveals meiosis-specific organization of
disjunction. The broad network of recombination–meiosis chromosome associations at the meiotic prophase I. A three-
mutants identified in yeast and other fungi, Drosophila, C. part proteinaceous structure, called SC, is initiated at the
elegans, and Arabidopsis serves as a basis for cloning re- zygotene stage, maintains the paired homologs together dur-
combination genes and studying the molecular mechanisms ing pachytene, and disappears at the diplotene. Short stretches
of recombination. Numerous results were obtained on this of DNA of the homologs are involved in the central part of SC
rich material showing a deep overlapping of pathways related where intimate molecular pairing takes place. SCs are con-
to DNA recombination and DNA repair. Thus, mutations for sidered as structures that affect the number and distribution of
loci controlling meiosis and recombination also tend to ex- crossovers and convert crossovers into functional chiasmata.
hibit disturbed repair functions. Such a duality seems to be an In many species (e.g., human, mouse, zebrafish, and Arabi-
ancient phenomenon. In prokaryotes, the best example is the dopsis), the difference in total lengths of SCs in male and
RecA protein of Escherichia coli that is involved in early stages female meiocytes is correlated with the difference between
of recombination and, simultaneously, in induced SOS repair. male and female recombination map lengths. Small electron-
Recombination is also strongly affected by mutations in genes dense bodies, RNs, were detected within the central part of the
controlling other enzymatic steps in DNA metabolism, for SC. Early RNs that appear before pachytene are more or less
example, replication. Genes for DNA repair and recombin- evenly distributed along the bivalent and are supposed to
ation exhibit a high degree of sequence, structural, or func- participate in synapsis. Late, or pachytene, RNs seem to rep-
tional similarity throughout life. This conservation was resent recombination enzymes associated with exchanging
established using a combination of genetic and cytological DNA molecules. The distribution of late RNs corresponds well
methods with molecular cloning, sequencing, and genetic to chiasma distribution and is believed to represent the sites
transformation. Thus, defects in the recombination–repair for recombination. Numerous meiotic mutants were tested
system of one species may be complemented by a gene with respect to their ultrastructural and molecular phenotype
transferred from another species, sometimes a distant one – the morphology and biochemistry of SC. An interesting fact
 Recombination 359

is that only in a few organisms was meiotic recombination DSB-dependent recombination in yeast may be indicative for a
normally found without SC (S. pombe, A. nidulans, T. thermo- shared DSB mechanism of recombination initiation, as ex-
phila, and possibly U. maydis). However, in many cases SC was emplified by the Drosophila mei-W68 gene and yeast Spo11
found in achiamatic meiosis (with no recombination). In (McKim and Hayashi-Hagihara, 1998). However, new evi-
S. pombe and A. nidulans with recombination occurring with- dence obtained on different organisms points to limitations of
out SC, the frequency of exchanges per bivalent is high, and the Szostak model that focused only on the crossover-for-
especially interesting is the absence of interference. Thus, the mation pathway of recombination, whereas new models take
absence of SC may allow for more reshuffling of the homo- into account the noncrossover pathway (potentially leading to
logs. Analysis of SC appeared to be an important tool in gene conversion) and explain details of DSB initiation in both
determining the causes of altered fertility in farm mammals. pathways and their interference-sensitive and -insensitive
classes (reviewed by Cromie and Smith, 2007; Heyer et al.,
2010). In particular, recent studies on the fission yeast S. pombe
point to the existence of a novel mechanism for the regulation
Molecular Mechanisms of Recombination
of crossover distribution along the chromosome – controlled
When considering molecular mechanisms of recombination repair of DSBs based on partner choice model (differential
in eukaryotes, a broader spectrum of basic processes involved interaction with the sister chromatid or with the homologous
may also be addressed, including chromosome conjugation chromosome) (Hyppa and Smith, 2010). This mechanism
and segregation. This is partly due to the prevailing experi- explains the rather high stability of the crossover rate per DNA
mental methodology based on analyzing cytogenetic and physical unit length despite high variation of the DSB rate
molecular effects of numerous mutants defective for meiosis– along the genome. The molecular mechanisms of DSB repair
recombination and DNA repair. Historically, the studies of involved in meiotic and mitotic recombination differ in many
molecular mechanisms of recombination in prokaryotes aspects (for a recent review see Andersen and Sekelsky, 2010).
served as the main source of models to be tested in eukaryotes.
The basic enzymatic activities involved in prokaryote re-
combination include those of endo- and exonucleases, strand
Recombination Signals, Hotspots of Recombination, and
transfer, DNA unwinding, ligation, and resolving Holliday’s
Transcription
heteroduplex structure. Eukaryotic analogs were found and
(are to be) characterized in the best-studied eukaryotes, such Although recombination occurs all over the genome, the ac-
as yeast, Drosophila, Arabidopsis, and mammals. Clearly, eu- cumulated data points to the existence of hotspots and cold
karyotic recombination systems are supposed to be more spots of recombination. In certain cases, highly specific
complex due to additional problems related to meiosis, such interactions were found between recombination–repair mu-
as homologous pairing and segregation. tations and ‘‘signal’’ DNA sequences resulting in a sharp in-
The classical concept of molecular recombination included crease in rf within defined genomic segments. The best-known
the following steps: (1) intimate pairing of homologs; (2) recombination signal is the octamere 50 GCTGGTGG-30 of
formation of single-stand DNA nicks at homologous sites and E. coli referred to as w element. Heptamer 50 ATGACGT-30 A is
local denaturation of double-stranded DNA; (3) formation of known as a meiotic recombination hotspot of fission yeast
recombination intermediate (Holliday junction) by reciprocal S. pombe. A few recombination hotspots have been character-
strand transfer and renaturation, its migration from the initi- ized in mammals, with the major histocompatibility complex
ation site, and heteroduplex formation (with local mis- being one of the best-known examples. In humans, 8-oxo-
matched DNA due to heterozygosity of the parental guanine, resulting from endogenous oxidative stress, is a
chromosomes); and (4) resolution of the Holliday structure, hotspot signal spread over thousands of positions in the
either after or before isomerization. In the first case, the con- genome. In humans and Drosophila, a 13-bp sequence motif
tent from both sides of the considered regions remains un- CCNCCNTNNCCNC is associated with hotspots of fine-scale
changed, whereas in the second case it results in reciprocal recombination (Stevison and Noor, 2010). Recent studies on
exchange. In both cases, the mismatched positions should humans and mice showed that a major player for hotspot
undergo correction. Depending on the strand used as a tem- specification is PRDM9 gene (Baudat et al., 2010). PRDM9 is
plate in the repair process, two alternatives will be observed: predicted to recognize the mentioned 13-mer motif. Allelic
recovering the initial state and nonreciprocal information variants of PRDM9 zinc finger region are significantly associ-
transfer (gene conversion). ated with variability in genome-wide hotspot usage among
A new understanding of eukaryotic recombination mech- humans, indicating that the binding of PRDM9 protein to
anisms resulted from the discovery of the leading role of specific DNA sequences controls the initiation of recombin-
double-strand DNA breaks (DSBs) in yeast (Szostak et al., ation at specific locations in the genome. The minisatellite
1983) that became the leading concept (Szostak model) of structure of the PRDM9 zinc finger region points to its po-
crossover formation for more than two decades. Details on tential to increase variability leading to new PRDM9 variants
different pathways related to this process were also studied in with new DNA binding specificities, hence new genome-wide
mammalian cells, Drosophila, and plants based mainly on distributed hotspots.
recombination–repair mutants. Another strategy is to study In eukaryotes, mobile elements may act as recombination
specially engineered genetic models with a modified re- enhancers. Thus, with identical locations on homologous
combination system (Shalev et al., 1999). Homology of chromosomes, a pair of P elements of D. melanogaster can
recombination genes in one species to genes controlling induce recombination at a frequency of 20% in males in
360 Recombination

which recombination is normally absent (Sved et al., 1991). partially overlapping environmental regulation of transcrip-
This effect serves as a basis for an efficient method of fine tion and recombination (Koren et al., 2002). In humans, in-
genetic mapping, in which P elements are used to promote creased meiotic recombination coincides with high GC and
recombination at their insertion sites. However, usually Dros- gene-rich regions, but recombination rate within genes is
ophila genome regions with increased density of transposable lower compared to their upstream and downstream regions
elements display reduced recombination. In general, both (Coop et al., 2008). Surprisingly, a negative correlation was
positive and negative correlations between the abundance of found between crossover rate and gene expression in meiosis
transposable elements and the recombination rate were found (McVicker and Green, 2010).
in eukaryotic organisms. The accumulating evidence indicates In general, the available information shows high intrage-
that recombination hotspots are associated with DSB sites, nomic heterogeneity and nonrandomness of recombination.
depend on chromatin organization, and are influenced by These are caused by a hierarchical control system that involves
epigenetic changes. interaction of a complex web of recombination–repair genes
In the 1960s and 1970s, it was speculated that initiation of with peculiarities of macro- and micro-organization of the
recombination is attached to transcription promoters genetic material (Table 1). The important role of gene distri-
(Whitehouse, 1966). In moving from lower to higher organ- bution in this regulation indicates its adaptive role in the
isms, the recombination density over the genome is reduced evolution of genetic systems.
by three or four orders of magnitude with no marked trend in
structural genes. Thuriaux (1977) attributed this paradox to
preferential localization of recombination in structural genes.
Biodiversity of Recombination Systems
It was suggested that the requirement for common specific
changes in chromatin organization may cause an overlapping
Evolutionary Trends in Genomewise Recombination Rate
in control of transcription, recombination, repair, and mu-
tation (Whitehouse, 1966; Thuriaux, 1977; Korol et al., 1994; A considerable (by several orders of magnitude) decrease in rf
Laurencon et al., 1997). The evidence from many eukaryotes per unit physical length of DNA is observed when moving
indicates that regions with high rf correspond to gene-rich from prokaryotes and lower eukaryotes to higher organisms
regions (Lichten and Goldman, 1995), although an opposite (Thuriaux, 1977). Even among eukaryotes, the density of re-
trend was found in C. elegans. The results obtained during the combination events within the genome varies widely
past decades for yeast (Saccharomyces cerevisiae and S. pombe) (Table 2). The genetic length of the genome varies less and
support the idea of recombination–transcription relation- strongly depends on haploid chromosome number. A rela-
ships, but also show their complicated nature. In particular, tively low correlation is displayed by fine-scale recombination
genes near hotspots display a higher expression level, but not rates in genomic regions with parallel syntenic blocks of
in meiosis when their expression is lower. It appeared that different species pointing to fast evolution of recombination
upstream gene sequences acting as transcriptional control sites hot spots. This is reflected in human-chimpanzee comparisons
may also play a role of inducible recombination hotspots. This and even among different human populations. But at larger
implies that environmental effects on recombination reflect scales, including the genome level, recombination rates are

Table 1 Hierarchical control of recombination events

System effects:
• Control of the nucleus as a whole, manifested in the classical interchromosomal effect of rearrangements, as well as effects of eu- and aneyploidy
and supernumerary (B) chromosomes
• Effect of chromosome size
• Regulation based on cytonuclear interactions
• Sex differences in recombination frequency and distribution
• Effect of environmental conditions (stress) and age
• Position of euchromatin (as main target for recombination) relative to the centromere, telomeres, and heterochromatic blocks
DNA sequence organization of the target region:
• Heterozygosity for micro- and macroinversions, deletions, and translocations
• Microsite organization of DNA sequences including distribution of specific regulatory sequences like recombination hotspots, micro-, and
minisatellites
• Distribution of epigenetically modified (e.g., methylated) DNA stretches
• Effects of mobile genetic elements
• Distribution of gene-rich islands
Genes controlling DNA metabolism:
• Genome-, chromosome-, and segment-specific effects of major rec-genes of the coarse control system affecting the basic steps in recombination
mechanics (pairing, DNA strand exchanges, recombination repair)
• Segment-specific regulation of crossover frequency by genes of the fine control system with relatively small effects of individual components
• Genes encoding active transposases
Interaction of factors both within and between the foregoing groups plays an important role in observed patterns of recombination frequency and distribution within genome and its
variation within and between species.
 Recombination 361

Table 2 Recombination characteristics of some eukaryotic genomes

Species Genome length (cM) Haploid number (n) Chromosome length (cM) kb:cM ratio

Magnaporthe grisea 600 7 85 80

Pinus taeda 1700 12 140 12,000
Arabidopsis thaliana 600 5 120 230
Brassica rapa 1900 10 190 280
Fhaseolus vulgaris 1200 11 110 530
Hordeum spontaneum 1200 7 170 4000
Lycopersicum esculentum 1400 12 120 900
Drosophila melanogaster 300 4 75 570
Bombix mori 3400 28 120 130
Tribolium castaneum 600 10 60 350
Apis mellifera 3700 16 230 45
Danio rerio 2900 25 120 590
Gallus gallus 3200 29 110 310
Monodelphis domestica 650 9 70 5400
Mus musculus 1400 20 70 1800
Canis familiaris 2100 39 54 1200
Homo sapiens 3600 23 160 800

well correlated. Higher ‘‘regularity’’ of recombination due to individual rf values in the ml-ru segment of chromosome 3
frequent outcrossing, the obligatory occurrence of at least one varied between 2.2% and 42.8% (i.e., 20-fold). Differences
exchange per bivalent, and increasing genome size are con- between maize plants in rf in the sh-su-wx segment increased
sidered to determine the evolutionary importance of re- as the degree of relatedness decreased and there was cor-
combination in higher sexual eukaryotes. In lower organisms, respondence between micro- and macrosporogenesis. Poly-
recombination was supposed to play a less important role due morphism for B chromosomes (also known as supernumerary
to a slight chance of sexual contacts between individuals chromosomes) and heterochromatic knobs and their inter-
(Maynard Smith et al., 1991). Consequently, natural popu- action play a significant role in the variation of rf in maize. B
lations of such species should manifest a clonal structure, with chromosomes affect pairing and crossing-over in many plants
strong correlation (linkage disequilibrium (LD)) between al- and animals.
leles at different loci. However, this may not be the case in Abundant evidence for intraspecific genetic variation in
some protozoan species, such as malaria Plasmodium falci- crossing-over rate and/or chiasma frequency is available for
parum (Conway et al., 1999). The revealed population pattern many eukaryotes, both lower (Coprinus, Schizophyllum, Neuro-
of P. falciparum includes high polymorphism and a rapid spora, Saccharomyces, and Sordaria) and higher (pearl millet,
decay of LD with distance (reaching independence between pea, jute, ryegrass, barley, wheat, tomato, snails, silkworm,
polymorphic sites of 0.3–1.0 kb) that looks like that of higher grasshoppers, locusts, flour beetle, lizards, mouse, rat, and
sexually reproducing eukaryotes. cattle) (Korol et al., 1994). Human variation in fine-scale re-
combination patterns among individuals was examined based
on dense genome-wide single-nucleotide polymorphism data
in nuclear families (Coop et al., 2008). Similar recombination
Intraspecific Variation for the Rate of Recombination
hot spot usage in male and female meiosis was found com-
The analysis of genetic variation of recombination parameters bined with high heritable variation in the proportion of
allows a deeper insight into patterns, mechanisms, and the crossovers detected in these hotspots. Intraspecific variation in
evolutionary role of recombination. The genetic differences in recombination rate and genomic distribution may be of pri-
crossing-over rate seem to have been originally established by mary importance in a better understanding of the evolution of
Sturtevant in 1913 who discovered modifiers sharply reducing recombination, genetic basis of biodiversity in natural sys-
crossing-over during oogenesis in heterozygous females of tems, optimizing genetic conservation and breeding strategies,
Drosophila. Later, he suggested these to be inversions, a pre- designing experiments on genetic mapping of complex traits,
diction that was subsequently confirmed by cytological an- and map-based cloning.
alysis of salivary gland chromosomes. Among point mutations
affecting recombination, the first to be discovered was c(3)G,
which almost completely blocks crossing-over in Drosophila
Recombination and Domestication Evolution
females. Additional studies demonstrated high genetic vari-
ation in rf in D. melanogaster and other Drosophila species Two questions are of major interest when discussing the
(Korol et al., 1994; see Meiotic Mutants as an Analytical Tool in interface domestication–recombination: (1) How have re-
Recombination Studies: Overlapping of DNA Repair, Re- combination features of the progenitors affected the pattern of
combination, and Segregation Systems). Recombination selection response and selection advance? (2) How has se-
studies in Drosophila ananassae males revealed high variation lection for domestication traits affected the recombination
in crossing-over rate in the second and third chromosomes; system? Despite many theoretical studies on the interaction
362 Recombination

between selection and recombination, the very limited avail- 25–40% of variation in fitness in natural populations is re-
able evidence concerns only a few organisms out of a few generated by crossing-over in one or two generations from a
hundred domesticated plant and animal species. Burt and Bell randomly drawn chromosome pair. Thus, a temporary arrest
(1987) studied the relationship between male chiasma fre- of the mutation process would not result in a significant de-
quency and features of reproductive strategy, longevity, body crease in variation over many generations. But does this mean
size, etc. of more than 30 mammalian species. They found that that populations always harbor enough variation to react to
domesticated animal species exhibit higher rates of exchanges external challenges? The author believes that the answer is
than might be expected based on a general relationship negative; otherwise, it would not be possible to change re-
among these parameters. The authors attributed this difference combination by direct selection for quantitative traits un-
to the advantage of increased recombination that facilitated to related to recombination (see Experimental Evolution of
overcome unwanted correlations between characters in the Recombination).
course of selection. Similar conclusions were achieved in An important question is how recombination affects genetic
comparative analysis of chiasma frequency of domesticated variation. In a population at linkage equilibrium (LD¼ 0),
and wild plant species (Ross-Ibarra, 2004). changes in recombination rate do not affect the amount of
At the stage of primitive selection, recombination was a variation. However, this does not mean that the level of vari-
major factor of genetic variation allowing for a deep and very ation is independent of rf. Indeed, the stability of poly-
fast (on the evolutionary scale) reorganization of domesticated morphism varies with rf: The general trend is a reduction of
organisms. Consequently, this selection pressure could become polymorphism stability with increased recombination. This
a driving force of evolution of increased recombination, at least will occur when one models the dynamics of a diploid popu-
during the first phase of domestication (thousands of gener- lation subjected to stabilizing or directional selection with all
ations). However, at the second phase, after regular agriculture selected loci linked within a block. Stability of polymorphism
had become established with more stable environmental con- for these loci inversely depends on recombination within the
ditions, stabilizing selection had to play an ever-increasing role. block. However, the effect of linkage may be opposite for
This might reverse the direction of recombination evolution, at configurations with multiple blocks of selected loci (Nevo et al.,
least in some species. For example, Williams et al. (1995) es- 1997). Another example is selection (either diploid or haploid)
tablished a higher chiasma frequency in primitive forms of in a fluctuating environment (Korol et al., 1996). The effect of
maize compared to modern industrial lines. recombination on the amount of variation may also result from
Polyploid formation is considered a unique evolutionary a dependence of the steady-state LD on rf: LDa0 is achievable
pathway exploited by both animal and plant organisms. A few (if at all) only at tight linkage and/or strong selection.
of the most important crops, such as wheat, cotton, or potato,
are allopolyploids. The recombination–domestication inter-
face may have very interesting aspects when related to
Recombination as a Source of Evolutionary Novelties
polyploids. These include the evolution of diploid-like
chromosome behavior in meiotic pairing, recombination and Through gene duplications (by unequal exchanges), re-
segregation, the effect of various forms of recombination on combination creates prerequisites for increasing variability. In
the coevolution and sequence polymorphism of genomes fact, it is a general belief that the evolutionary formation of a
sharing one nucleus, and the effect of selection on the fore- new gene by mutation–selection interaction is preceded by the
going processes. Detailed mapping and DNA sequencing ef- duplication of the genetic material. Local duplications may
forts on the polyploid crops combined with comparative result from unequal crossing-over, unequal sister chromatid
evidence on their diploid relatives and/or ancestors may be exchange, polymerase slippage, or replicative transposition.
considered an important way of highlighting the role of re- Positive association between recombination and copy number
combination in evolution and domestication. variation is well documented for various eukaryotes. There is
abundant evidence (from both pro- and eukaryotes) sug-
gesting that structurally (and in many cases functionally) re-
lated genes are linked to one another. These may include
Evolutionary Effects of Recombination
tandemly repeated gene families, gene clusters (with arbitrary
spacing and orientation of the individual genes), and dis-
Recombination and Genetic Variation for Adaptation to
persed distribution of the family members across the genome.
Varying Environment
Different mechanisms were proposed for allele diversification
Producing variation is considered one of the major evo- after duplication. For example, a new allele may be produced
lutionary functions of recombination that is supposed to assist by intrachromosomal homologous recombination mediated
in adaptation to both spatial and temporal environmental by DNA sequences within the duplicated elements. The cre-
changes. Abundant genetic variation manifested by natural ative role of different forms of recombination (unequal
population is a result of a complex interaction between a few crossing-over, gene conversion, reciprocal exchange, and ec-
factors: mutation, selection, population subdivision, breeding topic and illegitimate recombination) is well documented for
system, and recombination in its various forms. All other different organisms.
things being equal, genetic variance for a sexual population Recombination was recognized to play an important role
with free recombination may be several times that for a in the evolution of gene families that are usually organized
population with no recombination (Charlesworth, 1993). in the genome as arrays of tandem repeats residing in one or
Classical estimates from Drosophila species indicated that up to a few chromosomes. Coordinated change of individual
 Recombination 363

elements of such families is referred to as concerted evolution. heterozygous for AT and GC (the biased gene conversion hy-
The diversification of the elements by mutations is opposed to pothesis). Also, we cannot exclude the possibility that another
homogenization processes caused by unequal crossing-over factor affecting both the evolution of intragenomic hetero-
and gene conversion. This process is opposite to the scenario geneity in GC content and recombination is responsible for
of evolution of new functions through ‘‘recombination (un- their positive association. An important aspect of the re-
equal crossing-over)-duplication-diversification (via mu- combination impact on genome organization is related to the
tation)’’. In the case of multigene families, selection amount and distribution of repetitive elements. Amplification
presumably favors the most homogeneous arrays. Different of retrotransposons is considered as an important source of
variants of homogenized arrays may coexist within a popu- repetitive DNA leading to an increase in genome size
lation, as shown on the rDNA cluster in Drosophila. This im- (‘‘genomic obesity’’) in many plant taxa. Recombination was
plies a low recombination between different arrays combined proposed as the mechanism counterbalancing this trend, in
with high intrachromosomal recombination within arrays. particular, unequal crossing-over and intrachromosomal re-
combination between the repeated elements (Shirasu et al.,
2000). Recombination is an important mechanism affecting
the formation of microsatellites. Owing to their abundance in
Recombination and Sequence Polymorphism
the genomes and high variability, microsatellites are con-
During the past decades, new evidence became available on sidered as accelerators of evolutionary adaptation (Kashi and
the association between recombination and DNA sequence King, 2006).
polymorphism in natural populations. The first group of data
came from Drosophila: Begun and Aquadro (1992) compared
variation in rf per physical DNA length in different regions of Horizontal Gene Transfer
D. melanogaster genome with polymorphism of DNA se- Numerous mechanisms exist that normally prevent the ex-
quences of these regions. Positive correlation between re- change of genetic material among distant species. Neverthe-
combination and sequence polymorphism was established. less, if this process is possible even at a low frequency, it may
Similar results were obtained in different Drosophila species, have important consequences on the long-term evolutionary
humans, and some plants. Recombination also seems to play scale. Only 30 years ago, such a possibility was considered an
an important role in sequence variation of natural popu- unrealistic one. Now, rich evidence is available showing that
lations for the mitochondrial genome. the opposite is true. Illegitimate recombination seems to be
Two opposite though not mutually exclusive hypotheses the main responsible mechanism. Extensive similarities have
were proposed as explanations for the association between been revealed between house-keeping genes of Archaea and
recombination and sequence polymorphism. The first, re- Eubacteria, although for the majority of informational genes
ferred to as selective sweeping, is based on selection of new (involved in transcription and translation) the difference is
favorable mutations. In genomic regions with a low re- very large. An interesting fact is that some of the foregoing
combination rate, the process of fixation of positively selected genes showing high similarity between the kingdoms are
mutations will result in fixation of closely linked neutral or clustered in the genome that may result from horizontal
even slightly deleterious sequence variations. The second ex- transfer events.
planation is based on the theory of background selection If horizontal transfer is indeed an evolutionary significant
(Nordborg et al., 1996) that considers the consequences of form of genetic recombination, one could consider ecological
purifying selection against deleterious mutations. This process interaction between species a replacement of sexual inter-
can also reduce sequence variation at closely linked sequences. action (vertical transfer). In other words, ecologically inter-
Combination of these two mechanisms seems to explain the acting species are the most probable exchanging partners. Viral
main pattern of polymorphism–recombination association, DNA integration in the host nuclear genome is considered a
although the contribution of additional factors cannot be relatively frequent event (on the evolutionary scale) for animal
excluded. hosts and recently also for plants. Likewise, there is evidence
A simple and universal measure of compositional organ- of an opposite, host-parasite flow of genetic information:
ization of nucleotide sequences is GC content. GC content Many transfers of signaling domains of pathogenesis-related
displays wide interspecific variation as well as high within- plant genes were detected in bacterial genomes.
genome heterogeneity, especially in mammals, where it is
presented in the form of GC isochors. GC content is strongly
correlated with biological features of genome organization
Evolution of Recombination
such as distribution of various classes of repeated elements,
gene density, level and tissue-specificity of transcription, and
Theoretical Models of Recombination Evolution
mutation rate. The available data on humans and other or-
ganisms point to a strong correlation between GC content and Artificial selection experiments suggest that almost every
recombination. This connection may be caused by increased population has enough stored genetic variability to ensure
recombination in GC-rich regions. An alternative assumption response to direct selection for changed recombination fre-
is that recombination machinery is responsible for the evo- quency (see Selection for Changed Recombination, Genetic
lution of the genomic GC distribution due to the bias of Modifiers of Recombination, and Intraspecific Variation for the
mismatch repair within the gene conversion tracts in favor of Rate of Recombination). The observed polymorphism at loci
GC, which preferentially converts A/T into G/C at sites that are affecting recombination could be either balanced (selected) or
364 Recombination

transient. Theoretical analysis shows that under a stable en- these models, the conditions favoring increased recombin-
vironment, a panmictic population should evolve toward a ation are associated with negative linkage disequilibria among
minimum possible level of recombination. This can be for- selected loci owing to stringent conditions for epistasis. Some
mulated in terms of the fate of a selectively neutral modifier mechanisms (e.g., selection of beneficial mutations) may favor
locus affecting recombination. recombination in the absence of epistasis, with linkage dis-
Understanding the forces maintaining sex and recombin- equilibrium caused by finite population size. The diversity of
ation is considered one of the most challenging problems in models proposed to explain the evolution of sex and re-
evolutionary theory (Michod and Levin, 1988). Shared genetic combination and their rather complicated ‘‘relationships’’
control and molecular mechanisms of DNA recombination with the evidence from studies of different organisms, lead to
and DNA repair across life (see Meiotic Mutants as an Ana- the idea of pluralistic explanation (Michod and Levin, 1988;
lytical Tool in Recombination Studies: Overlapping of DNA Korol et al., 1994; West et al., 1999). It suggests that different
Repair, Recombination, and Segregation Systems and Mo- mechanisms could have played different roles in the emer-
lecular Mechanisms of Recombination) indicate their com- gence, evolution, and, especially, maintenance of recombin-
mon evolutionary origin and functional overlap in extant ation and sex in diverse phyletic lines of eukaryotes.
organisms. It could probably be supposed that repair func-
tions played the leading role at early evolutionary stages,
having provided opportunities for a large increase in the Testing the Theoretical Assumptions
genome size and the transition from haploidy to diploidy. The
Despite voluminous literature, the problem has been studied
latter offered the possibility of recombination repair of two-
little experimentally or by observations in natural popu-
strand DNA lesions that is impossible in haploid systems.
lations. In the classical period of genetics, it was proposed that
Some authors hypothesize that the subsequent stages in the
sex and recombination may facilitate selection for advan-
evolution of recombination and sex were associated with re-
tageous and against deleterious mutations and assist in
pair alone. These explanations of repair functions were called
adaptation to changing environment. The major question is
‘‘physiological’’ (Maynard Smith, 1978). Another physio-
whether these processes in turn are able to promote evolution
logical explanation of recombination in sexuals is its associ-
of the recombination system. Are the resulting changed pat-
ation with chromosome segregation (see Meiotic Mutants as
terns of recombination detectable experimentally? Is it pos-
an Analytical Tool in Recombination Studies: Overlapping of
sible to find corresponding changes in nature? Theoretical
DNA Repair, Recombination, and Segregation Systems), al-
models answer positively to these questions. However, do
though numerous examples are known, such as male Dros-
these models (and/or their parameter values) fit the real
ophila and female silkworm, in which normal segregation is
world? Unfortunately, there is not enough evidence to answer
associated with achiasmatic (without crossing-over) meiosis.
this question. Two different and complementary approaches
Combinative, or generative, hypotheses consider the main
can be considered (Korol et al., 1994; Barton and Charles-
function of sex and recombination in shuffling genes. This
worth, 1998; Korol, 1999).
removes negative correlation between favorable alleles at dif-
ferent loci, thereby increasing the efficiency of natural selec-
tion. Generative models can be classified according to the Testing the Initial Assumptions
source of linkage disequilibria between selected loci: stochas- The aim is to test whether the real estimates of the main
tic or deterministic, caused by new mutations or variation of parameters underlying the proposed mechanism(s) fit the
external conditions (Kondrashov, 1993). In the 1930s, Fisher expectations. These include mutation rate, effective population
and Muller proposed that sex may be advantageous by com- size, mode of interaction of deleterious or beneficial mu-
bining beneficial mutations randomly occurring in different tations, relative role of stabilizing and directional selection,
individuals (Hartfield et al., 2010). A complementary version and cost of sex and meiosis. The major advantage of this ap-
of the stochastic–mutation explanation considers the role of proach is that it allows the accumulation of data on basic
recombination in selection against deleterious mutations parameters affecting the evolution of recombination. The
(Muller, 1964). According to Muller, deleterious mutations drawback is that it is difficult to believe that even the main
tend to be fixed in a finite asexual population due to random factors were taken into account so that the predicted direction
drift despite purifying selection (Muller’s ratchet), whereas of recombination evolution is determined correctly. As an
recombination helps to stop this process. In the 1980s, a few example of a successful application of this approach, com-
deterministic models of selection against deleterious alleles petition experiments with yeast S. cerevisiae (Greig et al., 1998)
were proposed, with the evolutionary advantage of re- are worth mentioning. Using sexual and asexual strains it was
combination being dependent on linkage disequilibria re- found, contrary to the expectations following from the con-
sulted from synergistic interaction between harmful mutations cept of costs of sex or meiosis, that sex can confer significant
produced at a high rate (Kondrashov, 1988). Other models of selective advantage to its carriers. Direct tests have also been
deterministic sources of linkage disequilibria include (1) conducted with the nematode C. elegans demonstrating
adaptation to temporarily or spatially varying environment, the possible role of recombination in purifying selection
(2) antagonistic species interaction (the Red Queen hypoth- (Zetka et al., 1987). It was shown that increased rf owing to
esis), and (3) instraspecific competition (sib competition or enhancer Rec-1 results in a higher fitness of genotypes carrying
tangled bank hypothesis) (Hamilton et al., 1990; Kondrashov, mutator gene mut-6 compared to analogous genotypes with a
1993; Korol et al., 1994; Feldman et al., 1997; Barton and normal recombination rate. Therefore, the presence of a gene
Charlesworth, 1998; Otto and Michalakis, 1998). In most of for increased recombination, which presumably has no effect
 Recombination 365

on mutation, reduces the genetic load caused by the mutator Especially intriguing is the high stability of number of ex-
system. However, studies of normal (wild-type) genotypes of changes per chromosome among higher eukaryotes (4–6 six-
C. elegans give very low estimates of mutation rate, which is fold) despite huge variation of Mb/cM ratio (more than two
inconsistent with the hypothesis of selection against dele- orders of magnitude – see Table 2).
terious mutations.

Experimental Evolution of Recombination

Searching for Changed Recombination
The second approach is to test the ‘‘final effect’’ by comparing Changes in Recombination Caused by Selection for
rates and genomic distribution of recombination in experi- Adaptive Traits
mental and natural populations subjected to contrasted se- Theoretical models indicate that directional or variable selec-
lection regimes during many generations. The advantage of tion for multilocus traits may promote evolution toward in-
such an analysis is that the target effect (changed pattern of creased recombination if the selected population is
recombination) is estimated directly (see Experimental Evo- polymorphic for recombination-modifying loci (Charlesworth,
lution of Recombination), although one cannot be sure that it 1993; Korol et al., 1994; Otto and Barton, 1997). Only a few
is causally connected only with the presumed process. experiments seem to have a direct bearing on this question.
The following are examples with regard to Drosophila (Korol,
1999). Clegg et al. (1979) attempted to test Fisher’s (1930)
Recombination System and Life History Traits prediction that stable favorable conditions may promote
evolution toward reduced recombination. The experiment
Rasmusson (1927) was possibly the first to discuss the genetic
started with a large cage population of D. melanogaster with
basis of putative differences in recombination systems of
maximum LD between allozyme loci of chromosome 3. No
outbreeding and inbreeding species. On the basis of extremely
general trend in rf for 20 generations was revealed, which was
high variation in the recombination rate in peas that is usually
explained as an indication of a low epistatic component of
not found in Drosophila, he advanced a hypothesis of possible
selection within the marked interval. Flexon and Rodell
differences in the system of genetic control of recombination
(1982) established a synthetic population of D. melanogaster
between outbreeders and selfers. In plants, elevated chiasma
and subjected it to selection for DDT resistance for 300 days.
frequencies were indeed found in selfers compared to related
By the time of estimation, resistance increased from 10- to
cross-pollinating species (Ross-Ibarra, 2006). The species ori-
100-fold relative to the control level. The control and treat-
gin and life conditions determine the strategy for adaptation
ment populations were compared with respect to recombin-
and hence the ‘‘optimum’’ amount of recombination. Evi-
ation rate across the genome. Increased rf between marker loci
dence of variation of chiasma frequency in nature and its
on chromosomes 2 and 3 but not X was found in the selected
correlation with ecological and life historical traits is available
population. Most experimental studies on recombination
for different organisms and may be employed to compare
evolution have employed one-way selection for fitness traits.
alternative theoretical models. In mammals with low repro-
The limitation of this scheme is the danger of a change in rf
duction rates, longer life cycles, and small progeny sizes, re-
caused by initial linkage disequilibria between selected genes
combination is higher than in species with short generation
and recombination modifiers. Therefore, to assess the effect of
times and large progeny sizes (Burt and Bell, 1987), fitting the
selection for geotaxis on recombination in D. melanogaster,
Red Queen model. Koella (1993) found that in plants with
Korol and Iliadi (1994) employed two-way selection. Forty
animal-dispersed seeds, recombination is higher than in other
generations of selection for altered geotaxis resulted in an
plants, and perennials show a higher recombination than
increment in recombination for the genome of 78 cM for
annuals when compared among genera but an opposite ten-
geo þ and 66 cM for geo  lines compared to nonselected
dency is displayed on the species level. Similarly, on the
control. Selection for negative geotaxis did not affect rf in
whole-genome level, higher cM/Mb ratios were found in trees
chromosome 2, whereas selection in the opposite direction
compared to short-lived herbs and shrubs (Jaramillo-Correa
caused a 4-fold increase in rf in the b-cn segment spanning the
et al., 2010).
centromere of chromosome 2. The observed change in rf,
The rapidly increasing number of genetically mapped or-
caused by two-way selection for geotaxis, was attributed to the
ganisms simultaneously demonstrates two opposite patterns.
advantage conferred by selection on recombinants. Similar
The first is a rather high similarity of some genome-wide
results were obtained by Rodell et al. (2004) in divergent se-
characteristics of the recombination system among different
lection for the sternopleural bristle number in D. melanogaster.
species, including the narrow spectrum of the crossover
Significantly, increased rates of recombination in chromo-
number per chromosome, positive interference, the cen-
somes 2 and 3 were found in both directions of selection, and
tromeric effect, and a positive correlation with GC content.
a tendency for reducing crossover rates was observed in
The second is the discovery of high variation of recombination
populations subjected to stabilizing selection.
features within and between species, with extreme outliers,
like honey bees (and other social insects), with their extremely
low Mb/cM ratio compared to other insects and some species, Experimental Evolution of Recombination Caused by
for example, Monodelphis or Pinus and Hordeum, with con- Adaptation to Stress
siderably higher cM/Mb compared to correspondingly other The author reviewed a series of studies with D. melanogaster
vertebrates and plants. In most of the cases, we still do not conducted in his laboratory and aimed to demonstrate
know the reasons for both, high similarity and high variation. that population adaptation to adverse conditions may
366 Recombination

promote evolution for increased recombination (Korol et al., selection against harmful mutations alone seems to be in-
1994; Korol, 1999). The experiments were performed on four sufficient for maintenance of the normal level of recombin-
large heterogeneous populations with the objective of simu- ation, at least in the genetic material that was tested.
lating the process of population adaptation to stress and
evaluating its effect on recombination. At the start, each
population was subdivided randomly into two parts, control Applied Aspects
(C) and treatment (T); T variants were then subjected to daily
fluctuating temperature with amplitude increasing with gen- Recombination as a Tool of Genetic Mapping
erations (from 23 to 27 1C at the beginning to 12–32 1C
at the end of the experiment). C variants were maintained at Recombination-based mapping as a method of studying the
25 1C. genetic material topography has been a major analytical ap-
proach in genetics for nearly a century. However, until recently
this technique was of little applied interest because of a lack of
Increased Recombination Resulting from Adaptation to Stressful easily scorable markers. The situation changed dramatically
Conditions with the appearance of DNA markers. Establishing high-
The control and treatment populations appeared to diverge quality maps was considered one of the most important ob-
genetically with respect to thermotolerance. Changes in re- jectives in the Human Genome Project, which also included a
combination were evaluated using multiple marked lines. All few model organisms. Mapping efforts allow a better under-
four populations showed one pattern – a segment-specific standing of the organization of genetic material and allow one
increase in rf, resulting from selection for thermoadaptation. to conduct far-reaching comparative mapping. This approach
Those that reacted most were the near-centromeric regions of is a power tool of ever-increasing importance for studies of
chromosomes 2 and 3 and the left arm of chromosome 3 genome structure, functioning, and evolution, and for diverse
(Figure 3). The T and C variants were also compared with medical and commercial applications of modern genomics.
respect to the frequency of double exchanges (interference). Fine genetic mapping is a basis of positional cloning, an ap-
Significant differences in the values of coefficients of coinci- proach proved successful in many organisms. Its efficiency
dence have been found for adjacent and nonadjacent intervals strongly depends on the Mb/cM ratio in the region harboring
of large autosomes. The general tendency was a reduced the target gene. If the target region is very rarely involved in
interference in T variants, corroborating the results for recombination, then map-based cloning is impractical. In-
recombination rate. duced recombination may be a nontraditional approach for
such situations. The availability of dense maps facilitates
Reduction of Recombination Rate under Optimal Stable Conditions mapping of genetic disease genes in humans with subsequent
The described divergence over time with respect to re- cloning, sequencing, and medical applications. An important
combination rate resulted from two independent trends. An application of mapping is genetic dissection of complex traits,
increase in rf was observed in populations adapting to adverse or quantitative trait loci mapping, especially in the framework
conditions, and simultaneously tightening of linkage in con- of new marker-assisted (or ‘‘genomic selection’’) strategies in
trol populations that were maintained under optimal tem- plant and animal breeding.
perature. The proportion of these two effects varied in
different ‘‘treatment–control’’ pairs corresponding to the
Recombination as a Major Source of Genetic Variation in
structure of the starting population. These results can be
Breeding
considered as the first experimental verification of Fisher’s
(1930) hypothesis that in a constant environment selection When viewed at the level of phenotypic traits, genetic innov-
favors tighter linkage. Indirectly, these results also indicate that ations produced by recombination on artificial hybridization

Chromosome 2 Chromosome 3
40
Recombination rate (%)

35
30
25
20
15
10
5
0
al-b b-pr pr-c c-px px-sp ru-h h-th th-st st-cu cu-sr sr-e e-ca
Segment
Figure 3 Adaptation to daily fluctuating temperature results in a segment-specific increase in recombination rate in large autosomes of
D. melanogaster (based on data from Korol AB, Preygel IA, and Preygel SI (1994) Recombination Variability and Evolution. London: Chapman
& Hall).
 Recombination 367

can arbitrarily be divided into three main types: (1) trans- also involved in meiotic recombination, suggesting that
gression for individual traits with the range of trait values in tumor-suppressive functions may have been co-opted
the segregating progeny exceeding the respective parental from primordial activities linked to recombination (Lu et al.,
ranges; (2) formation of new trait combinations of the crossed 2010).
components; and (3) the appearance of new traits (‘‘anom-
alous’’ variation) as a result of recombination in genetic
complexes with strong nonallelic interactions. These forms of Recombination Shuffling of Genes: New Technologies of
recombination variability play a significant role in breeding Artificial Sequence Optimization
programs providing the raw material for selection. The im-
Recombination combined with selection has become the basis
portance of controlling recombination increases with the ne-
for new technologies of ‘‘molecular breeding’’ aimed at im-
cessity of involving new genetic resources for breeding
proving protein functions via directed evolution in vitro. It
purposes, especially in light of ever-decreasing homeostasis
includes reshuffling of a few copies of the coding DNA by
and tolerance to abiotic and biotic stresses of elite animal
homologous or site-specific recombination with subsequent
breeds and plant cultivars. New genes are supposed to be
screening of the resulting sequences for functional (enzym-
introgressed from exotic germplasm based on recombination.
atic) activity. The reshuffling can be provided by template
This problem is complicated by reduced recombination on
switching in selfpriming polymerase chain reaction (PCR)
distant crosses and linkage drag (close linkage between the
referred to as sexual PCR. The efficiency of such artificial
target gene and undesirable genes). Hence, manipulating re-
evolution may be enhanced by increasing the initial genetic
combination may become an efficient tool of modern
variation involved in recombination, for example, by muta-
breeding.
genesis of the target sequences based on error-prone PCR.
Another possibility is to include a wider spectrum of parental
Recombination and Cancer sequences using the natural diversity of homologous genes of
the same or related species or even distant taxa (Crameri et al.,
There are many aspects of anomalous chromosome and cell 1998). This approach is especially promising for the im-
behavior and cancer related to DNA metabolism and provement of molecules that provide resistance against toxic
recombination–repair. A few are noted here. agents or increase survival on minimal media. Experiments
including transformation of the engineered sequences into
Mitotic Recombination single-cell organisms (E. coli) and selection of the transfor-
A surprisingly large proportion of human neoplasms appear mants showed dozens-to-hundreds- and even thousands-fold
to result from loss of function of tumor-suppressing genes, for increases in efficiency of the target molecules obtained after
example, due to loss of heterozygosity of tumor suppressors. recombination–selection cycles. DNA shuffling can improve
In retinoblastomas caused by loss of heterozygosity, about half the targeted pathway by complex mechanisms, even though it
of the homozygosity cases appear to be the result of mitotic may be impossible to predict the optimized sequence by a
recombination. rational design. Of special interest is the use of recombination
for engineering novel therapeutic drugs. Namely, homology-
Specificity of Chromosome Translocations independent recombination of human genes with genes from
Many types of human cancers are associated with non- other species may be used for construction of ‘‘humanized’’
homologous recombination events producing specific trans- proteins (with a part of nonhuman sequence) possessing the
locations. This results in an enhanced expression of oncogenes desired clinical activity (Griswold et al., 2005). The success of
residing in the vicinity of the breakpoint. Some of these using recombination as a strategy for protein engineering
oncogenes are transcription factors participating in normal corroborate the findings that recombination causes more tol-
development. It seems reasonable to assume that interaction erable changes in protein coding sequences compared to
between recombination and transcription (see Recombination random mutation, implying a higher importance of re-
Signals, Hotspots of Recombination, and Transcription) may combination as a source of meaningful evolutionary genetic
also involve ectopic recombination resulting in the specificity variation (Drummond et al., 2005).
of some tumorigenic translocations (Barr, 1998).

Targeted Gene Replacement

Defects in DNA Repair–Recombination
A few cancer-related genetic disorders of this type have been Gene targeting is an experimental approach aimed at pro-
characterized, including Bloom syndrome, which manifests ducing site-specific changes in the genome based on genetic
mutator and hyper-recombination phenotypes; Ataxia recombination. Predetermined changes can be introduced in
telangiectasia mutation, which increases V(D)J-mediated the target site, in both somatic and germline cells. Usually, the
chromosome rearrangement in T lymphocytes; and Omenn targeting system involves a gene mediating site-specific re-
syndrome, a disorder caused by mutations of recombination– combination and a specific target locus. The best-known ex-
activation genes Rag1 and Rag2 (related to V(D)J recombin- amples are Cre and Flp recombinases (of bacteriophage P1
ation) resulting in severe immune deficiency combined with a and yeast) with IoxP and FRT targets, respectively. The basic
complete lack of circulating T and B cells. Recent studies in idea is to first transform the target sites flanking a
Drosophila and mice show that one of the most important selectable marker into the recipient genome. Subsequent
tumor-suppressing genes controlling cell apoptosis, p53, is introduction of the recombinase activity leads to the excision
368 Recombination

of the marker. Many important applications have been pro- Burt A and Bell B (1987) Mammalian chiasma frequencies as a test of two theories
posed and many others are expected using the powerful of recombination. Nature 326: 803–805.
Charlesworth B (1993) Directional selection and evolution of sex and
strategies of targeted gene replacement. These include con-
recombination. Genetical Research 61: 205–224.
trolled expression of transformed genes, disruption of a cho- Clegg MT, Horch CR, and Kidwell JF (1979) Dynamics of correlated genetic
sen gene in vivo for a better understanding of its role in systems. VI. Variation in recombination rates in experimental population of
physiological and pathological processes, disruption of genes Drosophila melanogaster. Journal of Heredity 70: 297–300.
to rechannel the biosynthesis and achieve overproduction of Conway DJ, Roper C, Oduola AM, et al. (1999) High recombination rate in natural
populations of Plasmodium falciparum. Proceedings of the National Academy of
certain molecules, site-specific targeting of DNA into embry- Sciences USA 96: 4506–4511.
onic stem cells for production of gene-modified model or- Coop GC, Wen X, Ober C, Pritchard JK, and Przeworski M (2008) High-resolution
ganisms, targeted delivery of specific constructs aimed at mapping of crossovers reveals extensive variation in fine-scale recombination
producing antitumor effects or complementing a mutated patterns among humans. Science 319: 1395–1398.
Crameri A, Raillard SA, Bermudez E, and Stemmer WP (1998) DNA shuffling of a
gene, and testing of gene expression in donor versus recipient
family of genes from diverse species accelerates directed evolution. Nature 391:
tissues in biomedical transplantation studies. 288–291.
Cromie GA and Smith GR (2007) Branching out: Meiotic recombination and its
regulation. Trends in Cell Biology 17: 448–455.
Conclusions Danzmann RG, Davidson EA, Ferguson MM, et al. (2008) Distribution of
ancestral proto-Actinopterygian chromosome arms within the genomes of 4R-
derivative salmonid fishes (Rainbow trout and Atlantic salmon). BMC Genomics
Recombination plays a key role in a remarkably broad spec- 9: 557.
trum of genetic processes related to DNA repair; sexual re- Drummond DA, Silberg JJ, Meyer MM, Wilke CO, and Arnold FH (2005) On the
production; immunity; genetic diversity; evolution of genes, conservative nature of intragenic recombination. Proceedings of the National
gene families, chromosomes and the entire genome; and Academy of Sciences USA 102: 5380–5385.
Feldman MW, OttoS P, and Christiansen FB (1997) Population genetic perspectives
evolutionary adaptation and speciation. Recombination is the on the evolution of recombination. Annual Review of Genetics 30: 261–295.
basis for traditional breeding, genetic mapping, recombinant Fisher RA (1930) The Genetical Theory of Natural Selection. Oxford: Clarendon.
DNA technologies and genetic engineering, gene targeting and Flexon PB and Rodell CF (1982) Genetic recombination and directional selection
gene therapy, developmental biology, and cancer genetics. for DDT resistance in Drosophila melanogaster. Nature 298: 672–675.
Greig D, Borts RH, and Louis EJ (1998) The effect of sex on adaptation to high
Genetic studies of various organisms have accumulated
temperature in heterozygous and homozygous yeast. Proceedings of the Royal
abundant evidence on common features and peculiarities in Society London B 265: 1017–1023.
genetic control and molecular mechanisms of the main types Griswold KE, Kawarasaki Y, Ghoneim N, Benkovic SJ, Iverson BL, and Georgiou G
of genetic recombination, including reciprocal exchange (2005) Evolution of highly active enzymes by homology-independent
between homologous chromosomes (crossing-over), non- recombination. Proceedings of the National Academy of Sciences USA 102:
10082–10087.
reciprocal exchange (gene conversion), site-specific re- Hadany L and Beker T (2003) On the evolutionary advantage of fitness-associated
combination occurring between homologous sequences, and recombination. Genetics 165: 2167–2179.
illegitimate recombination between nonhomologous Hamilton WD, Axelrod RA, and Tanese R (1990) Sexual reproduction as an
sequences. However, despite a long history, thousands of adaptation to resist parasites (a review). Proceedings of the National Academy of
Sciences USA 87: 3566–3573.
studies, and great progress during the last two decades, re-
Hartfield M, Otto SP, and Keightley PD (2010) The role of advantageous mutations
combination continues to be a puzzle in the aspects of its in enhancing the evolution of a recombination modifier. Genetics 184:
mechanisms, evolutionary effects, contribution of biodiversity, 1153–1164.
and the factors determining its own evolution. Heyer WD, Ehmsen KT, and Liu J (2010) Regulation of homologous recombination
in eukaryotes. Annual Review of Genetics 44: 113–139.
Hyppa RW and Smith GR (2010) Crossover invariance determined by partner choice
for meiotic DNA break repair. Cell 142: 243–255.
See also: Adaptation. Coevolution. Diversity, Molecular Level. Jaramillo-Correa JP, Verdú M, and González-Martı́nez SC (2010) The contribution
Genetic Diversity. Life History, Evolution of of recombination to heterozygosity differs among plant evolutionary lineages and
life-forms. BMC Evolutionary Biology 10: 22.
Kashi Y and King DG (2006) Simple sequence repeats as advantageous mutators in
evolution. Trends in Genetics 22: 253–259.
Koella JC (1993) Ecological correlates of chiasma frequency and recombination
References index of plants. Biological Journal of the Linnean Society 48: 227–238.
Kondrashov AS (1988) Deleterious mutations and the evolution of sexual
Agrawal AF and Wang AD (2008) Increased transmission of mutations by low- reproduction. Nature 336: 435–440.
condition females: Evidence for condition-dependent DNA repair. PLoS Biology Kondrashov AS (1993) Classification of hypotheses on the advantage of
6: e30. amphimixis. Journal of Heredity 84: 372–387.
Andersen SL and Sekelsky J (2010) Meiotic versus mitotic recombination: Two Koren A, Ben-Aroya S, and Kupiec M (2002) Control of meiotic recombination
different routes for double-strand break repair. Bioessays 32: 1058–1066. initiation: A role for the environment? Current Genetics 42: 129–139.
Barr FG (1998) Translocations, cancer and the puzzle of specificity. Nature Genetics Korol AB (1999) Selection for adaptive traits as a factor of recombination evolution:
19: 121–124. Evidence from natural and experimental populations. In: Wasser SP (ed.)
Barton NH and Charlesworth B (1998) Why sex and recombination? Science 281: Evolutionary Theory and Processes: Modern Perspective, pp. 31–53. Dordrecht:
1986–1990. Kluwer.
Baudat F, Buard J, Grey C, et al. (2010) PRDM9 is a major determinant of meiotic Korol AB and Iliadi KG (1994) Recombination increase resulting from directional
recombination hotspots in humans and mice. Science 327: 836–840. selection for geotaxis in Drosophila. Heredity 72: 64–68.
Begun DJ and Aquadro CF (1992) Levels of naturally occurring DNA polymorphism Korol AB, Kirzhner VM, Ronin YI, and Nevo E (1996) Cyclical environmental
correlate with recombination rates in D. melanogaster. Nature 356: 519–520. changes as a factor maintaining genetic polymorphism. II. Two-locus diploid
Bhagat R, Manheim EA, Sherizen DE, and McKim KS (2004) Studies on crossover- selection. Evolution 50: 1432–1441.
specific mutants and the distribution of crossing over in Drosophila females. Korol AB, Preygel IA, and Preygel SI (1994) Recombination Variability and
Cytogenetic and Genome Research 107: 160–171. Evolution. London: Chapman & Hall.
 Recombination 369

Landner L (1970) Variation of recombination frequency in Neurospora crassa Rodell CF, Schipper MR, and Keenan DK (2004) Modes of selection and
following meiosis and evidence for a premeiotic sensitive stage. Molecular and recombination response in Drosophila melanogaster. Journal of Heredity 95:
General Genetics 27: 385–392. 70–75.
Laurencon A, Gay F, Ducau J, and Bregliano J-C (1997) Evidence for an inducible Ross-Ibarra J (2004) The evolution of recombination under domestication: A test of
repair-recombination system in the female germ line of Drosophila melanogaster. two hypotheses. American Naturalist 163: 105–112.
III. Correlation between reactivity levels, crossover frequency and repair Ross-Ibarra J (2006) Genome size and recombination in angiosperms: a second
efficiency. Genetics 146: 1333–1344. look. Journal of Evolutionary Biology 20: 800–803.
Lenormand T and Dutheil J (2005) Recombination difference between sexes: A role Sandler L and Szauter P (1978) The effect of recombination-defective meiotic
for haploid selection. PLoS Biology 3: e63. mutants on fourth-chromosome crossing-over in Drosophila melanogaster.
Lichten M and Goldman ASH (1995) Meiotic recombination hotspots. Annual Genetics 90: 699–712.
Review of Genetics 29: 423–444. Schoustra S, Rundle HD, Dali R, and Kassen R (2010) Fitness-associated sexual
Loidl J and Scherthan H (2004) Organization and pairing of meiotic chromosomes reproduction in a filamentous fungus. Current Biology 20: 1350–1355.
in the ciliate Tetrahymena thermophila. Journal of Cell Science 117: 5791–5801. Shalev G, Sitrit Y, Avivi-Ragolski N, Lichlenstein C, and Levy AA (1999) Stimulation
Lu W-J, Chapo J, Roig I, and Abrams JM (2010) Meiotic recombination of homologous recombination in plants by expression of the bacterial resolvase
provokes functional activation of the p53 regulatory network. Science 328: ruvC. Proceedings of the National Academy of Sciences USA 96: 7398–7402.
1278–1281. Shirasu K, Schulman AH, Lahaye T, and Schulze-Lefert P (2000) A contiguous 66-
Lukaszewski AJ and Curtis CA (1993) Physical distribution of recombination in kb barley DNA sequence provides evidence for reversible genome expansion.
B-genome chromosomes of tetraploid wheat. Theoretical and Applied Genetics Genome Research 10: 908–915.
86: 121–127. Stevison LS and Noor MAF (2010) Genetic and evolutionary correlates of fine-scale
Lupu A, Nevo E, Zamorzaeva I, and Korol AB (2006) Ecological-genetic feedback in recombination rate variation in Drosophila persimilis. Journal of Molecular
DNA repair in wild barley, Hordeum spontaneum. Genetica 127: 121–132. Evolution 71: 332–345.
Maynard Smith J (1978) The Evolution of Sex. Cambridge, UK: Cambridge Sved JA, Blackman LM, Gilchrist AS, and Engels WR (1991) High level of
University Press. recombination induced by homologous P elements in Drosophila melanogaster.
Maynard Smith J, Dowson CG, and Spratt BG (1991) Localized sex in bacteria. Molecular and General Genetics 225: 443–447.
Nature 349: 29–31. Szostak JW, Orr-Weaver TL, Rothstein RJ, and Stahl FW (1983) The double-strand-
McKim KS and Hayashi-Hagihara A (1998) mei-W68 in Drosophila melanogaster break repair model for recombination. Cell 33: 25–35.
encodes a Spoll homolog: Evidence that the mechanism for initiating meiotic Takeda J, Uematsu N, Shiraishic S, Toyoshima M, Matsumotoaand T, and Niwae O
recombination is conserved. Genes and Development 12: 2932–2942. (2008) Radiation induction of delayed recombination in Schizosaccharomyces
McVicker G and Green P (2010) Genomic signatures of germline gene expression. pombe. DNA Repair 7: 1250–1261.
Genome Research 20: 1503–1511. Thuriaux P (1977) Is recombination confined to structural genes of the eukaryotic
Michod RE and Levin BR (eds.) (1988) The Evolution of Sex: An Examination of chromosome? Nature 268: 460–462.
Current Ideas. Sunderland, MA: Sinauer. Trivers R (1988) Sex differences in rates of recombination and sexual selection. In:
Molinier J, Ries G, Zipfel C, and Hohn B (2006) Transgeneration memory of stress Michod RE and Levin BR (eds.) The Evolution of Sex: An Examination of
in plants. Nature 442: 1046–1049. Current Ideas, pp. 270–286. Sunderland, MA: Sinauer.
Morrell PL, Toleno DM, Lundy KE, and Clegg MT (2006) Estimating the Tsaousis AD, Martin DP, Ladoukakis ED, Posada D, and Zouros E (2005)
contribution of mutation, recombination and gene conversion in the generation Widespread recombination in published animal mtDNA sequences. Molecular
of haplotypic diversity. Genetics 173: 1705–1723. Biology and Evolution 22: 925–933.
Muller HJ (1964) The relation of recombination to mutational advance. Mutation West SA, Lively CM, and Read AF (1999) A pluralist approach to sex and
Research 1: 2–9. recombination. Journal of Evolutionary Biology 12: 1003–1012.
Neel JV (1941) A relation between larval nutrition and the frequency of crossing Whitehouse HLK (1966) An operator model of crossing over. Nature 211: 708–713.
over in the third chromosome of Drosophila melanogaster. Genetics 26: Williams CG, Goodman MM, and Stuber CW (1995) Comparative recombination
506–516. distances among Zea mays L. inbreds, wide crosses and interspecific hybrids.
Nevo E, Kirzhner V, Beiles A, and Korol A (1997) Selection versus random drift: Genetics 141: 1573–1581.
Long-term polymorphism persistence in small populations (evidence and Yamamoto T (1961) Progenies of sex-reversal females mated with sex-reversal
modelling). Philosophical Transactions of the Royal Society B 352: 381–389. males in the Medaka, Oryzias Iatipes. Journal of Experimental Zoology 146:
Nordborg M, Charlesworth B, and Charlesworth D (1996) The effect of 163–179.
recombination on background selection. Genetics Research 67: 159–174. Zetka MC, Rose AM, and Baillie DL (1987) A test of the Muller Ratchet Hypothesis
Otto SP and Barton NH (1997) The evolution of recombination: Removing the limits using the REC-1 strain of Caenorhabditis elegans. Genetics 116(supplement):
to natural selection. Genetics 147: 879–906. 30.
Otto SP and Michalakis Y (1998) The evolution of recombination in changing Zhuchenko AA and Korol AB (1983) Ecological aspects of the recombination
environments. Trends in Ecology and Evolution 13: 145–151. problem. Theoretical and Applied Genetics 64: 177–185.
Pecinka A, Rosa M, Schikora A, et al. (2009) Transgenerational stress memory is Zhuchenko AA, Kovtyukh LP, and Korol AB (1983) Testing the ‘‘ontogenetic
not a general response in Arabidopsis. PLoS One 4: e5202. memory’’ hypothesis on Drosophila. Doklady Academii Nauk USSR 269:
Rasmusson J (1927) Genetically changed linkage values in Pisum. Hereditas 10: 1493–1495 (in Russian).
1–52.