Cassava Program

RSA1 - Enhancement of Genetic Resources

Cassava Breeding and Prebreeding

Lead: Dr Augusto Becerra López-Lavalle

Partner institutions

• CIAT Genetic Resource Unit (URG-CIAT), IITA, Leibniz Institute-DSMZ, KU, BMGF Germany, Ingredion, Colombia: AGROSAVIA
• Peru: UNDAC – Universidad Nacional Daniel Alcides Carrión (UNDAC), Instituto Nacional de Innovación Agraria (INIA), Federación de Comunidades Nativas Yaneshas (FECONAYA), Instituto del Bien Común (IBC)
• UK: NRI – Natural Resources Institute (NRI), Royal Holloway Univ of London (RHUL), Univ of Greenwich
• Nicaragua: Instituto Nicaragüense de Tecnología Agropecuaria
• Tanzania: Tanzania Agricultural Research Institute (TARI)
• Uganda: National Crops Resources Research Institute (NaCRRI)
• USA: Cornell University, Univ of California Riverside (UCR)
• Vietnam: Agricultural Genetics Institute (AGI), RDRDC – Root Crop Research and Development Center (RDRDC), Hung Loc Agricultural Research Center (HLARC)

The goal of the RSA-1 team is to contribute to decreasing hunger by providing solutions targeting agronomic or economic problems that can have either genetic or agronomic solutions, thus increasing productivity, sustainability, and use of cassava in Latin America and the Caribbean (LAC) and Southeast Asia (SEA).

Plant genetic resources underpin crop improvement. Thousands of cassava landraces have been collected by CIAT in their center of origin, which constitute the global cassava reference collection, but add-value traits have been discovered and integrated into variety development, for example, high pro-Vitamin A, waxy starch and disease resistance. The challenge is to efficiently and effectively use the untapped genetic resources to continuously improve cassava production and sustainability. The Alliance Bioversity and CIAT Cassava Program has strategically invested both in genomics-based germplasm characterization and population improvement and in high throughput and accurate phenotyping to understand the genetic diversity of the global cassava reference population, discover and integrate the favorable haplotype into elite populations, and deliver improved populations or final products to the global cassava community.

Current research in breeding and prebreeding focuses on next-generation breeding technology, carotene expression profiling in cassava roots, starch composition (waxy starch), cassava mosaic disease resistance, cassava ORANGE protein characterization, induction of doubled haploids, genetic transformation and gene editing, and genomic-based characterization and identification of genetic diversity in the global cassava reference collection.

Breeding, Prebreeding and Next-Generation Breeding

Lead: Xioafei Zhang

Collaborators: Sandra Salazar, Nelson Morante, Hernán Camilo Vargas, Jorge Iván Lenis, Thierry Tran

Clones with dual resistance to cassava mosaic disease (CMD) and cassava brown streak disease (CBSD)

Eighteen full-sib families of CBSD x CMD clones were analyzed together with Dr Stephan Winter (DSMZ, Germany) to screen for CBSD resistance. Segregation within full-sib families indicated that CBSD resistance could be a dominant trait. Moreover, we identified seven clones with dual resistance to CMD and CBSD and shared these with African national programs. These clones serve as CBSD donors for breeding programs, thus providing a new solution to the CBSD pandemic in Africa.

Understanding target population of environment (TPE) in global cassava production regions

In collaboration with the Alliance’s Climate Action Team, we analyzed the climate similarity of global cassava production regions. We observed that the Caribbean coast of Colombia (in the sub-humid and semi-arid lowland tropics) represents cassava growing regions that account for about 50% of global production, including Africa and SEA. Thus, we are focusing our breeding activities in this area to develop cassava varieties or populations that are adapted to major cassava production regions, especially in Asia and Africa.

Validating CMD2 markers in breeding populations

We screened a multi-parental population and the elite parents and found two markers that explain 51% of the population variance in CMD severity. These two markers will be used for marker-assisted selection (MAS) to develop CMD-resistant varieties for Africa and Asia. We also observed a high rate of co-segregation (73%) between the two markers on different chromosomes (12 and 14), which requires further investigation.

Implementing genomic prediction

We developed a training population for genomic prediction using seeds harvested from the 2019-2020 pollination season. The training population was derived from 22 progenitors with high dry matter, 9 with good cooking quality, and 12 with CMD or whitefly resistance. In total, 392 clones from 44 full-sib families were increased for stake production. We established replicated yield trials in two target environments in May 2021 to develop genomic prediction models.

Using CassavaBase to manage yield trial data

In 2020, 64 trials were conducted covering all stages of seven breeding pipelines, three of which are dedicated to Africa. All the agronomy and quality data were managed in CassavaBase (online analysis tool of the NexGen Cassava Program). Shared checks among all breeding pipelines allow connectivity between trials within and among breeding populations, which facilitates the calculation of genetic gains.

Identifying CMD-resistant clones for SEA

Among cassava clones introduced in Vietnam from CIAT and IITA, we identified nine clones showing CMD resistance and 30% higher starch yield than KU50 (10.5 vs. 7.9 t/ha) at two locations in Vietnam. KU50 is the predominant variety in SEA. The clones were advanced to regional yield trials at seven locations under medium or high CMD pressure. The best clones will be released as the first generation of varieties for cassava farmers in the region.

Cassava Molecular Genetics Lab

Adriana Bohórquez Chaux and Luis Augusto Becerra López-Lavalle

Collaborators: María Isabel Gómez Jim&eacutr;nez, Anestis Gkanogiannis, Carmen Bolaños, Carlos Ordóññez, Adriana Vásquez, Gerardino Pérez, Daniel Encarnación, Janneth Patricia Gutiérrez

The cassava genetics group supported the cassava research community in maintaining high-quality control of population development. The Ugandan 5CP lines (Mkumba, NASE3 and NASE14 and their susceptible comparators, Sauti and Mkuranga) are the parental materials for crosses directed toward improved whitefly resistance (WFR), associated with virus-disease resistance. Although the genetic and biochemical pedigree of the selected 5CP genotypes is poorly understood, they are likely to be hybrids or non-“true-breeding.” Sequence polymorphisms and molecular features were identified to enable their use as quantitative trait markers in breeding programs. The 5CP accessions (510 samples from Africa) were processed using a SNP chip for the Nanofluidic Dynamic Arrays (SNPY-Array; Fluidigm^®, USA) developed by our group, which contains 96 single nucleotide polymorphisms (SNPs) for genotyping in cassava. The technique allowed simultaneous collection of both endpoint and real-time data from a unique chip cell with 97% confidence.

Genetic duplicates test: In our variety identification test, all samples that are genetic duplicates belong to the same group. In total, we found 40 genetic duplicate groups (GD-groups) that represent 40 different genotypes. This set of duplicate samples contains 508 samples from Uganda, Tanzania and Malawi.

African Cassava Whitefly Project (ACWP)

Advanced intercrossing is in progress to create pre-breeding cassava progenies homozygous for the whitefly resistance (WFR) loci and possessing superior resistance to ECU72 (the original WFR donor). Cassava lines homozygous for WFR are needed to transfer superior resistance into regionally preferred, African-adapted cassava germplasm. First, we generated two advanced crosses (CM8996 and GM8586). From these F1s, we developed 12 “advanced crossover” F2 populations for WFR. Seeds from the offspring were multiplied in the greenhouse. Analysis of the progeny harboring all three WFR regions, a subset of these regions and one completely lacking these regions will determine the ability of these markers to identify the superior WFR seen in the F2 progeny of the CM8996 and GM8586 crosses.

We developed a high-throughput, quantitative phenotyping method for WFR in cassava, which consists of a greenhouse experimental design and an ImageJ plugin called Nymphstar, for automated counting of nymphs. The progeny (F1) of ECU72 (WFR) and COL2246 (whitefly susceptible or WFS) crosses (CM8996), as well as ECU72 (WFR) and TMS60444 (WFS) crosses (GM8586), were phenotyped over four consecutive years. Phenotyping for five of these F2s has been completed (AM1588, GM12200, GM12201, GM12202 and GM12199). The true-F2 AM1588 (CM8996-199 x CM8996-199) was selected for its high resistance to the whitefly Aleurotrachelus socialis. DNA was extracted from 200 offspring of true-F2 AM1588, and a RAD-sequencing approach was used. The 15 top resistant and 15 top susceptible offspring of this F2 were selected for assessment. Whitefly infestations were performed and paired samples split for RNA and metabolite analyses. RNAs were sent to UCR for RNA-seq (eQTLs) analysis, while tissue was sent to RHUL for chemotyping. The SNPs analysis, mapping and quantitative trait locus (QTL) analysis are ongoing. From these we aim to identify molecular markers located within the cassava WFR genes. The advanced intercrossed cassava lines will next be used to (i) identify, verify and refine QTLs for WFR, and (ii) identify potential epistasis for WFR QTL.

Cassava Genetics Tissue Culture Lab

Two families (GM13489 and GM13494) from the cassava breeding program, with 300 individuals showing CMD-CBSD resistance, were conserved in vitro to support CIAT-Asia breeding programs. Other lines conserved in vitro include 104 individuals from advanced selection for beta-carotene content (2019) to support of the cassava breeding program.

WFR in a unique cassava germplasm array in the Latin American Manihot esculenta tribe comprising seven subpopulations

Of the 6,240 cassava accessions held at CIAT, nearly 20% (1,200) were selected for genetic profiling based on the georeference information and frequency of cassava landraces used to generate CIAT elite breeding lines or cultigenes. In all, 18,286 SNPs were obtained from 292 LAC cassava landraces. Seven subpopulations were resolved, and 219 unique individuals were identified and will be phenotyped for WFR to undertake a genome-wide association study. This approach will allow us to unravel the genomic regions that may be responsible for whitefly resistance in cassava.

Of the 219 unique genotypes, 172 were evaluated using the phenotyping method described above, including Nymphstar image analysis. Genotype COL1468, commonly known as CMC40, was used as a susceptible check to later perform genome-wide association studies (GWAS) with data on the SNPs.

Five significantly different groups in the response to A. socialis were identified. Among those, 20 unique genotypes (11.6%) showed the highest levels of resistance to the whitefly A. socialis, according to the parameter measured (nymph count), which included genotype ECU72, known for its high levels of resistance. Another 20 genotypes (12%) showed the highest levels of susceptibility, including the susceptible check COL1468. Fourteen genotypes showed intermediate levels of resistance (green), and 37 (yellow) showed intermediate levels of susceptibility. Nearly half (48%) of the genotypes were classified as intermediate in their response to whitefly.

A collection of 54 varieties were sent to Agrosavia (Colombia) for the Llanos Orientales region (Eastern Plains).

Cassava genomics, metabolomics and proteomics

Lead: Luis Augusto Becerra López-Lavalle

Collaborators: Diana Katherine Castillo, Tatiana Melissa Ovalle, Diana Victoria Marín, Janneth Patricia Gutiérrez, Daniel Álvarez, Angélica Jaramillo, Jairo Rodríguez

Carotene expression profiling

This work aims to characterize the expression of nine candidate genes associated with high carotene content in cassava. The expression patterns of six cassava high-carotene genotypes were evaluated monthly. In 2020, we recorded expression data from all samples at different growing stages. One of the candidate genes, BETA6 (Manes.09G032800), showed a 200-fold expression change in the six genotypes evaluated relative to a calibrator (PAR39). This gene is a good candidate for identifying genotypes with high carotene content in a cassava population. Six primer sets belonging to genes regulated by BETA6 were designed and standardized to evaluate its contribution to BETA6 expression.

Cassava waxy starch variation

Based on evidence collected in 2019, we designed an amplification protocol to obtain the full length of the GBSSI gene (3500bp) by PCR amplification. The fragment obtained was sequenced using the MinION sequencing platform. The aim is to capture internal variation of the gene by sequencing the full fragment without overlapping segments.

Molecular characterization of CMD resistance

Cassava Mosaic Disease (CMD) is caused by a virus belonging to the Geminiviridae virus family and can be transmitted via stakes from infected cassava mother plants or by the whitefly Bemisia tabaci insect vector. The plant material for this study consisted of 42 crosses between a Nigerian variety TME-3 that represents a new source of CMD resistance and an improved disease-tolerant line (TMS 30555), which was classified as susceptible for this cross, given the extreme nature of the resistance under study. Candidate amplicons were cloned and sequenced for further analysis.

Carotenoids and the cassava ORANGE protein characterization

In recent years, ORANGE protein (OR) has received special attention for its role in carotene biosynthesis. OR is a chaperone not directly involved in carotenoid biosynthetic pathways. It promotes the accumulation of carotenoids and their stability in several plants. In this project, our aim was to identify, characterize and investigate the role of OR in the biosynthesis and stabilization of carotenoids in cassava and its relationship with phytoene synthase (PSY), the rate-limiting protein in carotenoid biosynthesis. OR expression levels and protein amounts were measured along carotenoid levels in roots of one white (60444) and two yellow cassava cultivars (GM5309-57 and GM3736-37). Four OR variants were found in yellow cassava roots. Expression levels in variants 1 (MeOR_X1) and 2 (MeOR_X2) remained unchanged, but significantly higher OR protein amounts were observed in the yellow varieties. Cassava PSY1 gene expression was significantly higher in the yellow cultivars although protein levels remained unchanged.

Next-generation breeding and gene editing

Lead: Paul Chavarriaga

Collaborators: Franciso Sánchez, Juan Pablo Arciniegas, Didier Marín, Orlando Vacca, Aníbal Peñaloza, Jhon Larry Moreno, Thierry Tran, Anestis Gkanogiannis, Alejandro Brand

Doubled-haploid induction

In maize, haploids are routinely produced at a rate of 2-3% maternal haploid seed when self-pollinated or outcrossed as a male, ie 10-20 times higher than the spontaneous haploid induction rate (HIR), which occurs through a process called gynogenesis, where the male gametes induce haploid embryo formation from the female chromosomes only. The genetic mechanism behind Stock 6 haploid induction line has been linked to a mutation in gene GRMZM2G471240. CRISPR/Cas9-mediated genome editing was used to knock out ZmPLA1 (Liu et al., 2017) and conduct pollination assays using self-pollinated knockout lines. HIR in these experiments ranged from 3.7% to 6.67%. Offspring from breeding male knockout lines with wild-type lines led to haploids with chromosomes stemming exclusively from the maternal genomes, as in the case of the original gene mutation in Stock 6.

Candidate homologs to GRMZM2G471240 were identified in cassava two genes (Manes.12G093400 and Manes.12G102200) being the most closely related to the maize gene and thus suitable candidate genes to create knockout mutations via gene editing. Candidate events are being further characterized molecularly and agronomically.

As expected for a haploid inducer line, regenerated cassava showed reduced vigor, dwarfism, thin stems, and partially wrinkled leaves with narrow lobes, whereby somaclonal variation is still a possibility that needs to be excluded. Plants were exposed to red light in the greenhouse to stimulate early flowering once established in the field. When transferred to the field, plants conserved the altered phenotype. Some lines flowered and produced both female and male flowers. Male flowers from T310718 lines were used to pollinate TMS60444 flowers. Plants from these crosses established in the field to be crossed with elite cultivars.

Waxy starch cassava

The waxy phenotype is due to a mutation or down-regulation of the granule-bound starch synthase I GBSSI) gene, a protein responsible for amylose synthesis. The cassava clone AM 206-5, containing 0% amylose, has been characterized (Ceballos et al., 2007), and it was found that the allelic variant in GBSSI is due to a single-nucleotide deletion that causes a frameshift, producing a premature stop codon in the sixth exon (Aiemnaka et al., 2012). This suggests that gene editing can be used to generate similar knockout alleles in other cassava cultivars.

Three guide RNAs (gRNAs) were designed based on the consensus sequence between the KU50 GBSSI coding sequence (Genbank accession JF708948.1) for cloning into a binary vector. The three corresponding gRNA constructs were tested via independent transformation essays. A total of 163 plants were regenerated from transformations with three different gRNAs constructs, among which 129 independent transgenic events were identified. In the iodine staining for starch, approximately 95% of the independent transgenic lines showed shades of the characteristic brown-red coloration of typical waxy stems, just like AM206-5, while the control wild type variety 60444 stained dark blue. In total, 82% of the lines showed an unmistakable brown-red color in the stems, thus indicating changes in amylose content. Among lines grown for 12 months in the field, 90% were waxy lines, with the root yields between 20 and 21.5 kg.

Varying root dry matter content (DMC) was detected depending on the construct used. In general, PHSE401GBSSI#4 lines had lower DMC TMS60444 (all below 30%). This construct targeted exon 6, the same exon mutated in AM 206-5, while PHSE401GBSSI#1 and pHSE401HBSSI#5 produced individuals with over 30% DMC. These constructs targeted exons 9 and 2, respectively, suggesting that mutations in these two regions may produce waxy cassava lines without greatly affecting DMC. Nevertheless, a large-scale field trial is required to statistically confirm the tendency observed.

Selected waxy lines (PHSE401GBSSI#1-32, PHSE401GBSSI#4-54, PHSE401GBSSI#5-4, PHSE401GBSSI#5-5 and PHSE401GBSSI#5-8) were used in a large-scale field trial to obtain data on yield, DMC, stake germination, and amylose content. The waxy lines used have between one and two T-DNA copies and one and three alleles.

HCN content in all lines was within the accepted range <100 µg HCN/g cassava WB). PSHE401GBSSI#5 lines had the highest root DMC, starch in flour, and flour in fresh roots, and are possibly the most promising for high yields.

To obtain non-transgenic waxy lines, T-DNA must be segregated through sexual reproduction. The challenge in this case is the asynchronicity of male and female flowering in cassava for self-pollination purposes and therefore, crossing between “sister” lines has been implemented, until self-pollinations can be performed regularly. The waxy phenotype is a recessive trait (Ceballos et al., 2007), meaning that non-waxy alleles must be eliminated during crossing and selection.

Editing cassava SWEET genes for resistance to Xanthomonas

One commercially important bacterial disease is cassava bacterial blight (CBB), caused by the phytopathogenic bacterium Xanthomonas axonopodis pv. Manihotis (Xam), which can lead to total crop loss. Pathogenesis mechanisms of Xam bacteria are proteins of the AvrBs3/PthA family or TAL effectors (TALEs), for Transcriptional Activator-Like proteins, which modulate gene expression in host cells. TALEs can activate susceptibility genes in the plant, which play an important role in the plant-pathogen interaction. On the other hand, TALEs can induce or repeat resistance genes that are activated by the plant to defend against infection. One gene related to this disease is of the SWEET/Nodulin-3 gene family, known to be targeted by TALEs of X. oryzae, the causal agent of bacterial leaf blight in rice. The SWEET genes are transporters providing a source of sucrose for plants. During infection by bacteria such as Xanthomonas, these genes are overexpressed, making them an important factor in the infection and colonization process.

The sequences of the cassava SWEET homologs MeSweet10a and MeSweet10b genes were identified and guide RNAs designed designed for CRISPR/Cas9 assisted gene editing. For the MeSweet10a gene, the guide RNA was directed to the EBE site (effector-binding element) recognizing the TAL20 of strain Xam668. For the MeSweet10b gene, the guide RNA was directed to exon 1 and 3 of the gene to generate a knockout mutant. Two designs with double targets were also developed. The first aimed to generate a complete knockout (deletion) of the MeSweet10a gene, with targets designed for the upstream and downstream untranslated RNA regions, while the second aimed to generate a double knockout for the MeSweet10a and 10b genes in exon 1 and another in exon 2 of the MeSweet 10b gene. So far, a total of 182 lines of the six constructs used have been generated. Characterization of the knockout mutants is ongoing.

RSA-1 Publications

1. Becerra López-Lavalle, LA, Bohorquesz-Chaux A, ZHANG X (2021) Identification of cassava varieties in ex-situ collections and global farmer’s fields: An Update from 1990 to 2020. In: Frediansyahk A (ed.). Cassava - Biology, Production, and Application. IntechOpen.
2. Behnam B, Higo A, Yamaguchi K, Tokunaga H, Utsumi Y, Selvaraj MG, Seki M, Ishitani M, Becerra López-Lavalle LA, Tsuji H (2021) Field-transcriptome analyses reveal developmental transitions during flowering in cassava (Manihot esculenta Crantz). Plant Molecular Biology 106:285-296.
3. Fernando A, Selvaraj M, Chavarriaga P, Valdés S, Tohme J (2021) A clearinghouse for genome-edited crops and field testing. doi.org/10.1016/j.molp.2020.12.010
4. Hackathon to develop market segments and product profiles for breeding programs. https://cgspace.cgiar.org/handle/10568/110979
5. Ige AD, Olasanmi B, Mbanjo EG, Kayondo IS, Parkes EY, Kulakow P, Egesi C, Bauchet GJ, Ng E, Becerra López-Lavalle LA, Ceballos H, Rabbi IY (2021) Conversion and validation of Uniplex SNP markers for selection of resistance to cassava mosaic disease in cassava breeding programs. Agronomy 11.
6. Irigoyen, ML, Garceau, DC, Bohorquez-Chaux A et al (2020) Genome-wide analyses of cassava Pathogenesis-related (PR) gene families reveal core transcriptome responses to whitefly infestation, salicylic acid and jasmonic acid. BMC Genomics 21:93. https://doi.org/10.1186/s12864-019-6443-1
7. Jaramillo AM, Sierra S, Chavarriaga-Aguirre P, Castillo DK, Gkanogiannis A, Becerra López-Lavalle LA, Arciniegas JP, Sun T, Li L, Welsch R, Boy E, Álvarez D (2020) Characterization of cassava ORANGE proteins and their capability to increase carotenoid accumulation in cassava. PLOS ONE. In press.
8. Monroe JG, Arciniegas JP, Moreno JL, Sánchez F, Sierra S, Valdés S, Torkamaneh D, Chavarriaga P (2020) The lowest hanging fruit: Beneficial gene knockouts in past, present, and future crop evolution. Current Plant Biology. Doi.org/10.1016/j.cpb.2020.100185
9. Moreno-Cadena P, Hoogenboom G, Cock JH, Ramírez-Villegas J, Pypers P, Kreye C, Tariku M, Ezui KS, Becerra López-Lavalle LA, Asseng S (2021) Modeling growth, development and yield of cassava: A review. Field Crops Research 267: 108140.
10. Ocampo J, Ovalle T, Labarta R, Le DP, De Haan S, Vu NA, Kha LQ, Becerra López-Lavalle LA (2021) DNA fingerprinting reveals varietal composition of Vietnamese cassava germplasm (Manihot esculenta Crantz) from farmers’ field and genebank collections. Plant Molecular Biology.
11. Pérez-Fons L, Ovalle TM, Maruthi MN, Colvin J, López-Lavalle LAB, Fraser PD (2020) The metabotyping of an East African cassava diversity panel: A core collection for developing biotic stress tolerance in cassava. PLoS ONE 15(11): e0242245. Doi.org/10.1371/journal.pone.
12. Pineda M, Morante N, Salazar S, Cuásquer J, Hyde PT, Setter TL, Ceballos H (2020) Induction of earlier flowering in cassava through extended photoperiod. Agronomy 10: 1273. doi:10.3390/agronomy10091273
13. Pineda M, Yu B, Tian Y, Morante N, Salazar S, Hyde PT, Setter TL, Ceballos H (2020) Effect of pruning young branches on fruit and seed set in cassava. Front. Plant Sci. 11:1107.
14. Thierry T, Zhang X, Ceballos H, Moreno JL, Luna J, Escobar A, Morante N, Belalcazar J, Becerra LA, Dufour D (2020) Correlation of cooking time with water absorption and changes in relative density during boiling of cassava roots. International Journal of Food Science & Technology. doi.org/10.1111/ijfs.14769
15. Uke A, Tokunaga H, Utsumi Y, Vu NA, Nhan PT, Srean P, Hy NH, Ham LH, Becerra López-Lavalle L A, Ishitani M, Hung N, Tuan LN, Van Hong N, Huy NQ, Hoat TX, Takasu K, Seki M, Ugaki M (2021) Cassava mosaic disease and its management in Southeast Asia. Plant Molecular Biology.