Review: Genome editing as a tool to achieve the crop ideotype and de novo domestication of wild relatives: case study in tomato

Full text

UNCORRECTED PROOF
ARTICLE INFO
Article history:
Received 8 October 2016
Received in revised form 21 December
2016
Accepted 23 December 2016
Available online xxx
Keywords:
Breeding
CRISPR/Cas9
Genome editing
Stress resistance
TALENs
Yield
ABSTRACT
The ideotype is a theoretical model of an archetypal cultivated plant. Recent progress in genome editing is aiding the
pursuit of this ideal in crop breeding. Breeding is relatively straightforward when the traits in question are monogenic in
nature and show Mendelian inheritance. Conversely, traits with a diffuse, polygenic basis such as abiotic stress resistance
are more difficult to harness. In recent years, many genes have been identified that are important for plant domestication
and act by increasing yield, grain or fruit size or altering plant architecture. Here, we propose that (a) key monogenic
traits whose physiology has been unveiled can be molecularly tailored to achieve the ideotype; and (b) wild relatives of
crops harboring polygenic stress resistance genes or other traits of interest could be de novo domesticated by manipu-
lating monogenic yield-related traits through state-of-the-art gene editing techniques. An overview of the genomic and
physiological challenges in the world’s main staple crops is provided. We focus on tomato and its wild Solanum (section
Lycopersicon) relatives as a suitable model for molecular design in the pursuit of the ideotype for elite cultivars and to
test de novo domestication of wild relatives.
© 2016 Published by Elsevier Ltd.
Plant Science xxx (2016) xxx-xxx
Contents lists available at ScienceDirect
Plant Science
journal homepage: www.elsevier.com
Review article
Genome editing as a tool to achieve the crop ideotype and de novo domestication of
wild relatives: Case study in tomato
Agustin Zsögön a, 1, Tomas Cermak b, 1, Dan Voytas b, Lázaro Eustáquio Pereira Peres c, ⁎
aLaboratory of Molecular Plant Physiology, Departamento de Biologia Vegetal, Universidade Federal de Viçosa, 36570-900 Viçosa, MG, Brazil
bDepartment of Genetics, Cell Biology and Development, Center for Genome Engineering, University of Minnesota, Minneapolis, MN 55455, USA
cLaboratory of Hormonal Control of Plant Development, Departamento de Ciências Biológicas, Escola Superior de Agricultura “Luiz de Queiroz”, Universidade de São Paulo,
CP 09 13418-900 Piracicaba, SP, Brazil
1. Introduction
Genome editing is an alternative to conventional breeding in crops
where a large amount of genetic/genomic resources are available.
The biggest advantage of this technology is that it enables molecu-
lar breeding of crops with specific properties. Molecular breeding re-
quires precise a priori knowledge of plant physiology and molecu-
lar genetics. With this knowledge and with state-of-the-art gene edit-
ing techniques, custom modifications, defined here as molecular de-
sign, can be targeted to specific genes to improve particular traits in a
predictive manner. Molecular design differs from conventional breed-
ing, which is empirical and where new beneficial traits are achieved
through harnessing the variation resulting from traditional breeding
methods, such as interspecific and intergeneric crosses or through nat-
ural, radiation and chemical mutagenesis.
The fast progress of genome editing in all fields of molecular biol-
ogy was possible thanks to the development of sequence-specific nu
Abbreviations: CRISPR, clustered regularly interspaced short palindromic re-
peats; ORF, open reading frame; QTL, quantitative trait locus; SNP, single nu-
cleotide polymorphism; TALENs, transcription activator-like effector nucleases;
ZFN, zinc finger nucleases
⁎Corresponding author.
Email address: lazaro.peres@usp.br (L.E.P. Peres)
1These authors contributed equally.
cleases, which introduce targeted double strand breaks in target loci
[1]. As opposed to the older sequence-specific nucleases systems such
as zinc finger nucleases (ZFNs) and meganucleases, whose specificity
is difficult to engineer, TALENs and more recently CRISPR/Cas9,
make it possible to target any desired sequence in the genome thanks
to their easily customizable DNA binding specificities [2]. The tar-
geted double strand breaks is repaired by non-homologous end joining
of the broken DNA or by homology directed repair. The former results
in short insertions or deletions (indels) causing frameshift mutations
that inactivate the gene of interest, whereas the latter enables precise
introduction of custom modifications. The intrinsic characteristics of
TALENs, CRISPR/Cas9 and DNA repair mechanisms leading to gene
editing have been described in great detail elsewhere [3,4]. Traits of
commercial value have already been created using genome editing in
rice [5], wheat [6], potato [7], soybean [8], and maize [9].
Here, we propose that the suite of genome editing techniques de-
scribed above could be used for two purposes. Firstly, to achieve the
ideotype, a guiding theoretical model of ideal cultivated plant, through
manipulation of key monogenic traits (whose physiology has been
unveiled) in elite varieties. Secondly, to manipulate monogenic do-
mestication-related traits in wild relatives of crops harboring poly-
genic traits of interest. The rationale for this is that while a large
number of monogenic determinants of yield have been characterized
in cultivated plants, the genetic basis for valuable polygenic traits
present in wild relatives (such as abiotic stress resistance) is diffuse
http://dx.doi.org/10.1016/j.plantsci.2016.12.012
0168-9452/© 2016 Published by Elsevier Ltd.

UNCORRECTED PROOF
2 Plant Science xxx (2016) xxx-xxx
and difficult to manipulate. Thus, instead of introducing alleles from
wild relatives into cultivated crops, as has been conventionally done
in classical breeding or in modern “rewilding” and “back-to-nature”
crop breeding, we aim at directly manipulating wild species at the
gene level to domesticate them de novo and harness their adaptation
to adverse environments. We call this approach de novo domestica-
tion. Here, we present an overview of molecular design of the ideo-
type and de novo domestication for the world’s top six staple crops
and develop them in-depth for a genetic model species with well de-
veloped genomic resources, ease of transformation, and known target
genes controlling specific traits, the tomato.
2. Molecular design in major crops and de novo domestication of
their wild relatives
The ideotype is a theoretical model of what an ideal crop plant
could be, proposed by Donald in 1968 [10]. Although originally pro-
posed for wheat, it can be extrapolated to any domesticated crop plant.
The idea behind the ideotype is that instead of selecting against defec-
tive traits, crop breeding should seek the achievement of model char-
acters.
Cereals have been subjected to divergent selective pressure during
domestication [11]. Selection in maize, sorghum and pearl millet has
led to a strong increase in apical dominance, suppressing side branch-
ing and concentrating seed production on a single, large terminal head
[11]. In maize, axillary branch number and length also decreased dur-
ing domestication, leading to the formation of the lateral ear (which
is itself an axillary branch). Wheat and rice, on the other hand, have
been selected for multiple tillers (lower level branches) that distrib-
ute grain production evenly, with relatively simultaneous maturation.
As will be discussed below, their height has also been reduced, as
a means to avoid lodging and thus, grain losses before harvest. Per-
haps the fact that the former crops have C4-type photosynthesis and
the latter have C3-type is more than just a coincidence. More tillering
(branching), and thus more self-shading, tends to be more detrimen-
tal to plants whose photosynthetic rate is optimized at higher irradi-
ances, such as C4 plants. On the other hand, tillering is advantageous
for weed control and to optimize the planting density. This is partic-
ularly true for crops such as paddy rice in Asia, where manual trans-
plantation of seedling is a millenary practice which would be very in-
efficient for crops with a single stem. Increased agricultural mecha-
nization for sowing and transplanting could thus lead to a concomitant
alteration of the ideotype, whereby a plant with a large, thick single
culm, that supports planting at higher density, would be desirable to
avoid unproductive tillers, enhanced grain yield per panicle and an el-
evated lodging resistance [12]. This proves that the ideotype is a fluid
concept which depends on the agronomic and social context.
Yield of some of the world's most important crops is restricted
by abiotic stresses such as salinity and drought [13,14]. In spite of
very few exceptions [15,16] resistance to such stresses tends to be
of complex polygenic nature, as it involves various levels of adjust-
ment of plant development, from the cell (e.g. Na+exclusion mech-
anisms) to the whole plant (e.g. source-sink relationships) [17]. In
most crops, the existence of wild relatives adapted to challenging en-
vironments provides suitable raw material for de novo domestication
through molecular design (Table 1) [18]. Effective methods for deliv-
ery of DNA into a species of interest are also a pre-requisite for mole-
cular design, and plant transformation has been successfully achieved
for the major crops discussed in this article [19]. A further condition
is in-depth knowledge of the genetic basis of the traits to be domesti-
cated in the crops of interest, to provide suitable targets for manipula-
tion of their wild relatives [20]. Below we discuss these topics for the
Table 1
Wild relatives of world's top crops with potential for de novo domestication.
Crop
Wild relatives with
domestication potential Relevant traits/references
Maize (Zea mays) Teosinte and Tripsacum Biotic and abiotic stress
resistance and apomixis [109].
Wheat (Triticum
aestivum)
Triticum dicoccoides,T.
turgidum L. ssp. durum
Grain protein. Low molecular
weight glutenins [109].
Rice (Oryza
sativa)
Oryza rufipogon and O.
longistaminata
Tolerance to aluminium and
drought avoidance [110]
Potato (Solanum
tuberosum)
Solanum demissum,S.
stoloniferum
Disease resistance [111]
Cassava (Manihot
esculenta)
Manihot glaziovii, M.
neosana
Enlarged roots and apomixis
[31]
Soybean (Glycine
max)
Glycine soja Improved root architecture
[112]
six major staple food crops: maize, wheat, rice, potato, cassava and
soybean. A summary of relevant genes, the traits they control and the
corresponding references is provided in the Supplementary Table S1.
2.1. The genetic basis of domestication of the three major cereal
crops
Maize (Zea mays), wheat (Triticum aestivum) and rice (Oryza
sativa) are the three most important crops in terms of worldwide grain
production. Maize domestication best illustrates the kind of modifi-
cations that transform a wild plant into a cultivated one. These mod-
ifications involve changes in plant architecture, plant source-sink re-
lationships and altered response to environmental cues. Although the
morphology of the maize plant is drastically different from that of its
putative wild progenitor, teosinte, classical association mapping work
pinned down the differences to just six regions in the genome [21].
Subsequent studies provided a thorough characterization of some of
the relevant QTLs and genes.
Among the genes that distinguish cultivated maize from its wild
progenitor, two control plant architecture: teosinte branched1 (tb1)
and grassy tillers1 (gt1). In the former, a gain-of-function mutation in
a TCP-family transcription factor leads to inhibition of side branch-
ing, altering source-sink relationships and increasing yield. In the case
of gt1, a more subtle cis-regulatory change affecting gene expression
is responsible for the reduced number of ears on a plant. Leaf inser-
tion angle is another maize trait that strongly influences yield. Upright
leaves in modern varieties are due to mutations in two genes, ligule-
less1 and 2(lg1 and lg2). Lg1 encodes a squamosa promoter-binding
protein and lg2 abZIP transcription factor. The expression of both is
reduced in domesticated maize varieties compared to wild relatives.
Increased number of kernel rows is also a hallmark of domestication in
maize. Point mutations that cause amino acid substitutions in the LRR
domain of the FASCIATED EAR2 (FEA2) gene, the maize orthologue
of CLAVATA2, can produce substantial increases in kernel row num-
ber. The transition from encased to naked kernels is under the con-
trol of an SBP-box transcriptional regulator, teosinte glume architec-
ture1 (tga1), whose causative polymorphism was recently identified
as a SNP that creates an amino-acid substitution.
Teosinte flowers in tropical short-day conditions, so the spread
of maize to more temperate areas necessitated rendering cultivated
maize photoperiod insensitive. This was achieved by selection of a
loss-of-function allele of a CCT-containing protein associated with
photoperiod response. The mutant allele harbors a transposable ele-
ment in its promoter, which significantly reduces flowering time in
temperate cultivars of domesticated maize. In summary, a large num-
ber of major genes responsible for the domestication syndrome from

UNCORRECTED PROOF
Plant Science xxx (2016) xxx-xxx 3
the grass-like Teosinte to cultivated maize have been well character-
ized at the molecular level. Targeted editing of these genes could be
used to engineer other wild relatives of maize (e.g.Tripsacum) as a
first step towards the production of a domesticated form.
The domestication of rice followed a complex pattern. Its direct
ancestor is the perennial wild species Oryza rufipogon, which, has
prostrate growth habit, with many tillers inserted at wide angles and
‘easy-to-shatter’ grain types, with few seeds. A number of key do-
mestication-related genes have been cloned in rice, although they are
still not sufficient to provide a complete picture of the domestica-
tion process [22]. A major transition was the switch to erect growth
habit, which has been ascribed to an amino acid substitution in the
PROG1 transcription factor, which also affects tiller number and an-
gle. Grass-specific genes controlling tiller angle have been identified,
including LAZY1 (LA1) and TAC1. Many genes were cloned which af-
fect seed shattering (Table S1). This trait is of particular interest, as the
seed-shattering habit is still a target in the breeding of new O. sativa
subsp. indica cultivars [23].
Among genes affecting seed-shattering is qSH1, a BEL1-type
homeobox gene that controls the formation of the abscission layer at
the base of the rice grain. The allele in domesticated varieties contains
a SNP in the 5′ regulatory region that leads to loss of seed shatter-
ing. Sh4, a Myb3 transcriptional regulator, harbors a SNP that causes
an amino acid substitution. Other domestication alleles have arisen
through more drastic genetic alterations. Such is the case of Rc, which
leads to white pericarp (opposed to red in the wild species), GW2,
which increases grain width and weight, and BAD2, which influences
grain fragrance and flavor. A dramatic increase in rice productivity
was achieved more recently with the introduction of semi-dwarf rice
cultivars during the so-called Green Revolution [24]. Later, the semi-
dwarf1 (sd1) mutant was characterized as a loss-of-function muta-
tion in a key gibberellin (GA) biosynthesis gene, GA20-oxidase. An
indirect way to increase rice productivity is the facilitation of grain
processing and storage. This was attained during rice domestication
through a frame-shift deletion in the LONG AND BARBED AWN1
(LABA1) gene, which encodes a cytokinin-activating enzyme homo-
logue to the LONELY GUY 1 (LOG1) gene (see below). Cultivated
rice harbors the laba1 allele, which reduces active cytokinin concen-
tration in awn primordia, disrupting barb formation and awn. The bar-
bless phenotype increases the efficiency of rice processing and ease of
rice storage, although it is not present in some temperate japonica lan-
draces, where the long barbed phenotype is preferred, probably due to
its effect deterring seed predation by birds.
In most cases discussed here, the alleles in domesticated rice carry
deletions causing protein truncation, an alteration that could be easily
reproduced using available gene editing technology.
The major gene responsible for domestication in wheat is Q, a ho-
mologue of the APETALA2 family of transcription factors. Point mu-
tations in Qresult in pleiotropic effects including free threshing and
speltoid spike format [12]. GPC-B1, a NAC-family transcription fac-
tor, controls leaf senescence and nutrient remobilization to the seeds.
The functional allele for GPC from wild accessions improves seed
protein, zinc and iron content. This positive allele was lost during do-
mestication in durum wheat. A more recent breakthrough was the in-
troduction of mutations in Rht, a DELLA repressor protein. Such mu-
tations stabilize the protein, rendering it insensitive to gibberellin-in-
duced degradation, and cause phenotypes similar to those conferred
by gibberellin-deficient Sd1 rice (see above). Thus, fine-tuning of
plant shoot architecture through the introduction of gibberellin-related
mutant alleles appears to be an overarching theme in increased ce-
real productivity. This can now be achieved through targeted alter
ation of the DELLA gene, as has recently been done in tomato using
TALENs [25].
2.2. The genetic basis of domestication of the three major non cereal
crops
After the cereals described above, potato (Solanum tuberosum),
cassava (Manihot esculenta) and soybean (Glycine max) are the most
important staple crops worldwide. Potato and cassava were domesti-
cated mainly to provide sources of starch. For soybean, high oil and
protein content are the traits of interest. Besides being the world’s fifth
most important crop, cassava is one of the main staple food crops in
tropical America and Africa.
As for many crops domesticated in South America, potato and cas-
sava were selected to accumulate the edible portion belowground, a
relatively safer environment that is protected from most herbivores
and some forms of abiotic stress. In both cases, tuberization does not
involve de novo organogenesis; rather pre-existing organs (stolons in
potato and roots in cassava) are converted into strong sinks. This im-
plies that cytokinins, the main hormone class controlling sink estab-
lishment [26], probably has a pivotal role on tuberization [27]. Indeed,
overexpression of the LONELY GUY 1 (LOG1) gene, encoding a cy-
tokinin-activating enzyme, converts axillary shoots into tuber-like or-
gans in transgenic tomato plants [28]. However, besides cytokinin ac-
tion to establish a strong sink, a specific signal triggering tuberization
is also required. Grafting experiments using tomato mutants as scions
onto potato rootstocks suggested that the former probably fail to pro-
duce such a mobile trigger involved in the conversion of potato stolons
into tuber sinks [27]. Consistent with this idea, it was later found that
SP6A, a member of the FLOWERING LOCUS T florigenic signal fam-
ily, whose tomato allele is inactive [29], is the main shoot-derived sig-
nal that triggers tuber formation in the underground stolon.
Potato, like many annual and biannual species that evolved in tem-
perate climates, develops storage organs that remain dormant in the
soil during adverse cold and dry periods [30]. This nutrient reserve
is reactivated in the next favorable season and nurses the vigorous
aboveground growth. Thus, South American natives took advantage of
a pre-existing mechanism in nature to domesticate potato. Contrary to
popular belief, this is not the case of cassava, whose wild relatives do
not normally form tuberous roots [31]. The edible part of cassava is an
artificial novelty created by pre-Columbian South Americans through
selection for roots that accumulate high quantities of starch over long
periods of time, ranging from 7 to 14 months. The common denomi-
nator, therefore, is the manipulation of source-sink relationships, and
it is irrelevant whether the sink already exists in nature in a weak form
(e.g. potato and carrot) or not (e.g. cassava and side buds in LOG1-OX
tomato). Unraveling the genetic determinants that convert organs into
storage sinks in model plants like potato or tomato could reveal target
genes for de novo domestication and improvement of other non-tuber-
ous wild species related to cassava (Table 1). Among such candidate
cassava wild relatives are its related tree-like species M. glaziovii,
which grows very large roots, and M. neosana, which can be propa-
gated by apomictic seeds, a way to reduce the incidence of many dis-
eases, mostly bacterial, that occur due to conventional propagation by
cuttings [31].
One of the most important genes in soybean domestication, Dt1
(Table S1) leads to determinate growth habit, as opposed to indeter-
minate growth in wild relatives such as Glycine soja. Soybean pod
ripening is synchronized with senescence (monocarpic senescence).
Growth habit variation combined with monocarpic senescence adjusts
the plant life cycle, which is shorter in determinate and longer in

UNCORRECTED PROOF
4 Plant Science xxx (2016) xxx-xxx
indeterminate plants. Soybean harbors at least five other Dt1 par-
alogues, whose manipulation could lead to the production of more ad-
vantageous, perhaps intermediate plant growth habits and life cycles.
This is an attractive prospect, as growth habit variation could help ad-
just the soybean crop to the agricultural calendar, following the old
adage “sow in the rain and harvest in the sun”. Variation in growth
habit, along with photoperiodic response changes, is the key driver be-
hind the spread of the soybean crop across an immense latitudinal gra-
dient, where both temperature and daylength can vary greatly.
As for other grain crops, non-shattering is an essential trait in soy-
bean, which was achieved through the selection of allelic variants
of a NAC-family transcription factor, SHAT1-5. Increased lignifica-
tion of the fiber cap cells prevents grain release from mother plants
and is conferred by a 3-bp deletion in the 5′ cis-regulatory region of
SHAT1-5. It remains to be determined if a similar mechanism, and the
orthologous genes, were also selected during domestication of other
important leguminous grains such as common bean (Phaseolus vul-
garis), lima bean (P. lunatus), mung bean (Vigna radiata) and cowpea
(V. unguiculata). It is worth noting that the non-shattering character
loses relevance when legumes are harvested as green pods, as is the
case of cowpea in northeast Brazil and the lima bean in Paraguay. In
these cases, the opposite phenotype, i.e. non-lignification of the pods,
would help the manual release of the beans from green pods, which
are preferred to dry beans, mainly because they require less energy for
cooking. Since the mutation described in soybean involves increased
expression of SHAT1-5, the opposite phenotype can be expected to oc-
cur in loss-of-function alleles created by gene editing.
The general mechanism underlying morphological evolution is be-
lieved to be alteration in gene expression levels [32]. These can be en-
gineered relatively easily using currently available gene editing tech-
nology. More subtle changes, such as alterations in protein function
resulting from single amino acid substitutions, are also involved in
a considerable number of domestication traits. These alterations are
technically more challenging to engineer, but certainly feasible in the
near future.
3. Molecular design to achieve the ideotype in tomato
Tomato is an ideal crop for molecular design, due to the extensive
knowledge of its basic biology and genetics, gained from decades of
conventional breeding and laborious “intuitive” phenotypic and
marker-assisted genotypic selection [33]. Here, we describe the most
important tomato traits and the underlying genes and discuss how they
can be edited to create the ideotype (Fig. 1). The ideotype can be bro-
ken down into the main themes of tomato breeding by dividing the
tomato plant into vegetative and reproductive modules. For the veg-
etative module, the main breeding goal is attaining the optimal plant
architecture, defined as the three dimensional organization of the plant
body [34]. In this review, we will consider only shoot architecture. For
the reproductive module, the targets are yield (which can be increased
by either fruit quantity or fruit size) and Brix (the content of soluble
solids, mostly sugars and organic acids in the fruit, measured as spe-
cific gravity).
In the last few years there has been an increased demand for
high-nutritional content of the fruits, and this can be achieved by rais-
ing the levels of carotenoids (β-carotene and lycopene), flavonoids
(anthocyanins) and vitamin C, among others. Tomato production is di-
vided in two main segments: fresh market and processing tomatoes.
For fresh market tomatoes, long shelf life is desired (although some-
times this is detrimental to flavor), as it reduces losses and allows the
producer to reach distant markets. For processing tomatoes, mechan-
ical harvesting is facilitated by synchronous ripening and fruits with
the “jointless” trait (see below). One further theme, which is related
to the whole-plant structure and life cycle, is resistance to stress, ei-
ther biotic (insect and pathogen) or abiotic (cold, drought and salinity),
which will not be covered in this article for the sake of simplicity.
3.1. The vegetative ideotype
3.1.1. Goals: reduction of sympodium and/or internode length,
reduction of lateral branching
Tomato is a sympodial species with either determinate or inde-
terminate growth habit [35]. “Sympodial” refers to the sympodium,
which, in the cultivated tomato, is the modular unit of three leaves
and an inflorescence (Fig. 1B), whose repeated concatenation (through
apical meristem conversion into a inflorescence and the resume of
vegetative growth in the proximal axillary meristem) makes up the
tomato plant [36]. “Indeterminate” growth means that plants grow
indefinitely by repeating sympodia, leading to a slender vine archi-
tecture; fruit set and ripening occurs sequentially, from bottom to
Fig. 1. Towards the tomato ideotype through molecular design. The ideotype is a theoretical model of what an agronomically ideal tomato plant could be. Most tomato cultivars
are far from this ideal phenotype. (A) Representative tomato plant of the cultivar Moneymaker, a genotype of indeterminate growth used for fresh market tomato production. Scale
bar = 20 cm. Molecular design could be used to bring about modifications to this and other similar tomato cultivars. (B) Tomato sympodia are typically composed of three leaves
(numbers) and an inflorescence (arrowheads). The succession of sympodia (white/yellow) form the core of the plant shoot. Reducing internode length and the number of leaves per
sympodium are highly sought breeding goals to increase productivity (see text). (C) Side branches (arrows) tend to act as sinks (net consumers) of photosynthates and are thus detri-
mental to plant yield. Side branching could be eliminated or reduced manipulating the correct gene targets. (D) Increased Brix (soluble solids), nutrient content (chiefly flavonoids,
carotenoids and vitamin C), parthenocarpy (fruit development without prior fertilization), concentrate ripening, long shelf life and elimination of the abscission layer (arrows) are
agronomically desirable goals that can be achieved through molecular design. Concentrate ripening and elimination of the abscission layer are important for mechanical harvesting of
processing tomatoes.

UNCORRECTED PROOF
Plant Science xxx (2016) xxx-xxx 5
top. Conversely, “determinate” plants, progressively reduce the num-
ber of leaves per sympodium until it is terminated in two consecutive
inflorescences; growth continues from more distal axillary meristems,
which further form terminal inflorescences, leading to a bushy, com-
pact plant with concentrated fruit set. Determinate cultivars are widely
used in the tomato processing industry, which has an once-over de-
structive harvest, whereas indeterminate ones are preferred for green-
house fresh tomato production, which has a continuous-harvest. How-
ever, in some places (e.g. California and Florida), determinate culti-
vars or hybrids are used for fresh market tomato production in the
open field. Wild type tomatoes, as well as tomato wild relatives are
indeterminate due to the presence of a functional allele of the SELF
PRUNING (SP) gene [37]. The loss-of-function sp allele creates de-
terminate plants. A variation on the determinate growth habit theme,
semi-determinate growth, is controlled by other members of the SELF
PRUNING family (called CETS family after CENTRORADIALIS from
Antirrhinum, TERMINAL FLOWER-1 from Arabidopsis and
SELF-PRUNING). There are indications that semi-determinate growth
combines advantages of the determinate and indeterminate growth
habit, including increased yield, Brix and water-use efficiency [38].
The genes of the CETS family can also be manipulated to achieve
sympodium reduction, i.e. two instead of three leaves plus an inflores-
cence per sympodium [39]. This family has both inducers and repres-
sors of flowering, and the difference between them can be as small as
a single amino-acid residue [40]. Increased expression of the flower-
ing inducer SINGLE FLOWER TRUSS (SFT) in an indeterminate (Sp)
background produces plants with two leaves per sympodium instead
of three [35]. Since this family has multiple members in tomato and
other Solanaceae [41], different allelic combinations could be tested
to find the best way to create compact, indeterminate plants with two
leaves per sympodium for fresh tomato production or determinate
and semi-determinate plants for the tomato processing industry. Such
modification in growth habit alters source-sink relationships and influ-
ences productivity [38].
The use of mutations reducing internode length, such as alterations
in gibberellin signaling in wheat and metabolism in rice, is wide-
spread in Green Revolution cereals, as mentioned previously. The tar-
get genes for this phenotype in tomato are the single-copy PROCERA
(PRO) gene encoding a DELLA protein [42] and the GA20 oxidase
(GA20-ox) gene family. The most relevant GA20ox genes in vegetative
tissues are GA20ox-1 and GA20ox-4 [43]. These two copies of GA20ox
have a relatively high expression in stems and low expression in fruits.
Although the DELLA mutant procera has abnormally long internodes,
an undesirable trait for field-grown tomatoes, it has increased Brix
[44] and parthenocarpic fruits [45]. This suggests that the construc-
tion of weak alleles for PRO could combine positive effects on Brix
and parthenocarpy (i.e. the development of fruits without prior fertil-
ization) while minimizing the pleiotropic internode elongation effect
[46]. In contrast, loss-of-function alleles of the GA20ox genes would
create compact plants without negative effects on the reproductive or-
gans.
A different way to reduce internode length was discovered in
maize and sorghum and consists in loss-of-function mutations in the
auxin MDR/ABC (multidrug resistant/ATP-binding cassette) trans-
porters [47]. MDR/ABC transporter homologues exist in tomato al-
though mutant alleles have not yet been described. It is possible that a
novel tomato phenotype with shorter internodes and higher productiv-
ity, as well as better suited for growth in confined greenhouse space,
could be created by editing such targets.
Removal of lateral branching is a long-sought breeding goal in
tomato, as it would eliminate the costly need for pruning of side
branches (suckers, Fig. 1C). A loss-of-function mutation in a regu-
latory GRAS-domain transcription factor (named after the first three
members of the family, GIBBERELLIC-ACID INSENSITIVE,
REPRESSOR of GAI and SCARECROW), called lateral suppressor,
produces such a phenotype. However, the mutations have negative
pleiotropic effects, such as very poor fruit set [48], which reduces
agronomic value. Editing the LATERAL SUPPRESSOR gene might
be a way to create weak alleles in order to prevent such undesirable
side effects. A more reliable way to achieve unbranched tomato plants
would be to edit the transcription factor BRC1a [49]. This branch-
ing inhibitor belongs to the TCP family (named after the orthologs
TEOSINTE BRANCHED from maize, CYCLOIDEA from Antir-
rhinum and PROLIFERATING CELL FACTORS from rice) whose ex-
pression is normally restricted to axillary buds. BRC1a expression,
however, is low, although not completely absent, in the cultivated
tomato [49]. The fact that expression of the BRC1a allele from the
wild relative S. pennellii is four times higher than that of tomato and
causes suppression of side branching, suggests that the expression of
the tomato orthologue could be increased using artificial transcrip-
tional activators or by editing the promoter of this gene [50,51].
3.2. The reproductive ideotype
3.2.1. Goals: improved fruit set, increased fruit Brix and nutritional
value, improved mechanical harvesting and long shelf life of fruits
Greenhouse tomato production is frequently hampered by insuf-
ficient pollination. Attempts to solve this are mechanical pollination
(shaking), which is rather impractical, and the use of bumblebees. Al-
though commercial sources of bumblebees are available, insects need
UV light for spatial navigation, so this practice conflicts with the
anti-UV technology generally used in plastic covers. The best way to
solve the problem of low fruit set in tomato is through parthenocarpy
[52]. This trait is not only relevant for greenhouse tomatoes, but also
for open-field production, as one of the expected consequences of cli-
mate change is a reduction of pollen germination and viability [53].
Although mutations leading to parthenocarpy have long been known
in tomato [54,55], genes such as pat have negative pleiotropic effects
or are difficult to introgress into relevant cultivars or hybrids (pat-2)
due to their poor penetrance. The molecular nature of these defective
alleles is hitherto unknown.
The tomato genome harbors 22 AUXIN RESPONSE FACTORS
(ARFs) [56], that are both negative and positive regulators of fruit set.
The Sl-ARF19 gene (formerly ARF7) is a negative regulator of fruit set
and its down-regulation in transgenic plants produced parthenocarpic
fruits [57]. Thus, non-transgenic parthenocarpic tomatoes could be ob-
tained by editing this gene to create a null allele. Other members of
the ARF family are positive regulators of fruit set whose expression
is controlled by micro-RNAs. Dominant alleles promoting partheno-
carpy could be created by editing these genes so as to eliminate the site
for microRNA recognition. Useful dominant alleles are always desir-
able in tomato breeding since they need only be incorporated into one
parent, facilitating the creation of hybrid lines.
A hurdle for tomato breeding is that Brix is inversely correlated
to yield. Breeders consider the product of Brix × ripe yield (BRY) the
single most relevant agronomic parameter for processing tomatoes.
The physiological explanation of this negative correlation is that both
Brix and yield compete for photosynthates, and thus this is a problem
of source/sink relationships. Mature leaves are net exporters of pho-
tosynthate (hence, sources), whereas flowers and fruits are net con-
sumers of it (sinks). An ingenious way to partially overcome this co-
nundrum is to make sinks contribute to the photosynthate budget by

UNCORRECTED PROOF
6 Plant Science xxx (2016) xxx-xxx
engineering photosynthesis in loco (i.e. fruit photosynthesis) through
chloroplast accumulation. Genes that control chloroplast development
have been described, e.g. Golden2-like (GLK2), HIGH PIGMENT1
(HP1) and HP2 [58–60]. The hp1 and hp2 mutants, in spite of produc-
ing fruits with increased Brix, carotenoids, and ascorbic acid [61,62],
are not agronomically viable due to their constitutive light response,
which reduces plant growth and fruit set. GLK2 expression is concen-
trated in the fruit and its level could be increased by targeted activation
or promoter swapping to induce production of higher levels of pho-
toassimilates resulting in increased Brix. It should be mentioned, how-
ever, that uniform ripening (u), which is a loss-of-function allele for
GLK2, has been incorporated into modern tomato varieties to prevent
uneven ripening. Thus, the current trend of consumer preference for
nutritional quality and flavor in detriment of appearance will define
the success of varieties harboring modifications on the GLK2 gene.
One further way to increase Brix is to increase sink strength. As
mentioned earlier, one of the main controllers of sink strength are
the cytokinin hormones along with their molecular target, the enzyme
invertase [28,63]. Manipulating either invertases or cytokinins can
lead to increased sink strength and therefore higher Brix [64,65]. Nat-
ural allelic variation in the gene for the extracellular invertase LIN5,
which is expressed in the ovary, can also increase fruit Brix. A SNP
in the LIN5 gene from S. pennellii LA716 created an invertase with
enhanced catalytic activity when introgressed into cultivated tomato
[65]. Gene editing could be used to further enhance LIN5 perfor-
mance. Alternatively, transcriptional activators could be engineered or
promoter mutations introduced to increase LIN5 expression.
Modulating cytokinins accumulation can lead to either increased
yield (accumulation early in inflorescence development, as in rice
[66]) or Brix (late accumulation in ovary development, [64]). Cy-
tokinins levels can be increased by lowering the expression of the
cytokinin-degrading enzyme CYTOKININ OXIDASE (CKX) [66] or
increasing expression of the cytokinin hydrolase enzyme LOG [67].
Whereas CKX inactivates cytokinins by irreversible cleavage of the
isoprenoid side chain of natural cytokinins (e.g. isopentenyl adenine,
zeatin), LOG releases cytokinins inactivated by ribosylation (e.g.
isopentenyl adenosine, zeatin ribosyde) [68]. A CKX gene expressed
in the fruit is a potential target for gene editing to create a loss-of-func-
tion allele. Gain-of-function alleles of the tomato LOG gene would be
expected to also result in fruits with higher Brix.
As mentioned above, novel alleles increasing expression of GLK2
or specifically lowering the expression of HP1 and HP2 in the fruit
would improve tomato nutritional value. In the case of HP1 and
HP2, this would simultaneously increase the content of carotenoids
and ascorbic acid. The hp2 mutant has also been used, in combina-
tion with the natural genetic variations atroviolacium (atv) and An-
thocyanin fruit (Aft), to create non-transgenic purple tomatoes accu-
mulating the anthocyanins delphinidin and petunidin [69] without the
high irradiance requirement for pigment production. The anthocyanin
levels are lower than in transgenic plants [70], but the conventionally
bred variety has the added advantage of increasing ascorbic acid and
carotenoids [69].
Anthocyanin consumption has been associated with protection
against a broad range of human diseases, including prevention of obe-
sity, diabetes and cardiovascular diseases, and improvement of vi-
sual and brain functions through modulation of signaling cascades and
gene expression [71]. A drawback is that anthocyanins accumulate
mostly in the tomato fruit skin, and less in the pulp, which is proba-
bly due to the low expression of chalcone isomerase (CHI) in tomato
fruit flesh [72]. Tomato harbors six CHI genes, all of which have
lower expression in the fruit tissue [73,74]. Solyc06g084260 has 4-
to 5-fold higher expression in the fruit tissue of the S. pennellii de
rived introgression lines IL6-3 and IL6-4, compared to the control
tomato cultivar M82 [73,75]. Bin 6G, defined by the ILs 6-3 and 6-4,
harbors the CHI allele from S. pennellii, a species whose higher CHI
activity is associated with improved flavonoid accumulation in fruits
[72]. Based on this information, the Solyc06g084260 gene could be
a suitable target for improving anthocyanin production in fruits, as
tomato carries an ortholog of that gene.
Three other traits could be considered in the reproductive ideo-
type: concentrate ripening, jointless pedicels and long shelf life,that
do not directly affect tomato productivity. The first two traits are im-
portant for processing tomato, since they improve mechanical harvest-
ing. Unfortunately, the genetic basis of concentrated ripening, which
depends on synchronous fruit set and/or the capacity of early set fruit
to be held on the plant until the late set fruit are ripe enough for
picking, is largely unknown. The jointless pedicel allows mechani-
cal harvesting of fruits avoiding pedicels and sepals, which are re-
tained in the mother plant due to the absence of the abscission layer.
Two genes, JOINTLESS1 (J1) and JOINTLESS2 (J2), control forma-
tion of the abscission layer in the pedicel. J1 is not used in tomato
breeding due to the pleiotropic effects caused by its loss-of-function
allele j1, such as indeterminate inflorescence (reversion into shoots)
and reduced flowering, which fit to the molecular nature of J1 as a
MADS-box family gene [76]. j2 is a natural genetic variation, de-
rived from S. cheesmanae, not allelic to j1, whose molecular identity
is unknown [77]. It is possible that alleles with less detrimental effects
could be created via gene editing of J1. Alternatively, epistasis with
the flowering inducer SFT (see above) might compensate for some of
the negative effects caused by the j1 mutation, such as reversion of in-
florescences to shoots.
Long shelf life can be obtained in tomato through gene dosage of
the mutations non ripening (nor), which codes for a NAC transcrip-
tion factor, and ripening inhibitor (rin), which codes for a MADS-box
family gene [78]. Although homozygous nor or rin plants produce
fruits lacking ripening attributes such as softening, flavor, pigment and
sugar accumulation, their heterozygous forms can be used to extend
the ripening period and delay deterioration. It is interesting to note that
tomato landraces from Spain and Portugal with extended shelf life har-
bor the alcobaça mutation (named after the namesake city in Portu-
gal), an allele of nor [79]. A concern exists that the long shelf life pro-
vided by rin and nor decreases fruit flavor, so alternative physiological
pathways, or less pleiotropic components downstream of the rin and
nor pathways [80], could be explored to edit the corresponding genes.
4. De novo domestication of wild relatives of tomato by gene
editing
During the domestication and breeding of tomato, a great number
of traits were improved through stacking of QTLs and individual al-
leles. Further improvement can be achieved through the use of exotic
material in tomato, since the tomato clade (Solanum section Lycoper-
sicon) is richly endowed in natural genetic variation: 13 related wild
species exist that can be readily crossed and interbred [81]. The use
of wild species as sources of useful genetic variation is only suitable
if a monogenic trait is to be introgressed into cultivated tomato. Con-
versely, polygenic traits would quickly be lost through segregation in
the first backcrosses to the recurrent tomato parent. Thus, we propose
molecular edition of a suite of domestication genes in the wild species
to capture polygenic traits of agronomic interest, an approach we call
de novo domestication (Fig. 2).
The main hurdles to be overcome in most wild species are self-in-
compatibility, exerted style and high alkaloid accumulation in the
fruit. Some wild species are difficult to reproduce, since they are

UNCORRECTED PROOF
Plant Science xxx (2016) xxx-xxx 7
Fig. 2. De novo domestication compared to traditional breeding. Diagram comparing the tomato (Solanum lycopersicum) traditional breeding approach (top) and the molecular de
novo domestication approach presented in this article (bottom). Notice that there is no real ‘gene flux’ from domesticated tomato to wild species, rather knowledge on the genetic
basis of tomato will be used to modify the ortholog genes in the wild species, as exemplified with S. galapagense (see Fig. 3) in the bottom. Potential target genes to be edited in the
process of domestication of S. galapagense are indicated below the arrow and detailed in Table 2.
self-incompatible and have exerted styles, which precludes self-polli-
nation and seed production. These traits could be overcome by creat-
ing null alleles for the S-RNAse and STYLE2.1 genes (Table 2). Such
modifications could potentially turn wild self-incompatibility species
into self-compatible ones and produce recessed styles in their flow-
ers. Regarding alkaloid content, it should be noted that even cultivated
tomato has high levels in unripe fruits, which are eventually lost dur-
ing ripening. This is also the case for many of the red-fruited wild rel-
atives in the Lycopersicon section, which have edible fruits albeit very
small ones [82].
A case study could be Solanum galapagense (Fig. 3), a self-com-
patible species endemic to the Galapagos Islands which grows close
to the sea shore and tolerates as much as 70% of sea water salinity
(ca. 350 mM NaCl) [83]. S. galapagense also shows one of the highest
Brix values (ca. 15°) of all wild tomato relatives and high resistance
to white fly, due to the presence of type IV trichomes and acyl sugars
[83,84]. The seeds show strong dormancy, that is usually broken by
treatment with bleach [82]. The fruits, though edible, are very small
(pea-size) and do not accumulate lycopene, but rather beta-carotene,
due to the presence of the Beta (i.e. functional) allele of LYCOPENE
BETA CYCLASE (Cyc-B) [85].
Mutations in FW2.2 [86], LOCULE NUMBER [87] and FASCI-
ATED [88] are in part responsible for the increased fruit size in cul-
tivated tomato, so their S. galapagense orthologues could be edited
as a first step towards domestication. While genetic variation in the
negative regulator of cell division, FW2.2, can increase fruit size by
30%, mutations at the LOCULE NUMBER (LC) and FASCIATED
(FA) loci increase the number of locules in the fruit and have even
larger effects on fruit weight than mutations in FW2.2 (up to 50%
increase in fruit size) [89]. The lc phenotype is due to a regulatory
mutation downstream of WUSCHEL, which affects the binding of
transcription factors and thus stem cell fate [87]. The fas mutation is
caused by an inversion in the regulatory region of a tomato homo
logue of the CLAVATA3 gene, whose product also participates in stem
cell determination [88]. Creating altered alleles of these genes in S.
galapagense could result in cherry-size tomatoes which, coupled to
a high Brix and orange color (high content of beta-carotene, a pre-
cursor of vitamin A), would fit the growing market for high nutrient
and gourmet tomatoes. A red-fruited version, which retains the highly
valuable salt and insect resistance could subsequently be created by
editing Cyc-B, whose loss-of-function allele in tomato is known as
crimson/old gold due to the high lycopene accumulation [85]. Yield
would probably be still low in the new domesticate due to the high
amount of energy spent by the plants to cope with salinity, but would
still allow cultivation in areas with high salt content in the irrigation
water.
We have compiled a list of additional genes that could be edited
(Table 2) to create favorable alleles. Among them are ovate and com-
pound inflorescence, whose mutant alleles will, respectively, create
grape-type tomatoes and increase the number of flowers/fruits per in-
florescence [90,91]. For cherry and grape tomatoes, enhanced number
of fruits per cluster is a highly desirable trait to compensate for their
small size. The ovate mutation is believed to improve the organoleptic
characteristics of the fruit, since grape-type fruit provide a “popping in
the mouth” sensation, probably due to the reduced number of locules
and water accumulation.
5. Technical challenges of molecular design
To exploit the full potential of the molecular design approach, all
types of genome editing techniques will need to be utilized. These
include gene inactivation or repression, base editing to create
amino-acid substitutions resulting in new allelic variants or to re-cre-
ate wild alleles in elite tomato genotypes, and activation of gene ex-
pression. Thanks to the rapid advancement of the genome engineer-
ing field, the technology for making different types of modifications

UNCORRECTED PROOF
8 Plant Science xxx (2016) xxx-xxx
Table 2
Genes affecting productivity and fruit quality in tomato discussed in this article, which
could be targeted for molecular design and de novo domestication.
Gene name Gene ID Gene function Reference
AUXIN RESPONSE
FACTOR 19
(ARF19)
Solyc07g042260 Negative regulator of fruit
set
[57]
BRANCHED1a
(BRC1a)
Solyc03g119770 Controls axillary
branching
[49]
CHALCONE
ISOMERASE (CHI)
Solyc06g084260 Participates in flavonoid
biosynthesis
[72]
COMPOUND
INFLORESCENCE
(S)
Solyc02g077390 Affects number of flower/
fruits per inflorescence
[91]
CYTOKININ
OXIDASE (CKX)
Solyc10g079870 Cytokinin inactivation [66]
FASCIATED (FAS) Solyc11g071380 Controls number of
carpels/locules in the fruit
[88]
FRUIT WEIGHT 2.2
(FW2.2)
Solyc02g090730 Quantitative variation of
fruit size
[86]
GA20-OXIDASE 1 Solyc03g006880 Gibberellin biosynthesis [43]
GA20-OXIDASE 4 Solyc01g093980 Gibberellin biosynthesis [43]
GOLDEN2-LIKE
(GLK2)
Solyc10g008160 Controls chloroplast
development in fruits
[60]
HIGH-PIGMENT 1
(HP1)
Solyc02g021650 Controls
photomorphorgenic
responses, including
anthocyanin
accumulation in fruits
[58]
HIGH-PIGMENT 2
(HP2)
Solyc01g056340 Controls
photomorphorgenic
responses, including
anthocyanin
accumulation in fruits
[59]
JOINTLESS1 (J1) Solyc11g010570 Controls abscission zone
formation in pedicels
[76]
LATERAL
SUPPRESSOR (LS)
Solyc07g066250 Controls axillary
branching
[113]
LIN5 Solyc09g010080 Extracellular invertase
expressed in the fruit
ovary
[65]
LOCULE NUMBER
(LN)
Vicinity of
Solyc02g083950
Regulates locule number
in fruits
[87]
LONELY GUY 1
(LOG1)
Solyc08g062820 Cytokinin hydrolase –
releases cytokinins
inactivated by
ribosylation
[67]
LYCOPENE BETA
CYCLASE (Cyc-B)
Solyc06g074240 Conversion of lycopene
into B-carotene
[85]
NON RIPENING
(NOR)
Solyc10g006880 Controls the initiation of
the normal fruit ripe
program
[78]
OVATE (O) Solyc02g085500 Regulates fruit shape [90]
PROCERA (PRO) Solyc11g011260 Repressor of gibberellin
signaling
[42]
RIPENING
INHIBITOR (RIN)
Solyc05g012020 Controls ripening-related
ethylene biosynthesis
[78]
SELF-PRUNING
(SP)
Solyc06g074350 Responsible for sympodial
and indeterminate growth
habit
[37]
SINGLE FLOWER
TRUSS (SFT)
Solyc03g063100 Flowering inducer [114]
S-RNAse Solyc01g055200 Controls Self-
incompatibility
[115]
STYLE2.1 Solyc02g087860 Responsible for style
exsertion from the
anteridial cone
[116]
is in place. Knock-outs of tomato genes can be readily isolated us-
ing TALENs [46] and CRISPR/Cas9 [92]. In cases where a full gene
knock-out is not desirable, gene expression can be tuned down using
TALE- [93] or CRISPR/Cas9- [94] based transcriptional repressors.
Although targeted transcriptional repression has not yet been demon-
strated in tomato, it was successfully used in Arabidopsis thaliana
Fig. 3. A potential target species for de novo domestication in the tomato clade. Repre-
sentative flowering plant of the wild tomato relative Solanum galapagense. This species
is endemic to the Galapagos Islands in South America and can withstand high salinity
levels in the soil, and is insect resistant due to the presence of trichome-produced insec-
ticide compounds. Scale bar = 5 cm. Inset: mature fruits. Scale bar = 1 cm. Though edi-
ble and with increased soluble solids (high Brix), the fruits are small, round and orange
due to the accumulation of beta-carotene, instead of lycopene as in cultivated tomato.
[50,93] and Nicotiana benthamiana [51], a relative of tomato. Further,
changes in gene expression and cell tissue specificity can be achieved
by mutating promoters and regulatory sequences.
While the above mentioned types of modifications proceed through
the non-homologous end joining, which is the most frequent DNA re-
pair mechanism in somatic cells, so far the most efficient way to in-
duce customized base substitutions, targeted insertions or other types
of precise modifications has been the use of homology directed repair.
In such experiment, the sequence-specific nucleases engineered to in-
duce a double strand breaks in the gene of interest and is delivered
to the cell along with a DNA donor template containing the desired
modification flanked by homology to the break site. The homologous
recombination repair pathway can use this template to copy informa-
tion and seal the break, introducing the custom sequence modifica-
tion into the gene of interest. This approach is much more challenging
in plants due to the fact that homology directed repair is not the pre-
ferred type of repair and due to inefficient methods of DNA delivery
to plant cells. Nevertheless, we have recently successfully applied this
approach in tomato and achieved relatively high frequency of precise
modification without random integration of foreign DNA through the
use of geminiviral replicons [95]. Alternatively, base substitutions can
also be induced by the recently developed reagents that are fusions be-
tween Cas9 and cytosine deaminases [96].
Additionally, there are two ways of engineering increased lev-
els of gene expression. Gene activation with engineered TALE- or
CRISPR/Cas9 transcriptional activators has been demonstrated in to-
bacco [97], N. benthamiana and Arabidopsis [50,51]. Another way of

UNCORRECTED PROOF
Plant Science xxx (2016) xxx-xxx 9
inducing stable expression is targeted insertion of a strong promoter
upstream of the gene of interest. We have used this approach to ac-
tivate the ANT1 gene controlling anthocyanin biosynthesis in tomato
and achieved robust expression [95]. Although we have used a vi-
ral promoter in this work, a number of constitutive and tissue spe-
cific native tomato promoters have been characterized [98,99] and can
be used to avoid regulatory hurdles with plants containing viral se-
quences.
To create more complex traits by editing several genes simulta-
neously, advantage can be taken of the multiplexing capability of
the CRISPR/Cas9 system. The most important difference between
CRISPR/Cas9 and the other types of sequence-specific nucleases is
that the targeting specificity of the CRISPR/Cas9 system is that DNA
targeting is determined by the small guide RNA (gRNA) molecule and
not by the protein itself. Therefore, targeting multiple sites is done
simply by expressing gRNAs targeting several genes along with the
Cas9 protein. Several systems have been developed for simultaneous
production of multiple gRNAs in plant cells and up to 8 genes have
been simultaneously edited [50,100]. Moreover, the recent finding that
gRNAs shorter than 16 nucleotides are still capable of tethering the
Cas9 protein to the target sequence, but cause drastic reduction in
the nuclease activity [101], opens up the door for combining different
types of modifications in different genes.
With this in mind, one can simultaneously induce double strand
breaks in one gene and activate or repress the expression of another
by the use of a single, catalytically active Cas9 protein fused to a
transcription activation or repressor domain and synchronous produc-
tion of full length, 20 bp gRNAs for double strand breaks induction
and shorter, 16 bp gRNAs for transcriptional regulation. Furthermore,
the gRNA sequence can be extended with modular RNA domains
that recruit effector (e.g. activator or repressor) domains fused to an
RNA-binding protein to the Cas9-bound target sites [102]. This re-
moves the need for fusing the activator domain directly to the Cas9
protein and allows for synchronously combining more than two func-
tions. For example, DNA cleavage, transcription activation, and re-
pression could be triggered at the same time, which would offer more
flexibility in molecular design. In theory, all types of modifications in-
cluding gene knock-outs, allele replacements, targeted insertion, gene
activation and silencing could be combined in a single experiment,
making it possible to model of more advanced traits that involve inter-
action of several genes.
Another requirement for genome editing in tomato is the deliv-
ery of the vector containing the sequence-specific nucleases and the
donor template to the plant cell. Tomato transformation protocols are
available, although the great majority of them are based on multiple
steps and involve the use of expensive chemicals such as the cytokinin
zeatin. The wild allele Rg1 [103] or the use of an auxin-containing
root-inducing medium during the two days of acquisition of cell com-
petence in tomato greatly improves its regeneration capacity and thus
the genetic transformation efficiency [104]. These improved proto-
cols could allow the use of the less expensive cytokinin benzyl amino
purine and reduce the number of steps involved. As for the genetic
transformation of wild species for de novo domestication, most of the
wild species have greater in vitro regeneration capacity compared to
cultivated tomato [105]. In summary, all the necessary tools are avail-
able to make molecular design in tomato feasible.
6. Future perspectives and concluding remarks
The advent of genome engineering should greatly enhance the ca-
pacity to fast-track breeding in multiple crops through molecular de-
sign of the ideotype. Being both a genetic model organism and a
leading horticultural crop, tomato lies at the crossroads between basic
and applied research. We suggest that the most profitable way to ap-
ply the revolutionary gene editing toolkit is to combine state-of-the-art
biotechnological tools and decades of accumulated knowledge on ge-
netics, physiology and ecology of tomato and its wild relatives. Ma-
nipulation of target genes should be the result of careful planning, re-
lying mostly on mining of the vast literature on all aspects of tomato
biology, particularly its physiology [106]. We also propose, and are
currently working towards, domestication of wild tomato relatives us-
ing genome engineering. This molecular de novo domestication ap-
proach should help us exploit valuable genomic resources extant in
the wider gene pool of tomato. Its extension to other horticultural
crops of interest would represent a logical next step. However, it is
worth noting that tomato is a true diploid species, unlike many sta-
ple crops, which are allopolyploid, as many cereals, or autopolyploid,
as the potato. Moreover, most examples presented here are recessive
mutations with high impact on productivity, which is not always the
rule, since most quantitative traits are semi-dominant in their action
[107]. These complex genomic contexts, in which gene dosage [107]
and epitasis [108] play a large role, will demand fine-tuned gene edit-
ing which can only be achieved through a synergistic effort of plant
physiologists, geneticists and breeders.
Acknowledgments
AZ and LEPP received fellowships from FAPESP (2013/11451-2)
and CNPq (307040/2014-3), respectively. This work was partially
funded by FAPESP (2015/50220-2). We thank Professor Adriano
Nunes-Nesi (University of Viçosa – Brazil) for critical reading of an
early version of the manuscript. We thank Dr. Frederico Almeida de
Jesus, Mateus Henrique Vicente and Diego Sevastian Reartes for help
with photos. We are also grateful to Professor Joerg Kudla (Univer-
sity of Münster, Germany), Dr. Leonardo Boiteux (EMBRAPA Hor-
taliças, Brazil) and Dr. Matías González (INIA, Uruguay) for helpful
discussion in developing the concepts presented in this article. Five
anonymous reviewers and the editor are gratefully acknowledged for
insightful comments on the manuscript.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in the
online version, at http://dx.doi.org/10.1016/j.plantsci.2016.12.012.
References
[1] D.F. Voytas, Plant genome engineering with sequence-specific nucleases,
Annu. Rev. Plant Biol. 64 (2013) 327–350.
[2] S.M. Schaeffer, P.A. Nakata, CRISPR/Cas9-mediated genome editing and
gene replacement in plants: transitioning from lab to field, Plant
Sci. 240 (2015) 130–142.
[3] D.F. Voytas, C. Gao, Precision genome engineering and agriculture: opportu-
nities and regulatory challenges, PLoS Biol. 12 (2014) e1001877.
[4] N.J. Baltes, D.F. Voytas, Enabling plant synthetic biology through genome en-
gineering, Trends Biotechnol. 33 (2015) 120–131.
[5] T. Li, B. Liu, M.H. Spalding, D.P. Weeks, B. Yang, High-efficiency
TALEN-based gene editing produces disease-resistant rice, Nat. Biotech-
nol. 30 (2012) 390–392.
[6] Y. Wang, et al., Simultaneous editing of three homoeoalleles in hexaploid
bread wheat confers heritable resistance to powdery mildew, Nat. Biotech-
nol. 32 (2014) 947–951.
[7] B.M. Clasen, et al., Improving cold storage and processing traits in potato
through targeted gene knockout, Plant Biotechnol. J. 14 (2015) 169–176.
[8] W. Haun, et al., Improved soybean oil quality by targeted mutagenesis of the
fatty acid desaturase 2 gene family, Plant Biotechnol. J. 12 (2014) 934–940.

UNCORRECTED PROOF
10 Plant Science xxx (2016) xxx-xxx
[9] J. Shi, et al., ARGOS8 variants generated by CRISPR-Cas9 improve maize
grain yield under field drought stress conditions, Plant Biotechnol. J. (2016)
1–10.
[10] C.M. Donald, The breeding of crop ideotypes, Euphytica 17 (1968) 385–403.
[11] J.R. Harlan, Crops and Man, American Society of Agronomy, Madison, WI,
1975.
[12] Y. Jiao, et al., Regulation of OsSPL14 by OsmiR156 defines ideal plant archi-
tecture in rice, Nat. Genet. 42 (2010) 541–545.
[13] J. Gressel, Genetic Glass Ceilings: Transgenics for Crop Biodiversity, Johns
Hopkins University Press, Baltimore, MD, 2008.
[14] A. Fischer, D. Byerlee, G. Edmeades, Crop Yields and Global Food Security:
Will Yield Increase Continue to Feed the World?, Australian Centre for Inter-
national Agricultural Research, Canberra, Australia, 2014. http://aciar.gov.au/
publication/mn158.
[15] M.P. Apse, G.S. Aharon, W.A. Snedden, E. Blumwald, Salt tolerance con-
ferred by overexpression of a vacuolar Na+/H+ antiport in Arabidopsis, Sci-
ence 285 (1999) 1256–1258.
[16] R. Munns, et al., Wheat grain yield on saline soils is improved by an ancestral
Na+transporter gene, Nat. Biotechnol. 30 (2012) 360–364.
[17] A. Albacete, et al., Ectopic overexpression of the cell wall invertase gene CIN1
leads to dehydration avoidance in tomato, J. Exp. Bot. 66 (2015) 863–878.
[18] R. Hajjar, T. Hodgkin, The use of wild relatives in crop improvement: a survey
of developments over the last 20 years, Euphytica 156 (2007) 1–13.
[19] K. Wang, Agrobacterium Protocols, vol. 2, Springer, New York, 2015.
[20] J.F. Doebley, B.S. Gaut, B.D. Smith, The molecular genetics of crop domesti-
cation, Cell 127 (2006) 1309–1321.
[21] J. Doebley, A. Stec, J. Wendel, M. Edwards, Edwards, Genetic and morpho-
logical analysis of a maize-teosinte F2 population: implications for the origin
of maize, Proc. Natl. Acad. Sci. U. S. A. 87 (1990) 9888–9892.
[22] M.J. Kovach, M.T. Sweeney, S.R. McCouch, New insights into the history of
rice domestication, Trends Genet. 23 (2007) 578–587.
[23] T. Sang, S. Ge, Genetics and phylogenetics of rice domestication, Curr. Opin.
Genet. Dev. 17 (2007) 533–538.
[24] P. Hedden, The genes of the green revolution, Trends Genet. 19 (2003) 5–9.
[25] S. Livne, et al., Uncovering DELLA-Independent gibberellin responses by
characterizing new tomato procera mutants, Plant Cell 27 (2015) 1579–1594.
[26] T. Roitsch, R. Ehneß, Regulation of source/sink relations by cytokinins, Plant
Growth Regul. 32 (2000) 359–367.
[27] L.E.P. Peres, et al., Grafting of tomato mutants onto potato rootstocks: an ap-
proach to study leaf-derived signaling on tuberization, Plant Sci. 169 (2005)
680–688.
[28] T. Eviatar-Ribak, et al., A cytokinin-activating enzyme promotes tuber forma-
tion in tomato, Curr. Biol. 23 (2013) 1057–1064.
[29] L. Carmel-Goren, Y.S. Liu, E. Lifschitz, D. Zamir, The SELF-PRUNING gene
family in tomato, Plant Mol. Biol. 52 (2003) 1215–1222.
[30] C. Navarro, E. Cruz-Oró, S. Prat, Conserved function of FLOWERING LO-
CUS T (FT) homologues as signals for storage organ differentiation, Curr.
Opin. Plant Biol. 23 (2015) 45–53.
[31] N.M.A. Nassar, D.Y.C. Hashimoto, S.D.C. Fernandes, Wild Manihot species:
botanical aspects, geographic distribution and economic value, Genet. Mol.
Res. 7 (2008) 16–28.
[32] S.B. Carroll, Evolution at two levels: on genes and form, PLoS Biol. 3 (2005)
e245.
[33] Y. Bai, P. Lindhout, Domestication and breeding of tomatoes: what have we
gained and what can we gain in the future?, Ann. Bot. 100 (2007) 1085–1094.
[34] V. Sarlikioti, P.H.B. De Visser, G.H. Buck-Sorlin, L.F.M. Marcelis, How plant
architecture affects light absorption and photosynthesis in tomato: towards an
ideotype for plant architecture using a functional-structural plant model, Ann.
Bot. 108 (2011) 1065–1073.
[35] E. Lifschitz, Y. Eshed, Universal florigenic signals triggered by FT homo-
logues regulate growth and flowering cycles in perennial day-neutral tomato, J.
Exp. Bot. 57 (2006) 3405–3414.
[36] A. Samach, H. Lotan, The transition to flowering in tomato, Plant Biotech-
nol. 24 (2007) 71–82.
[37] L. Pnueli, et al., The SELF-PRUNING gene of tomato regulates vegetative to
reproductive switching of sympodial meristems and is the ortholog of CEN
and TFL1, Development125 (1998) 1979–1989.
[38] M.H. Vicente, A. Zsögön, A.F.L. de Sá, R.V. Ribeiro, L.E.P. Peres, Semi-de-
terminate growth habit adjusts the vegetative-to-reproductive balance and in-
creases productivity and water-use efficiency in tomato (Solanum lycoper-
sicum), J. Plant Physiol. 177 (2015) 11–19.
[39] R.C. McGarry, B.G. Ayre, Manipulating plant architecture with members of
the CETS gene family, Plant Sci. 188–189 (2012) 71–81.
[40] Y. Hanzawa, T. Money, D. Bradley, A single amino acid converts a repressor
to an activator of flowering, Proc. Natl. Acad. Sci. U. S. A. 102 (2005)
7748–7753.
[41] C. Qin, et al., Whole-genome sequencing of cultivated and wild peppers pro-
vides insights into Capsicum domestication and specialization, Proc. Natl.
Acad. Sci. U. S. A. 111 (2014) 5135–5140.
[42] S. Jasinski, et al., PROCERA encodes a DELLA protein that mediates control
of dissected leaf form in tomato, Plant J. 56 (2008) 603–612.
[43] E. Maza, P. Frasse, P. Senin, M. Bouzayen, M. Zouine, Comparison of normal-
ization methods for differential gene expression analysis in RNA-Seq experi-
ments: a matter of relative size of studied transcriptomes, Commun. Integr.
Biol. 6 (2013) e25849.
[44] R.F. Carvalho, et al., Convergence of developmental mutants into a single
tomato model system: Micro-Tom as an effective toolkit for plant development
research, Plant Methods 7 (2011) 18.
[45] E. Carrera, O. Ruiz-Rivero, L.E.P. Peres, A. Atares, J.L. Garcia-Martinez,
Characterization of the procera tomato mutant shows novel functions of the
SlDELLA protein in the control of flower morphology, cell division and ex-
pansion, and the auxin-signaling pathway during fruit-set and development,
Plant Physiol. 160 (2012) 1581–1596.
[46] V.S. Lor, C.G. Starker, D.F. Voytas, D. Weiss, N.E. Olszewski, Targeted mu-
tagenesis of the tomato PROCERA gene using transcription activator-like ef-
fector nucleases, Plant Physiol. 166 (2014) 1288–1291.
[47] D.S. Multani, et al., Loss of an MDR transporter in compact stalks of maize
br2 and sorghum dw3 mutants, Science 302 (2003) 81–84.
[48] S.P.C. Groot, L.C.P. Keizer, W. de Ruiter, J.J.M. Dons, Seed and fruit set of
the lateral suppressor mutant of tomato, Sci. Hortic. 59 (1994) 157–162.
[49] M. Martín-Trillo, et al., Role of tomato BRANCHED1-like genes in the con-
trol of shoot branching, Plant J. 67 (2011) 701–714.
[50] L.G. Lowder, et al., A CRISPR/Cas9 toolbox for multiplexed plant genome
editing and transcriptional regulation, Plant Physiol. 169 (2015) 971–985.
[51] A. Piatek, et al., RNA-guided transcriptional regulation in planta via synthetic
dCas9-based transcription factors, Plant Biotechnol. J. 13 (2015) 578–589.
[52] N. Ficcadenti, et al., Genetic engineering of parthenocarpic fruit development
in tomato, Mol. Breed. 5 (1999) 463–470.
[53] E.A. Ainsworth, D.R. Ort, How do we improve crop production in a warming
world?, Plant Physiol. 154 (2010) 526–530.
[54] A. Mazzucato, A. Taddei, G. Soressi, The parthenocarpic fruit (pat) mutant of
tomato (Lycopersicon esculentum Mill.) sets seedless fruits and has aberrant
anther and ovule development, Development 125 (1998) 107–114.
[55] M. Fos, et al., The gene pat-2, which induces natural parthenocarpy, alters the
gibberellin content in unpollinated tomato ovaries, Plant Physiol. 122 (2000)
471–480.
[56] M. Zouine, et al., Characterization of the tomato ARF gene family uncovers a
multi-levels post-transcriptional regulation including alternative splicing, PLoS
One9 (2014) e84203.
[57] M. de Jong, M. Wolters-Arts, R. Feron, C. Mariani, W.H. Vriezen, The
Solanum lycopersicum auxin response factor 7 (SlARF7) regulates auxin sig-
naling during tomato fruit set and development, Plant J. 57 (2009) 160–170.
[58] P.J. Cookson, et al., Increases in cell elongation, plastid compartment size and
phytoene synthase activity underlie the phenotype of the high pigment-1 mu-
tant of tomato, Planta 217 (2003) 896–903.
[59] I. Kolotilin, et al., Transcriptional profiling of high pigment-2dg tomato mutant
links early fruit plastid biogenesis with its overproduction of phytonutrients,
Plant Physiol. 145 (2007) 389–401.
[60] A.L.T. Powell, et al., Uniform ripening encodes a Golden 2-like transcription
factor regulating tomato fruit chloroplast development, Science 336 (2012)
1711–1715.
[61] T. Mochizuki, D. Ishiuchi, S. Kamimura, Early seedling growth, yield compo-
nents and fruit chemical composition of high-pigment processing tomato lines,
Acta Hortic. 220 (1988) 85–92.
[62] R.J. Bino, et al., The light-hyperresponsive high pigment-2dg mutation of
tomato: alterations in the fruit metabolome, New Phytol. 166 (2005) 427–438.
[63] T. Roitsch, M.-C. González, Function and regulation of plant invertases: sweet
sensations, Trends Plant Sci. 9 (2004) 606–613.
[64] B. Martineau, K.R. Summerfelt, D.F. Adams, J.W. DeVerna, Production of
high solids tomatoes through molecular modification of levels of the plant
growth regulator cytokinin, Nat. Biotechnol. 13 (1995) 250–254.
[65] E. Fridman, F. Carrari, Y.-S. Liu, A.R. Fernie, D. Zamir, Zooming in on a
quantitative trait for tomato yield using interspecific introgressions, Sci-
ence 305 (2004) 1786–1789.
[66] M. Ashikari, et al., Cytokinin oxidase regulates rice grain production, Sci-
ence 309 (2005) 741–745.
[67] T. Kurakawa, et al., Direct control of shoot meristem activity by a cy-
tokinin-activating enzyme, Nature 445 (2007) 652–655.
[68] S. Perilli, L. Moubayidin, S. Sabatini, The molecular basis of cytokinin func-
tion, Curr. Opin. Plant Biol. 13 (2010) 21–26.
[69] I. Sestari, et al., Near-isogenic lines enhancing ascorbic acid, anthocyanin and
carotenoid content in tomato (Solanum lycopersicum L. cv Micro-Tom) as a
tool to produce nutrient-rich fruits, Sci. Hortic. 175 (2014) 111–120.

UNCORRECTED PROOF
Plant Science xxx (2016) xxx-xxx 11
[70] E. Butelli, et al., Enrichment of tomato fruit with health-promoting antho-
cyanins by expression of select transcription factors, Nat. Biotech-
nol. 26 (2008) 1301–1308.
[71] T. Tsuda, Dietary anthocyanin-rich plants: biochemical basis and recent
progress in health benefits studies, Mol. Nutr. Food Res. 56 (2012) 159–170.
[72] M.G. Willits, et al., Utilization of the genetic resources of wild species to cre-
ate a nontransgenic high flavonoid tomato, J. Agric. Food Chem. 53 (2005)
1231–1236.
[73] D. Koenig, et al., Comparative transcriptomics reveals patterns of selection in
domesticated and wild tomato, Proc. Natl. Acad. Sci. U. S. A. 110 (2013)
2655–2662.
[74] Tomato Genome Consortium, The tomato genome sequence provides insights
into fleshy fruit evolution, Nature 485 (2012) 635–641.
[75] Z.B. Lippman, Y. Semel, D. Zamir, An integrated view of quantitative trait
variation using tomato interspecific introgression lines, Curr. Opin. Genet.
Dev. 17 (2007) 545–552.
[76] L. Mao, et al., JOINTLESS is a MADS-box gene controlling tomato flower ab-
scission zone development, Nature 406 (2000) 910–913.
[77] M.A. Budiman, et al., Localization of jointless-2 gene in the centromeric re-
gion of tomato chromosome 12 based on high resolution genetic and physical
mapping, Theor. Appl. Genet. 108 (2004) 190–196.
[78] G.B. Seymour, N.H. Chapman, B.L. Chew, J.K.C. Rose, Regulation of ripen-
ing and opportunities for control in tomato and other fruits, Plant Biotechnol.
J. 11 (2013) 269–278.
[79] J. Casals, et al., Genetic basis of long shelf life and variability into Penjar
tomato, Genet. Resour. Crop Evol. 59 (2012) 219–229.
[80] M. Liu, et al., Comprehensive profiling of ethylene response factor expression
identifies ripening-associated ERF genes and their link to key regulators of
fruit ripening in tomato, Plant Physiol. 170 (2016) 1732–1744.
[81] A. Gur, D. Zamir, Unused natural variation can lift yield barriers in plant
breeding, PLoS Biol. 2 (2004) e245.
[82] C.M. Rick, Potential genetic resources in tomato species: clues from observa-
tions in native habitats, in: A.M. Srb (Ed.), Genes, Enzymes and Populations,
Springer, New York, 1973, pp. 255–269.
[83] D.W. Rush, E. Epstein, Genotypic responses to salinity: differences between
salt-sensitive and salt-tolerant genotypes of the tomato, Plant Phys-
iol. 57 (1976) 162–166.
[84] A.F. Lucatti, et al., Differences in insect resistance between tomato species en-
demic to the Galapagos Islands, BMC Evol. Biol. 13 (2013) 175.
[85] G. Ronen, L. Carmel-Goren, D. Zamir, J. Hirschberg, An alternative pathway
to beta −carotene formation in plant chromoplasts discovered by map-based
cloning of beta and old-gold color mutations in tomato, Proc. Natl. Acad. Sci.
U. S. A. 97 (2000) 11102–11107.
[86] A. Frary, et al., Fw2.2: a quantitative trait locus key to the evolution of tomato
fruit size, Science 289 (2000) 85–88.
[87] S. Muños, et al., Increase in tomato locule number is controlled by two sin-
gle-nucleotide polymorphisms located near WUSCHEL, Plant Phys-
iol. 156 (2011) 2244–2254.
[88] C. Xu, et al., A cascade of arabinosyltransferases controls shoot meristem size
in tomato, Nat. Genet. 47 (2015) 784–792.
[89] Z. Lippman, S.D. Tanksley, Dissecting the genetic pathway to extreme fruit
size in tomato using a cross between the small-fruited wild species Lycopersi-
con pimpinellifolium and L. esculentum var. Giant Heirloom, Genet-
ics 158 (2001) 413–422.
[90] J. Liu, J. Van Eck, B. Cong, S.D. Tanksley, A new class of regulatory genes
underlying the cause of pear-shaped tomato fruit, Proc. Natl. Acad. Sci. U. S.
A. 99 (2002) 13302–13306.
[91] Z.B. Lippman, et al., The making of a compound inflorescence in tomato and
related nightshades, PLoS Biol. 6 (2008) e288.
[92] C. Brooks, V. Nekrasov, Z. Lippman, J. Van Eck, Efficient gene editing in
tomato in the first generation using the CRISPR/Cas9 system, Plant Phys-
iol. 166 (2014) 1292–1297.
[93] M.M. Mahfouz, et al., Targeted transcriptional repression using a chimeric
TALE-SRDX repressor protein, Plant Mol. Biol. 78 (2012) 311–321.
[94] L.S. Qi, et al., Repurposing CRISPR as an RNA-guided platform for se-
quence-specific control of gene expression, Cell 152 (2013) 1173–1183.
[95] T. Čermák, N.J. Baltes, R. Čegan, Y. Zhang, D.F. Voytas, High-frequency,
precise modification of the tomato genome, Genome Biol. 16 (2015) 232.
[96] A.C. Komor, et al., Programmable editing of a target base in genomic DNA
without double-stranded DNA cleavage, Nature 533 (2016) 420–424.
[97] W. Liu, et al., Synthetic TAL effectors for targeted enhancement of transgene
expression in plants, Plant Biotechnol. J. 43 (2013) 436–446.
[98] A.I. Fernandez, et al., Flexible tools for gene expression and silencing in
tomato, Plant Physiol. 151 (2009) 1729–1740.
[99] E. Shani, et al., Stage-specific regulation of Solanum lycopersicum leaf matu-
ration by class 1 KNOTTED1-LIKE HOMEOBOX proteins, Plant
Cell 21 (2009) 3078–3092.
[100] H.-L. Xing, et al., CRISPR/Cas9 toolkit for multiplex genome editing in
plants, BMC Plant Biol. 14 (2014) 327.
[101] S. Kiani, et al., Cas9 gRNA engineering for genome editing, activation and re-
pression, Nat. Methods 12 (2015) 1051–1054.
[102] J.G. Zalatan, et al., Engineering complex synthetic transcriptional programs
with CRISPR RNA scaffolds, Cell 9 (2014) 1–12.
[103] S. Lombardi-Crestana, et al., The tomato (Solanum lycopersicum cv. Mi-
cro-Tom) natural genetic variation Rg1 and the DELLA mutant procera control
the competence necessary to form adventitious roots and shoots, J. Exp.
Bot. 63 (2012) 5689–5703.
[104] L.E. Pino, et al., The Rg1 allele as a valuable tool for genetic transformation of
the tomato Micro-Tom model system, Plant Methods 6 (2010) 23.
[105] F.N. Arikita, et al., Novel natural genetic variation controlling the competence
to form adventitious roots and shoots from the tomato wild relative Solanum
pennellii, Plant Sci. 199–200 (2013) 121–130.
[106] F.A. Jesus, A. Zsögön, L.E.P. Peres, Physionomics, in: A. Borém, R.
Fritsche-Neto (Eds.), Omics in Plant Breeding, John Wiley & Sons, Inc.,
Chichester, UK, 2014, pp. 103–125.
[107] J.A. Birchler, A.F. Johnson, R.A. Veitia, Kinetics genetics: incorporating the
concept of genomic balance into an understanding of quantitative traits, Plant
Sci. 245 (2016) 128–134.
[108] S.J. Park, et al., Optimization of crop productivity in tomato using induced mu-
tations in the florigen pathway, Nat. Genet. 46 (2014) 1337–1342.
[109] D. Hoisington, et al., Plant genetic resources: what can they contribute toward
increased crop productivity?, Proc. Natl. Acad. Sci. U. S. A. 96 (1999)
5937–5943.
[110] P.L. Sanchez, R.A. Wing, D.S. Brar, The wild relatives of rice: genomes and
genomics, in: Q. Zhang, R.A. Wing (Eds.), Genetics and Genomics of Rice,
Springer, New York, 2013, pp. 9–25.
[111] R. Machida-Hirano, et al., Diversity of potato genetic resources, Breed.
Sci. 65 (2015) 26–40.
[112] S.J. Prince, et al., Genetic variants in root architecture-related genes in a
Glycine soja accession, a potential resource to improve cultivated soybean,
BMC Genomics 16 (2015) 132.
[113] K. Schumacher, M. Ganal, K. Theres, Genetic and physical mapping of the lat-
eral suppressor (ls) locus in tomato, Mol. Gen. Genet. 246 (1995) 761–766.
[114] E. Lifschitz, et al., The tomato FT ortholog triggers systemic signals that regu-
late growth and flowering and substitute for diverse environmental stimuli,
Proc. Natl. Acad. Sci. U. S. A. 103 (2006) 6398–6403.
[115] J. Royo, Y. Kowyama, A.E. Clarke, Cloning and nucleotide sequence of two
S-RNases from Lycopersicon peruvianum (L.) Mill, Plant Physiol. 105 (1994)
751–752.
[116] K.-Y. Chen, B. Cong, R. Wing, J. Vrebalov, S.D. Tanksley, Changes in regula-
tion of a transcription factor lead to autogamy in cultivated tomatoes, Sci-
ence 318 (2007) 643–645.