Two Novel DNAs That Enhance Symptoms and Overcome CMD2 Resistance to Cassava Mosaic Disease

Abstract

Cassava mosaic begomoviruses (CMBs) cause cassava mosaic disease (CMD) across Africa and the Indian subcontinent. Like all members of the geminivirus family, CMBs have small, circular single-stranded DNA genomes. We report here the discovery of two novel DNA sequences, designated SEGS-1 and SEGS-2 (forsequencesenhancinggeminivirussymptoms), that enhance symptoms and break resistance to CMD. The SEGS are characterized by GC-rich regions and the absence of long open reading frames. Both SEGS enhanced CMD symptoms in cassava (Manihot esculentaCrantz) when coinoculated withAfrican cassava mosaic virus(ACMV),East African cassava mosaic Cameroon virus(EACMCV), orEast African cassava mosaic virus-Uganda(EACMV-UG). SEGS-1 also overcame resistance of a cassava landrace carrying the CMD2 resistance locus when coinoculated with EACMV-UG. Episomal forms of both SEGS were detected in CMB-infected cassava but not in healthy cassava. SEGS-2 episomes were also found in virions and whiteflies. SEGS-1 has no homology to geminiviruses or their associated satellites, but the cassava genome contains a sequence that is 99% identical to full-length SEGS-1. The cassava genome also includes three sequences with 84 to 89% identity to SEGS-2 that together encompass all of SEGS-2 except for a 52-bp region, which includes the episomal junction and a 26-bp sequence related to alphasatellite replication origins. These results suggest that SEGS-1 is derived from the cassava genome and facilitates CMB infection as an integrated copy and/or an episome, while SEGS-2 was originally from the cassava genome but now is encapsidated into virions and transmitted as an episome by whiteflies.

Importance: Cassava is a major crop in the developing world, with its production in Africa being second only to maize. CMD is one of the most important diseases of cassava and a serious constraint to production across Africa. CMD2 is a major CMD resistance locus that has been deployed in many cassava cultivars through large-scale breeding programs. In recent years, severe, atypical CMD symptoms have been observed occasionally on resistant cultivars, some of which carry the CMD2 locus, in African fields. In this report, we identified and characterized two DNA sequences, SEGS-1 and SEGS-2, which produce similar symptoms when coinoculated with cassava mosaic begomoviruses onto a susceptible cultivar or a CMD2-resistant landrace. The ability of SEGS-1 to overcome CMD2 resistance and the transmission of SEGS-2 by whiteflies has major implications for the long-term durability of CMD2 resistance and underscore the need for alternative sources of resistance in cassava.

FIG 1

Severe symptoms in resistant cassava cultivars in Tanzanian fields. Namikonga, an Amani hybrid (Manihot esculenta × M. glaziovii), was recommended in Tanzania as CMD resistant in the 1990s. AR40-6, an International Center for Tropical Agriculture (CIAT) breeding line with the CMD2 resistance locus (4), was introduced in Tanzania in 2004. Kibaha, which is most likely an Amani hybrid, was recommended in Tanzania as CMD tolerant in the 1990s. The top panels show plants with severe symptoms associated with EACMV infection, while the bottom panels show symptom-free plants. A yellow arrow marks the symptomatic AR40-6 leaf.

FIG 2

SEGS-1 and SEGS-2 clones, related cassava genomic sequences, and episomes. (A) Linear maps of SEGS-1 and SEGS-2 cloned sequences showing their GC-rich regions and flanking repeated sequences (green). The positions of cassava genomic sequences (FC1, PC2-1, PC2-2, and PC3-3; see Table S1 in the supplemental material) showing the strongest relationship to the SEGS-1 or SEGS-2 clones are marked by lines. The triangles indicate the positions and sizes of insertions or deletions detected in some genomic sequences related to SEGS-1. (B) Sequences of junction regions of SEGS-1 and SEGS-2 episomes compared to their respective clones and the longest related cassava genomic sequences. The red arrows mark the episome junctions in the linear sequences and the circular episomes (see Fig. S1 in the supplemental material for longer alignments.) The italic typeface and underlining in the clone sequences marks inverted repeats flanking the episome junctions of SEGS-1 and SEGS-2. (C) Comparison of SEGS-2 junction sequences with alphasatellite origin sequences. The vertical lines show the region of full or near identity between the two types of sequences. The apostrophe is the nick site in the alphasatellite origins, while the underlined sequences are the 5′ side of the stem structure in the origin hairpin. In panels B and C, blue typeface indicates sequence identity with the SEGS-2 clone, while black typeface marks differences. The alphasatellite sequences in panel C are associated with Radish leaf curl virus (RaLCV; GenBank accession number KF471057.1), Cotton leaf curl virus (CLCuV; HF564605.1 and HQ728354.1), Sida leaf curl virus (SiLCV; FR772088.1), Okra mosaic virus (OkMV; FN658718.1), Tomato yellow leaf curl China virus (TYLCCV; AJ888452.1), Tobacco curly shoot virus (TCSV; FN678903.1), Tomato yellow leaf curl Thailand virus (TYLCThV; FN678903.1), Bendi yellow vein mosaic virus (BYVMV; KF471059.1), Chili leaf curl virus (ChiLCV; KF471049.1), Ageratum enation virus (AgEV; FN543100.1), Malvastrum yellow mosaic virus (MaYMV; AM236767.1), and East African cassava mosaic Kenya virus (EACMKV; HE984148).

FIG 3

SEGS-1 and SEGS-2 enhance CMD symptoms. (A) Cassava cv. 60444 plants were bombarded with CMB DNA-A and DNA-B replicons alone or in combination with an SEGS-1 or SEGS-2 dimer plasmid under controlled conditions. (Left) Plants inoculated with ACMV, EACMCV, or EACMV-UG alone. (Middle) Each CMB coinoculated with SEGS-1. (Right) Each CMB coinoculated with SEGS-2. Photographs of representative leaves were taken at 7 dpi. (B) Range of leaf phenotypes (narrowing to extreme filiform) seen on cassava plants coinoculated with EACMV-UG and SEGS-1 at 21 dpi.

FIG 4

SEGS-1 can overcome CMD2 resistance of the cassava landrace TME3. TME3 plants were inoculated under controlled conditions and monitored for symptoms at 21 dpi. (A) Leaf from a plant inoculated with EACMV-UG DNA-A plus DNA-B showed no symptoms. (B) Leaf from a plant inoculated with EACMV-UG DNA-A plus DNA-B and SEGS-1 showed severe symptoms. (C) No symptoms were observed on leaves from plants inoculated with SEGS-1 alone. (D) Total DNA was extracted at 21 dpi from systemically infected leaves and equivalent leaves from symptom-free plants and was analyzed by DNA gel blotting with radiolabeled probes corresponding to EACMV-UG DNA-A. The lanes correspond to mock inoculation (M; lane 1), EACMV-UG (lane 2), EACMV-UG plus SEGS-1 (lane 3), and SEGS-1 alone (lane 4). The loading controls are ethidium bromide-stained total genomic DNA.

FIG 5

SEGS-2 enhances CMB infection in N. benthamiana. (A) N. benthamiana plants inoculated with EACMV-UG DNA-A plus DNA-B showed no symptoms, while plants coinoculated with EACMV-UG DNA-A plus DNA-B and SEGS-2 were symptomatic at 14 dpi. (B) Plants inoculated with EACMCV DNA-A plus DNA-B showed very mild symptoms, while those coinoculated with EACMCV DNA-A plus DNA-B and SEGS-2 displayed strong symptoms at 14 dpi. (C) Plants inoculated with SEGS-2 alone did not develop symptoms. (D) Total DNA was extracted at 14 dpi from systemically infected leaves and equivalent leaves from symptom-free plants and analyzed by DNA gel blotting with radiolabeled probes corresponding to EACMV-UG DNA-A (lanes 1 to 4) or EACMCV DNA-A (lanes 5 to 8). The lanes correspond to mock inoculation (M; lanes 1 and 5), CMB alone (lanes 2 and 5), CMB plus SEGS-2 (lanes 3 and 7), and SEGS-2 alone (lanes 4 and 8). The loading control is ethidium-stained, total genomic DNA.

FIG 6

Manihot genomes contain sequences related to SEGS-1 and SEGS-2. (A) PCR analysis of genomic DNA from South American Manihot genotypes using the SATIIF/R primer pair for SEGS-1 and the SATIIIF/R primer pair for SEGS-2. The table lists the 10 South American cassava cultivars and the 7 wild Manihot species that were analyzed. (B) PCR analysis of genomic DNA from African cassava cultivars using the 1-hp1F/1-6R primer pair for SEGS-1 and the 2-1F/2-5R primer pair for SEGS-2. The DNA samples were from plants rendered virus free by passage through tissue culture. (C) Multiple PCR products related to SEGS-1 were resolved for some African cultivars.

FIG 7

Amplification of SEGS-1 and SEGS-2 episomes in infected plants. (A) The convergent primer pairs 1-hp1F/1-6R and 2-7F/2-hp-0R amplify genomic copies of SEGS-1 and SEGS-2, respectively. The divergent primer pairs 1-2R/1-5F and 2-8R/2-5F amplify circular episomal or concatemeric copies of SEGS-1 and SEGS-2, respectively. Total DNA was the template for the genomic PCR products, while RCA DNA was the template for the episomal PCR products. (B) PCR products from CMB-infected cassava samples from Cameroon. The arrowheads mark bands with sequences that match SEGS-1 or SEGS-2. (C) PCR products from healthy cassava collected from Cameroon and passaged through tissue culture. The Arabidopsis thaliana (At) DNA was a control for potential contamination during DNA isolation. C+ is the positive PCR control using a cloned DNA template. C− is the negative PCR control that lacks template DNA. Bands marked with dots are nonspecific products that were also sequenced.

FIG 8

SEGS-2 episomes in infected leaves, virions, and whiteflies. (A) Divergent primer pairs were used to amplify episomal or concatameric copies of SEGS-1 and SEGS-2 from RCA template DNA. Convergent primer pairs were used to amplify CMB DNA and test for genomic DNA contamination of the RCA template DNA. The panels used the indicated primer pairs: SEGS-1 (1-4F/1-2R), SEGS-2 (2-4F/2-6R), CMB (EACMVAfor3/EACMVArev6 or UG3A-2/UG3A-3), and genomic DNA (Cass PeroxF/R). RCA template was produced using total DNA from infected cassava leaves (Cameroon, lane 2; Tanzania, lane 3), total DNA from whiteflies (Tanzania, lane 4), virion DNA from infected leaves (Tanzania, lane 5), and virion DNA from whiteflies (Tanzania, lane 6). Enhanced exposures of lanes 5 and 6 are shown at the right. C− is the negative PCR control that lacks template DNA (lane 7). C+ is the positive control using plasmid DNA corresponding to SEGS-1 (lane 8), SEGS-2 (lane 8), or CMB (lanes 8 and 9) or genomic DNA (lane 8) as the template. (B) Convergent primer pairs were used to amplify SEGS-1 or SEGS-2 in cassava leaves (Cv) from Tanzania using the same RCA template as that in lane 3 of panel A, which was shown to be free of genomic DNA. C− is the negative PCR control that lacks template DNA, while C+ is the positive control using the corresponding plasmid DNA. The arrowheads mark bands with sequences that match SEGS-2 (2-6F/2-4R) or SEGS-1 (1-2F/1-4R). Bands marked with a number sign are CMB PCR products. Bands marked with dots are nonspecific products that were also sequenced.