The Development of Forest Genetic Breeding and the Application of Genome Selection and CRISPR/Cas9 in Forest Breeding

Abstract

With the birth of classical genetics, forest genetic breeding has laid a foundation in the formation of the basic theories of population genetics, quantitative genetics, cytogenetics, and molecular genetics. Driven by the rapid growth of social demand for wood and other forest products, modern genetics, biotechnology, biostatistics, crop and animal husbandry breeding theories, and technical achievements have been continuously introduced for innovation, thus forming a close combination of genetic basic research and breeding practice. Forest tree breeding research in the world has a history of more than 200 years. By the middle of the 20th century, the forest tree genetic breeding system was gradually formed. After entering the 21st century, the in-depth development stage of molecular design breeding was opened. With the continuous improvement of traditional genetic breeding methods, emerging modern bioengineering technology has also continuously promoted the development of forest genetic breeding. This study mainly summarizes the research history of forest tree genetics and breeding, as well as discusses the application of modern bioengineering technology represented by genome selection and gene editing in forest tree breeding, so as to provide better reference for forest tree breeding research.

1. Introduction

Based on genetics, forest genetic breeding is to study the breeding and cultivation of new forest varieties, and to explore the theory and method of forest reproduction. Forest genetic research in the world has a history of more than 200 years [1,2,3]. It has experienced three stages: birth and foundation stage, formation stage, and further development stage (Figure 1). At present, it has gradually formed a mature system based on traditional breeding methods and supplemented by modern molecular biology techniques. Since the birth of forest genetic breeding, many excellent varieties have been cultivated through traditional breeding methods, which has become an important part of forest breeding history. However, traditional breeding methods have the disadvantages of long cycle, poor predictability, and low efficiency. The contradiction between the demand for rapid directional breeding of new varieties and the bottleneck effect of traditional breeding was increasingly apparent. With the development of biological technology, new breeding theories and methods are constantly emerging. Modern forest breeding technology represented by molecular breeding has greatly promoted the process of directional breeding of new forest varieties. Through modern forest tree breeding technology systems such as genome selection and gene editing, the process of forest tree genetic improvement can be accelerated, efficient forest tree breeding can be realized, and the innovation ability of forest tree seed industry can be improved.

2. Birth and Foundation of Forest Genetic Breeding

Although the practice of forest genetic breeding has a long history, modern forest genetic breeding research began to appear in the 19th century. In 1821, De Vilmorin studied the differences in growth, stem form, needles, buds and cones of different provenances of Pinus sylvestris in Paris, which became the beginning of modern forest genetic breeding [1]. In 1892, the International Union of Forest Research Organizations (IUFRO), established in Germany, assigned forest genetic breeding to the Department of Physiology and Genetics, and formulated the international provenance test plan for major tree species at the first meeting, leading the development of forest genetic breeding.

After entering the 20th century, with the sharp increase in demand for wood, forest resources cultivation and utilization of some European industrialized countries began to pay attention to provenance test. In 1907, IUFRO organized the first international provenance trial of Pinus sylvestris in Germany, Sweden, Austria, Switzerland, and other countries, and demonstrated that there were geographical provenance variations in tree growth, morphology, and adaptability, and that trait variations were continuous and heritable; the concepts of continuous gradient variation and gradient group were further proposed [2,3,4,5]. Since then, IUFRO organized the second international provenance trial project including Pinus sylvestris and Picea abies in 1938, and further incorporated Larix decidua into the provenance trial plan in 1944 [5]. Several provenance experiments have prompted the combination of genetic theory research and forest tree breeding practice. After decades of efforts, the main afforestation tree species in foreign countries, including those with small natural distribution areas, have been subjected to provenance tests.

The construction of seed orchard also gradually began. In 1787, German scientist F.A.L. von Burgsdorf proposed to construct seed orchards through asexual reproduction. To increase the quinine content of cinchona trees, the Dutch first established the Cinchona calisaya seed orchard in Java in 1880 [6]. Since then, Sweden, Finland, Denmark, Norway, and other Nordic countries have also built a number of Scots pine, Norway spruce, Larix decidua, and other seed orchards to promote the development of local plantations [7,8].

Hybridization was an active field of plant breeding in the 19th century. German scientist Klotzch first tested the hybridization between Pinus sylvestris and Pinus nigra Arn. in 1845. In 1912, Henry carried out interspecific hybridization of poplar (Populus spp.), oak (Quercus spp.), and other tree species, and selected fast-growing and adaptable poplar (P.generosa), representing the rise of hybrid breeding. Since then, the United States, the Soviet Union, Italy, Germany, Japan, and other countries have carried out large-scale hybrid experiments of Populus, Quercus, Salix, Betula, Larix, Robinia pseudoacacia, and Cryptomeria japonica [9,10,11,12]. Taking poplar hybridization as an example, in 1929, Jacometti used P. canadensis and P. deltoides angulata ‘Carolin’ to breed ‘I-214′ poplar (P. × canadensis ‘I-214′), which has been widely introduced around the world for its fast growth and strong adaptability [13]. In 1945, the hybridization experiment between P. hopeiensis and P. tomentosa started the journey of forest tree hybridization in China [14]. The discovery of triploid black poplar is the starting point of forest tree polyploid breeding in 1935 [15].

3. Forming Stage of Forest Genetic Breeding

In the 1950s, due to the sharp increase in wood consumption, forest land area was shrinking, and problems such as increasing the unit area of wood production became urgent, which greatly promoted the development of forest genetic breeding. In addition, the gradual development of biological technology, molecular markers, genetic engineering, and other biological techniques began to be widely used in forest genetic breeding. In order to meet the demand for wood, Australia, New Zealand, and other countries began to introduce pine trees. Since then, many countries have performed large-scale introduction, Eucalyptus spp., Populus nigra, Populus deltoides, Pinus radiata, Pinus elliottii, and other tree species, which have gradually become international afforestation tree species [16,17,18,19].

Seed orchards have been developed since the 1950s and became popular around the world in the 1960s. Nearly 90 seed orchards have been established in about 50 countries on five continents, involving Pinus radiata, Picea abies, Rhus typhina L., Cunninghamia lanceolata, and other tree species, but most of them were coniferous species; thus, countries around the world began to change from the main way of afforestation from relying on seed reproduction to seed orchards [20,21]. For example, in the United States, through the use of preferred genetic testing seed orchard technology, by 1991, a total of about 4000 hm² seed orchards had been built, gradually developing from primary seed orchards, thinning seed orchards, 1.5th- and second-generation seed orchards to third-generation seed orchards [22]. A large number of theoretical and technical achievements of seed orchards were reported in terms of breeding strategy, clone arrangement, pollen dispersal law, grafting technique, tree management, flowering and fruiting promotion, indirect selection, and pest control [23,24,25]. Forest breeding and seed orchard construction also began to shift to an enterprise-led approach [18]. China began to build Chinese fir clone seed orchards in 1964, and then began to receive the attention of the forestry production department; accordingly, seed orchard construction has rapidly developed. By the end of 1997, the area of seed orchards in mainland China reached 15,420 hm², with the construction of more than 40 kinds of tree species, but most of the primary seed orchards were established using 1.5th-generation seed orchard, a few species were built using second-generation seed orchard, and individual tree species were build using third-generation seed orchard [6].

Hybrid breeding was still popular at this stage, and a variety of heterosis tree species were successfully cultivated, among which Populus, Eucalyptus, Pinus, and Larix achieved remarkable results. E. grandis × E.urophylla, derived from Aracruz, Brazil, grows faster, germinates more vigorously, and is more resistant to canker disease [6]. E. grandis × E. camaldulensis, bred in South Africa, shows a significantly higher growth rate, drought resistance, and papermaking ability than the hybrid’s parents [26]. Since the 1950s, Germany, Belgium, Greece, Asia, Japan, and South Korea have implemented a large number of pine intergeneric hybridization studies, leading to the cultivation of P. rigida Mill × P. taeda L., P. taeda L. × P. banksiana Lamb., P. sylvestris L. × P. nigra Arn., and P. thunbergii Parl. × P. massoniana Lamb. The hybrid of L. decidua Mill. × L. kaempferi Carr. cultivated in Germany, has higher growth than its parents [6]. During this period, China performed extensive hybridization studies and cultivated a number of hybrid tree species. By the 1970s, China had cultivated a number of fast-growing, high-quality, disease-resistant poplar hybrid varieties, with the clones of P. euramericana CL. ‘Zhonglin46′ and P. euramericana CL. ‘Zhonglin Sanbei 1′ being widely used throughout the country [27]. The hybridization of Eucalyptus began in the 1960s. Although it developed slowly in the early stage, cross-breeding was gradually carried out on the basis of introduction and provenance experiments, with E. grandis and E. urophylla showing the best results [28,29,30]. The cross-breeding of Pinus also began in the 1960s, and the hybrid varieties obtained had higher economic benefits [6].

In the 1950s–1970s, forest ploidy breeding was widely investigated, and a series of haploid, triploid, and tetraploid varieties were obtained in Populus, Betula, Robinia pseudoacacia, Morus, and Hevea brasiliensis [31,32,33,34,35]. However, due to the limitations of theory and technical methods, there has been no significant breakthrough in ploidy breeding, with only polyploid varieties such as Morus and Robinia pseudoacacia and hybrid triploid varieties such as P. tremula L. × P. tremuloides being applied in production. In the 1990s, quite a few triploid P. tomentosa varieties with fast growth, long fibers, and low lignin content were successfully selected by backcrossing the natural 2n pollen of P. tomentosa with P. tomentosa × P. bolleana, which caused an upsurge in polyploid breeding [36,37,38]. Haploidy by anther culture was first achieved in Datura stramonium in 1964 [39]. Owing to the long reproductive cycle of forest trees, the physiological status of donor plants was greatly affected by the temperature and precipitation of the year, which in turn affected anther culture, resulting in poor repeatability of the experiment, seriously limiting the application and development of anther culture in forest trees. The study of forest haploid breeding has decreased rapidly since the 1990s, remaining in a semi-stagnant state for a long time.

Since the 1960s, the demands of wood pulp papermaking and other forest industries have surged; thus, countries began to focus on clone breeding. With the progress of modern biology, plant growth regulators and all-optical spray seedling technology have been widely used, such that the rooting rate of cutting seedlings has been significantly improved, with even coniferous tree species such as P. abies and P. radiata D. Don also achieving large-scale cutting propagation [40]. Tissue culture technology has been applied to forest tree asexual reproduction, and more than 200 kinds of trees from more than 30 families have obtained complete plants through tissue or organ culture. Among them, excellent varieties of Populus and Eucalyptus have achieved large-scale asexual reproduction by tissue culture [41]. Somatic embryo induction technology has also made a breakthrough in conifers, becoming an effective approach for the genetic improvement of tree species which can only reproduce via seeds [42]. P. radiata D. Don, P. abies (L.) Karst., P. taeda L., Pseudotsuga menziesii (Mirb.) Franco, and other tree species were introduced to large-scale afforestation applications through somatic embryo reproduction [43].

4. In-Depth Development of Forest Genetic Breeding

With the development of genome sequencing technology, forest genetic breeding research has entered the post-genome era. Since the first publication of the P. trichocarpa genome in 2006 [44], nearly 50 forest genomes have been published, namely, Norway spruce [45], Phyllostachys heterocycla [46], E. grandis [47], B. pendula [48], Eucommia ulmoides [49], Liriodendron [50], Chinese pine [51], and others [52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67]. Organelle genome sequencing represented by the mitochondrial genome and chloroplast genome has also developed rapidly and is widely used in forest trees [68,69,70,71,72,73,74,75,76,77,78,79,80,81,82]. With the development of single-cell sequencing technology, this technology has been successfully applied to reveal the intracellular dynamics of single cells in forest trees [83,84,85,86,87,88]. With the reduction in cost and the continuous improvement of technology, single-cell sequencing still has broad application prospects in forest trees. On this basis, combined with the results of transcriptomics, metabolomics, proteomics, and degradation, massive function genes regulating important traits have been mapped and cloned, and the genetic regulation mechanisms of some important traits such as growth, development, and stress resistance have been analyzed from different levels of transcriptional regulation and post-transcriptional regulation [89,90,91,92,93,94,95,96,97]. In addition, quantitative trait locus (QTL) mapping [98,99,100,101,102,103,104], genome-wide association study (GWAS) [105,106,107,108,109,110,111,112], marker-assisted selection (MAS) [113,114,115,116,117,118], genomic selection or genome-wide selection (GS), and other techniques have been widely used to analyze the difference of forest traits [119,120].

Transgenic studies became a research trend at this stage; however, limited by the establishment of the transgenic system, only a few species such as Populus, Pinus, and Eucalyptus were successfully implemented to obtain a transgenic forest [121,122,123,124,125,126,127]. Genetic engineering is controversial because of its potential harm to the environment, and the long-term stability of transgenic trees remains to be studied. However, no effect of exogenous genes on the biodiversity of arthropods or soil bacteria has been found in field safety tests on transgenic trees [128,129,130,131,132,133]. China is the only country that has commercialized genetically modified poplar species (Populus nigra lines with Bt and Populus × aldatomentosa Cl.741 with BtCry1Ac and API), causing widespread concern in the world [134]. Strict transgenic plant safety regulatory policies and measures temporarily limit the application of transgenic varieties. Transgenic forests are still in the growth and development stage, with further research on pest resistance and stress resistance mechanisms required before actual production. The development of gene editing technology has solved people’s concerns about the safety of transgenic plants, and the CRISPR/Cas9 system has shown strong fixed-point editing capabilities in the field of gene editing [135,136]. In 2015, researchers from the University of Georgia used Cas9 to successfully knock out lignin synthesis-related enzyme genes in P. tomentosa, the first successful application of the CRISPR/Cas system to edit offspring in woody plants [137].

At present, ploidy breeding is also valued, with triploid breeding results being particularly prominent. P. abies, Thuja plicata, Cryptomeria japonica, Encephalartos hildebrandtii, Ginkgo biloba, Morus alba L., Toxicodendron vernicifluum (Stokes) F. A. Barkl., Acacia dealbata, Salix, P. deltoides, P. alba, P. balsamifera, P. tomentosa, and other natural polyploids have been detected [37,138,139,140,141,142,143,144,145]. Triploids of Populus, Salix, Betula, Acacia auriculiformis, Morus, and others have also been obtained by hybridization, from which new varieties were selected for production [37,146,147,148]. In addition, triploids obtained by chromosome doubling of male and female gametes were successfully reported in P. tremula L., P. tremuloides, P. deltoides Marshall., P. balsamifera, P. alba L., P. tomentosa × P. bolleana, P. alba × P. glandulosa, P. tomentosa, P. adenopoda, P. canescens, Morus alba, Eucommia ulmoides, and Eucalyptus [33,149,150,151,152,153,154,155,156,157,158]. Triploid angiosperms tend to have fast-growing characteristics, but most of the triploid gymnosperms grow slowly and have poor fertility. In addition, forest triploids also have the genetic characteristics of huge cells and organs, significantly increased secondary metabolites, and enhanced resistance [141].

5. Application of Genome Selection in Forest Genetic Breeding

Genome-wide selection (GS) is a new method for breeding selection using high-density molecular markers, which was first proposed by Meuwissen et al. [159,160]. By measuring a large number of molecular markers across the whole genome, GS can estimate the effect values of different chromosome fragments or single markers, and then accumulate the chromosome fragments or all marker effect values in the whole genome to obtain the genomic estimated breeding value (GEBV). The individual breeding value can be estimated by combining phenotypic and marker information, and early individual prediction and selection can be made according to the breeding value to shorten the generation interval, accelerate the breeding process, improve the selection accuracy, and save costs [161]. After constructing two populations, i.e., the training population (TP) and candidate population/testing population, the effect value of each marker can be estimated on the basis of the phenotype and genotype data of the TP, and a prediction model of genotype and phenotype can be established. Then, the GEBV can be estimated using the known genotype data and the SNP effect estimates of individuals in the candidate population. Finally, the reserved individuals can be selected from the candidate groups according to the GEBV ranking. The theoretical hypothesis is that, in the high-density SNP markers distributed throughout the genome, at least one SNP can be in linkage disequilibrium (LD) with the quantitative trait locus (QTL) affecting the target trait, such that the effect of each QTL can be reflected by an SNP [162]. At present, the commonly used GS models include rrBLUP [163], synbreed [164], GBLUP [165], BGLR [166], GVCBLUP [167], GAPIT [168], sommer [169], BLUPGA [170], and Bayes Alphabet [171]. It is necessary to design breeding programs according to the actual needs and construct corresponding GS models for specific groups and target traits.

The research on forest tree genome selection has mainly been based on perennial cross-pollinated tree species with a long growth cycle and high heterozygosity. At present, genome selection research has been carried out in Picea, Pinus, Populus, Hevea, and Eucalyptus (

Table 1. Application of GS in woody plants.

Genus	Species	Reference
Elaeis	E. guineensis	[175,176]
	E. oleifera	[177]
Eucalyptus	E. grandis × E. urophylla	[178,179,180,181]
	E. pellita F. Muell.	[182,183]
	E. robusta Sm.	[184]
	E. benthamii Maiden & Cambage.	[182]
	E. nitens (H.Deane & Maiden) Maiden.	[185,186]
	E. urophylla × E. grandis	[178]
	E. grandis	[187,188]
	E. globulus Labill.	[189,190]
	E. dunni Maiden	[191]
	E. cladocalyx F. Muell.	[192]
	E. camaldulensis	[193]
	E. grandis	[47]
	E. polybractea	[170]
Hevea	H. brasiliensis	[55,194,195,196,197]
Picea	P. abies	[198]
	P. glauca	[199,200,201,202]
	P. sitchensis	[203]
	P. mariana (Mill.) Britton.	[198,204,205]
Pinus	P. pinaster	[206]
	P. taeda	[207,208,209]
	P. contorta Douglas ex Loudon	[210]
	P. radiata D. Don	[186,211]
	P. sylvestris Thunb.	[212]
Populus	P. deltoides W. Bartram ex Marshall	[213]
	P. trichocarpa	[44]
	P. euphratica	[214]
	P. nigra L	[215]
Castanea	C. dentata	[216]
Pseudotsuga	P. menziesii (Mirb.) Franco	[217]

). The application of genome selection in forest trees is helpful to shorten the breeding cycle, improve the selection efficiency, and accelerate the genetic improvement process. Wong and Bernardo [172] systematically compared phenotypic selection, MAS, and GS through Elaeis guineensis simulation data and found that GS had better selectivity and could still obtain certain genetic gain in breeding populations with a long generation interval and small sample size. In addition, genome selection has the advantages of low cost, high accuracy, high selection efficiency, and more genetic variation information in forest tree breeding [173,174].

Although forest genome selection has achieved certain results, compared with crops, development has been slow with low accuracy, limited by many factors. Reference genome quality is a prerequisite for GS to be applied to the genetic improvement of target species. However, the genome quality of forest tree species is generally lower than that of crops, which limits the application of GS in forest tree genetic improvement. Moreover, forest trees, especially coniferous trees, have larger genomes, leading to a significant increase in GS costs [218]. The prediction model can affect the accuracy of genome prediction; however, currently developed statistical models and analytical tools struggle to meet the application requirements of forest tree genome selection. Forest tree species generally have the characteristics of a long generation time, long juvenile phase, and giant plant size, which makes it difficult to apply forest tree genome selection in multigeneration families for breeding [219]. The genetic relationship between the reference population and the candidate population used to construct the genomic selection model also affects the accuracy of the genomic breeding estimates to some extent. A closer genetic relationship between two populations results in more accurate genomic breeding estimates [220,221,222,223,224]. Trees have complex target traits and heritability, and the heritability of growth traits and wood traits changes with the increase of tree age, resulting in low prediction accuracy [225,226,227]. In addition, factors such as marker density, population structure, trait heritability, population size and relationship, and the number and distribution of target trait loci also affect the accuracy of tree genome selection [228]. These limitations of GS also affect its application in forest tree breeding. GS mainly considers additive effects, whereas dominant effects and interaction effects are not included in the breeding value estimation model. GS is currently mainly carried out in forest varieties, while, for forest hybrid varieties, GS prediction accuracy is reduced. The prediction effect of individual breeding value with a distant genetic relationship among forest tree varieties is not ideal. The cost of obtaining GS genome information is high, but the results of multi-omics research have not been fully utilized. Compared with the sparse matrix of traditional BLUP, GS needs to use the dense matrix calculated using genomic information to estimate the parameters of the mixed model and calculate the model, which is more complicated.

At present, GS technology has been widely used in the genetic improvement of crops and other plants. Although some achievements have been made in forest breeding, it still has great potential. With the continuous development of genome sequencing technology and the improvement of analysis platforms, the background data of genomics, phenomics, and genetics of forest trees are steadily improving, thereby providing strong technical support and data support for GS research of forest tree species. The gradually improved GS technology can also greatly promote the development of forest breeding. However, selective breeding of forest tree genomes still faces some challenges. Firstly, genome quality is the basis for GS research. However, the quality of genome assembly of forest tree species is generally low; thus, improving the accuracy and quality of the reference genome remains an important condition for high-quality forest GS research. Secondly, choosing a reasonable statistical model is necessary for GS research. The selection of statistical models should fully consider the characteristics of target tree species and target traits. The computational efficiency and accuracy of statistical models ensure the accuracy of GS research. The perennial attribute makes the target trait data of forest tree species longitudinal, and GS statistical models and analysis software with the ability to process longitudinal trait data are still needed. Thirdly, multi-trait composite selection is needed forest GS research. At present, the study of forest GS focuses on the estimation of the genomic breeding value of a single trait, but the cultivation of new varieties with multiple dominant traits is a necessity for forest genetic improvement. Fourthly, the emergence of new technologies has been the driving force of forest GS research. CRISPR/Cas9 technology can be used not only for genome editing, but also for identifying precise locations in the genome to induce mitotic recombination. The combination of GS and genome editing (GE) can accelerate forest tree breeding. With the continuous development of pan-genome genotyping and interpolation techniques, the development of structural variations or special markers of subpopulations can promote the wide application of GS in forest tree breeding.

6. Application of CRISPR/Cas9 Technology in Forest Tree Breeding

The CRISPR/Cas9 system is currently recognized as the most promising gene editing technology. The Cas9 protein binds to the artificially designed sgRNA to form an sgRNA–Cas9 protein complex and binds to a specific nucleotide sequence under its guidance. After sgRNA binds to Cas9 protein and targets a specific genomic site, Cas9 is cleaved to produce double-strand breaks (DSBs), which are repaired by cell-autonomous nonhomologous end joining (NHEJ) or homology-directed repair (HR), resulting in base insertion or deletion at the target site, thus inactivating gene function [136]. Compared with zinc finger nucleases (ZFNs) and transcription activator-like effector nucleases (TALENs), the CRISPR/Cas9 system has a simple design process, convenient operation, and high gene editing efficiency, and it has played an important role in the precision breeding of major food and economic crops. The CRISPR/Cas9 system has been applied in a variety of plants following success in gene targeting experiments in Arabidopsis thaliana, rice (Oryza sativa L.) and other plants. CRISPR/Cas9 system has good application prospects in improving crop yield, improving crop quality, and cultivating new varieties with disease resistance and stress resistance [229,230,231,232]. Forest trees have the biological characteristics of a long growth cycle, high genetic heterozygosity, and complex genome ploidy. Immature genetic transformation systems also lead to a limitation of CRISPR/Cas9 in forest tree genetic improvement. At present, CRISPR/Cas9 has been successfully applied to only a few forest species (

Table 2. Application of CRISPR/Cas system in woody plants.

Genus	Species	Genes	Phenotype	Reference
Populus	P. canescens P. tremula	SOC1, FUL, NFP-like genes, TOZ19		[245]
	P. alba × P. glandulosa	PdNF-YB21	Drought resistance	[246]
	P. alba × P. glandulosa	PagDA1	Promoting xylem formation	[247]
	P. alba var. pyramidalis	Hyg		[248]
	P. tomentosa	PtSGT1, PtSGT4	Regulating cellulose synthesis in cell wall	[249]
	P. tomentosa Carr. clone 741	PtoDET2	Xylem development and reduced wall thickness	[250]
	P. tomentosa Carr. clone 741	MYB115	Reduced proanthocyanidin accumulation	[251]
	P. tomentosa Carr. clone 741	PtoMYB156	Negative regulation of secondary wall formation	[252]
	P. tomentosa Carr. clone 741	PtoMYB170	Regulates lignin deposition	[253]
	P. tomentosa Carr. clone 741	PtrMYB57	Increased anthocyanins and procyanidins	[254]
	P. tomentosa Carr. clone 741	JMJ25	Increased anthocyanin accumulation	[255]
	P. tomentosa Carr. clone 741	MYB189	Regulating secondary cell-wall biosynthesis	[256]
	P. tomentosa Carr. clone 741	PtoDWF4	Reduced xylem development	[257]
	P. tomentosa Carr. clone 741	PtrWRKY18, PtrWRKY35	Melampsora resistance	[258]
	P. tremula × P. alba clone INRA 717-IB4	GNC	Drought stress tolerance	[259]
	P. tremula × P. alba clone INRA 717-IB4	4CL1, 4CL2	Decreased lignin content, discoloration of stems	[237]
	P. tremula × P. alba clone INRA 717-IB4	LEAFY		[260]
	P. tremula × P. alba clone INRA 717-IB4	BRANCHED1, BRANCHED2	Bud outgrowth control	[261]
	P. tremula × P. alba clone INRA 717-IB4	LHY2	Photoperiodic growth	[262]
	P. tremula × P. alba clone INRA 717-IB4	PeuBELL15	Improved accumulation of glucan and lignin	[263]
	P. tremula × P. alba clone INRA 717-IB4	SHR	Affecting endoderm single-cell layer	[264]
	P. tremula × P. alba clone INRA 717-IB4	MYB186, MYB138, MYB38	Non-glandular trichomes	[265]
	P. tremula× P. tremuloides clone T89	BRC1	Photoperiodic control of seasonal growth	[266]
	P. alba	PalCESA4	Affecting cellulose content	[267]
	P. alba var. pyramidalis	PalWRKY77	Salt resistance	[268]
	P. alba x P. glandulosa	CSE	Increased lignocellulose biomass	[269]
	P. alba x P. glandulosa	PagPDS1	Chlorophyll biosynthesis	[270]
	P. tomentosa	GATA19	Photosynthesis and growth	[271]
	P. tomentosa	PtoLAC14	Integrated enhancement on biomass enzymatic saccharification	[272]
	P. tomentosa Carr. clone 741	PtoPDS	Chlorophyll biosynthesis	[238]
	P. tremula	ARR17	Regulating gender	[273]
	P. tremula × P. alba	CSE1, CSE2	Reduced lignin and increased cellulose	[274]
	P. tremula L. × P. tremuloides Michx.	VNS	Secondary cell-wall thinning	[275]
	P. tremula × P. alba	PopSAP	Impaired growth, complete sterility with no initiation of inflorescences	[276]
	P. tremula × P. alba	PtaSUT4	Orchestration of ROS, antioxidant, starch utilization, and RWC dynamics	[277]
	P.tremula × P. tremuloides	FT1, FT2	Yearly growth cycle	[278]
	P.tremula × P. tremuloides	PGM	Starch biosynthesis	[279]
	P. trichocarpa	PtrHSFB3-1, PtrMYB092	Reduced lignin and increased cellulose	[280]
	P. trichocarpa	PtrADA2b-3	Drought resistance	[281]
	P. trichocarpa	PtrMYB161	Wood Formation	[282]
	P. trichocarpa	PtrMYB074, PtrWRKY19	Strong drought-tolerant	[283]
	P. trichocarpa	PtrCesA4, PtrCes7A/B or 8A/B	Cellulose biosynthesis	[284]
	P. trichocarpa L.	PHBMT1	Controlling the formation of p-hydroxybenzoylated lignin structures	[285]
Citrus	C. sinensis cv. Valencia	CsPDS	Chlorophyll biosynthesis	[233]
	C. sinensis Osbeck	CsLOB1	Canker resistance	[234]
	C. paradisi	CsLOB1	Canker resistance	[236,286]
Poncirus	P. trifoliate L. × C. sinensis L. Osb	PDS	Chlorophyll biosynthesis	[235]
	P. trifoliate L. × C. sinensis L. Osb	Cs7g03360	Leaf development	[287]
Manihot	M. esculenta	MePDS	Chlorophyll biosynthesis	[239]
	M. esculenta cv. 60444	nCBP-1, nCBP-2	Biotic stress response	[288]
	M. esculenta	AC2, AC3	Biotic stress response	[289]
Hevea	H. brasiliensis	FT, TFL1	Early flowering	[243]
Jatropha	J. curcas	JcCYP735A	Growth and flowering regulation	[241]
Parasponia	P. andersonii	PanHK4, PanEIN2, PanNSP1, PanNSP2	Nodule formation, layer activity, plant sex	[290]
Bambusa	B. oldhamii	PDS	Chlorophyll biosynthesis	[242]
Dendrocalamus	D. latiflorus Munro	DlmPSY1-A, DlmPSY1-B, DlmPSY1-C, GRG1	Whitening, increased plant height	[244]
Juglans	J. regia L.	JrPDS	Chlorophyll biosynthesis	[291]
Picea	P. glauca	DXS1	Albino somatic embryo (SE) plants	[292]
Cryptomeria	C. japonica	CjChlI	Whitening	[293]
Castanea	C. sativa Mill	PDS	Chlorophyll biosynthesis	[294]
Eucalyptus	Eucalyptus	FT	No statistical difference in seedling vegetative growth rate or leaf morphology	[295]

), such as Populus, Parasponia andersonii, Jatropha curcas, Hevea brasiliensis, and Citrus. Citrus sinensis (L.) Osbeck was the first woody plant to undergo gene editing using CRISPR/Cas9. Jia and Wang used Xanthomonas citri subsp.-promoted Agroinfiltration for the CsPDS in Citrus sinensis (L.) Osbeck and confirmed the feasibility of the system in Citrus sinensis (L.) Osbeck for the first time in detached leaves, but they did not obtain complete gene-edited plants [233]. Subsequently, several research teams have been committed to improving transformation methods and gene editing efficiency, and gene editing lines capable of stable inheritance have been obtained in C. sinensis Osbeck [234], Poncirus trifoliate L. Raf. × C. sinensis L. Osb. [235] and C. paradisi [236]. Poplar is not only an important timber species and model tree species for genetic research of woody plants, but also the most successful tree species for CRISPR/Cas9 application. In 2015, Tsai’s team from the University of Georgia in the United States successfully edited the lignin synthesis-related enzyme gene 4-coumarate coenzyme A ligase genes 4CL1 and 4CL2 in the hybrid P. tremula × P. alba, opening the door for poplar gene editing [237]. In the same year, a team led by Keming Luo from Southwest University of China successfully knocked out the phytoene desaturase gene (PDS) in P. tomentosa Carr. to obtain albino poplar material [238]. Since then, CRISPR/Cas9 has been widely used in the genetic improvement of poplar and has been widely used in the analysis of poplar growth and development, abiotic stress response, and biotic stress response (Table 2). Cassava (Manihot esculenta) is one of the important edible plants in the tropics and one of the main raw materials for industrial flour. Odipio et al. [239] first established the CRISPR/Cas9 gene editing system in cassava for the MePDS gene and obtained independent transgenic plants. Since then, CRISPR/Cas9 has also been used to breed cassava varieties resistant to Cassava brown streak virus (CBSV) and African cassava mosaic virus (ACMV) [240]. Jatropha (Jatropha curcas) seeds can be refined into biodiesel, which is a new source of biomass energy. JcCYP735, a rate-limiting enzyme gene for cytokinin synthesis in Jatropha, was edited by CRISPR/Cas, and the mutant showed plant dwarfing [241]. Hevea brasiliensis, Parasponia andersonii, and Dendrocalamus have also been used to establish perfect CRISPR/Cas9 systems [242,243,244,245,246].

Although CRISPR/Cas9 has achieved certain results in forest tree breeding, compared with crops, the application of CRISPR/Cas9 system in gene function research and variety improvement of woody plants was is lagging behind, and most of these plants are still in the initial stage of gene editing system establishment. The preparation, efficient transformation. and regeneration of protoplasts for most forest species remains a technical problem. Therefore, the development of efficient protoplast transformation and regeneration systems plays an important role in the development of gene editing in trees. The selection of efficient editing targets relies on the existing plant sgRNA online database, which currently contains only a few woody plant genome information. The establishment of a public network platform for sgRNA design and detection of off-target effects for woody plants, and the collection of more varieties of genome data, especially the genome information of commonly used forest varieties, will be beneficial to increasing the application range of the CRISPR/Cas9 system. At present, research on the CRISPR/Cas9 system in forestry has mostly focused on economic forests and model plants, with relatively few studies on gymnosperms. Therefore, genome sequencing, CRISPR/Cas9 system optimization, and transformation system improvement in gymnosperms are essential. With the help of the CRISPR/Cas9 system, the literature has mainly focused on the role of target genes in growth and development, whereas research on environmental stress is still scarce. The CRISPR/Cas9 system requires Agrobacterium-mediated genetic transformation to obtain stable transgenic lines, but it is also accompanied by the insertion of exogenous DNA; hence, it is difficult to obtain noncontaminated gene editing plants via hybridization in the short term. Forest trees have a long growth cycle and juvenile period, with slow growth. The transgenic trees obtained using CRISPR/Cas9 systems have a large number of chimeras, and it is difficult to obtain homozygous tree varieties via hybridization technology in a short time. The off-target effects of CRISPR/Cas9 are common in forest trees; therefore, it is important to reduce their rate of occurrence. In addition, many new technologies derived from CRISPR/Cas9 system, such as CRISPR/Cpf1 system, targeted gene activation and inhibition, gene directional knock-in, and base editing technology (cytosine base editor and adenine base editor), are still very limited in woody plants; thus, broadening the application scope of these new technologies in trees will also become a future development direction.

7. Conclusions

After more than 200 years of development, modern forest genetic breeding has formed a complete system based on traditional breeding methods, supplemented by current biological techniques. The traditional breeding system based on hybrid breeding and seed orchard selection still occupies the dominant position in the breeding and production of forest varieties. The development of emerging biotechnology has greatly promoted the process of forest genetic breeding. It is still the development trend and emphasis of forestry genetics and breeding to select forestry varieties with high yield, high quality, strong resistance, and extensive influence before applying them to production. There are still many problems to be solved related to modern forest tree breeding basic theories or technical methods in guiding the rapid development of forest breeding. At present, popular varieties often have a single excellent trait, which cannot satisfy the needs of society. It is particularly important to cultivate excellent varieties that take into account yield and resistance. With global warming and the deterioration of the ecological environment, it is becoming more necessary to breed and cultivate new forest varieties with wide planting range, strong adaptability, and good tolerance through genetic breeding technology in harsh environments such as salinity and drought. With the development and continuous change of social needs, some forest varieties have aged and can no longer meet the needs of modern society. Therefore, it is crucial to continue to upgrade and cultivate tree species with high economic benefits, in line with the needs of social development. In addition, we should pay attention to the collection, protection, evaluation, and utilization of gene resources, strengthen the evaluation and management of collected and preserved gene resources, deepen the analysis of genetic information of main forest germplasm resources, and construct a database of important traits of main forest germplasm resources. Additional goals include strengthening the exploration of forest tree natural polyploid germplasm resources, developing new polyploid induction technology, especially for fast-growing and precious tree species, further improving the polyploid induction rate, comprehensively utilizing heterosis and ploidy advantage, and achieving rapid improvement of forest tree multi-objective traits. More genetic transformation systems of forest varieties should be established, and biotechnologies such as transgenic and gene editing should be used to rapidly integrate superior traits and shorten the breeding cycle. Breeding methods should be continuously innovated, new breeding methods in animal and crop breeding should be actively absorbed, and forest tree breeding methods should be enriched in order to increase the diversity of forest tree breeding.