Tea (Camellia sinensis) is the major plantation crop in Sri Lanka contributing about 15% of total export earnings. Canker is the most widely prevalent stem disease of tea worldwide. Cankers are persistent, attacking twigs, stems and killing or girdling branches ultimately leading to death of tea bushes.

Phomopsis theae has been identified as the major pathogen causing stem canker disease in Japan (Kasai et al. 1965), India (Venkata Ram 1979), China (Chen and Chen 1982), Kenya (Onsando 1988) and Bangladesh (Ahmad et al. 2013). In Sri Lanka, Macrophoma theicola and Botryodiplodia theobromae (Sabanayagam et al. 1974) are the major canker causing pathogens in the low (<600 m mean sea level) and mid-elevations (1200–600 m mean sea level) whereas P. theae is problematic in the high elevations (>1200 m mean sea level) (Balasuriya 2008). Liyanage et al. (2013) reported Fusarium solani as a canker causing agent for the first time in the low elevation tea growing areas of Sri Lanka. Recently in Sri Lanka, the incidence of collar canker and associated bush dieback has become widespread. This study aimed at identifying the causal organism in different tea growing regions.

Dieback incidence in affected fields ranged from 50 to 85%. The affected tea bushes had fewer and smaller leaves, yellowing of leaves and premature falling. Heavy flowering and fruiting commonly followed canopy thinning. Progressive dieback of branches resulted in bare bushes (Fig. 1a). Foliar symptoms were always associated with cankers having cracks and peeling of bark at the collar region. The exposed cambium and wood were discoloured, tan, brown or black. Cankers at the collar region were irregular in shape and extended to several centimetres in length and depth (Fig. 1b). Reddish, tiny superficial perithecia were seen under damp and moist conditions (Fig. 1c). Fibrous root formation just above the cankered area was also observed.

Fig. 1
figure 1

a Tea bush naturally affected by collar canker and dieback (b) natural infection showing the prominent canker at the collar region (c) perithecia on an infected tea bush (d) necrotic tissue beneath the point of artificial inoculation (note the bark has been removed to expose internal tissue browning)

Symptomatic samples were collected from tea fields covering a range of rainfall and elevation during the period 2013–2014. Pieces (3–5 mm3) of tissue from the margin of canker lesions were separately plated on Fusarium selective pentachloronitrobenzene (PCNB) medium (Papavizas 1967), potato dextrose agar (PDA, Oxoid), and malt extract agar (MEA, Oxoid) supplemented with tetracycline (10 mg/ml). Fungi were allowed to grow for seven days in the dark at room temperature (22 ± 2 °C), sub-cultured and purified on PDA. The isolates identified as fusaria were grown on carnation leaf agar (CLA) at 27 °C to stimulate conidia development (Fisher et al. 1982). The colony characters of Fusarium isolates were compared with diagnostic characteristics specified by Leslie and Summerell (2006) and Nirenberg and O’Donnell (1998). Selected isolates were deposited at the International Collection of Microorganisms from Plants (ICMP).

Four isolates were selected for molecular characterisation, being one from Ruwanwella (ICMP 21513), one from Norton (ICMP 21515), and two from Yatinuwara (ICMP 21518, ICMP 21519). DNA was extracted using a DNeasy plant mini extraction kit (Qiagen) from mycelium of 7–10 days old monoconidial cultures grown on PDA. Internal transcribed spacer regions and intervening 5.8S rRNA gene (ITS) of the rDNA operon were amplified using the primer pair ITS-1F (5′-CTTGGTCATTTAGAGGAAGTAA-3′) / ITS-4 (5′-TCCTCCGCTTATTGATATGC-3′), and PCR conditions of White et al. (1990). Translation elongation factor 1- α (tef1) was amplified using the primer pair EF-1 (5′-ATGGGTAAGGA(A/G)GACAAGAC-3′) / EF-2 (5′-GGA(G/A)GTACCAGT(G/C)ATCATGTT-3′), following PCR conditions described by O’Donnell et al. (1998). Amplicons of about 560 bp and 700 bp (ITS and tef1, respectively) were sequenced through Macrogen (Korea) and deposited in GenBank (Accession No KJ009327 to KJ009330, KR707726 to KR707729).

Sequences generated in this study and others obtained from GenBank (Table 1), composed two datasets (ITS and tef1) which were separately aligned using Mafft v.7 (http://mafft.cbrc.jp/alignment/server/index.html). Ambiguous regions were removed and, whenever necessary, the alignments were manually improved. The individual and combined ITS and tef1 sequences were subjected to a Maximum Likelihood analysis (ML) using RaxMLGUI version 1.3.1 (Silvestro and Michalak 2012). The optimal ML tree search was conducted with 1000 separated runs, using the default algorithm of the program, from a random starting tree for each run. The final tree was selected among sub optimal trees, from each run, under the GTR + G substitution model. The resulting tree was edited and printed with TreeGraph 2 v. 2.11.1 (Stöver and Müller 2010).

Table 1 Sequences used in the phylogenetic analysis, newly generated sequences are in bold

Pathogenicity tests were carried out on vegetatively propagated 1 year and 10 years old tea plants (cv TRI 2023) with the selected isolates, ICMP 21513, ICMP 21515, ICMP 21518 and ICMP 21519. Each isolate/control was inoculated on to ten replicate plants. A wound (5 mm diameter) was made on the main stems of test and control plants by removing the bark with a cork borer. Test plants were inoculated by placing a mycelia plug (5 mm in diameter) cut from the margins of seven-day-old monoconidial cultures of each isolate, on the wounds. Control plants were inoculated with sterile PDA plugs without mycelia. The plants were maintained under recommended nursery or field conditions for tea plants. The pathogen was re-isolated from lesions after development of symptoms.

Ninety percent of 96 isolates obtained from symptomatic tissues were fusaria. The 4 selected isolates were identified as the members of Fusarium solani species complex (FSSC) based on morphological characters (Fig. 2 A-F, supplementary data) and analysis of two DNA loci (ITS and tef1). The combined phylogenetic analysis (Fig. 3) revealed that all 4 isolates from tea belonged to FSSC clade 3 (O’Donnell 2000). Two isolates (ICMP 21513 and ICMP 21518) formed a sub-cluster within the Ambrosia Fusarium Clade (AFC), together with F. pseudensiforme NRRL 46517 (Kasson et al. 2013) and F. cf. ensiforme NRRL 22354 (Kasson et al. 2013). The isolates ICMP 21519 and ICMP 21515 clustered on a distinct clade, together with F. solani AF14 and Nectria haematococca (Fig. 3).

Fig. 2
figure 2

Colony on PDA of (a) ICMP 21515 (b) ICMP 21518 (c) ICMP 21513 and (d) ICMP 21519, (e) micro and macro conidia of ICMP 21519 (f) perithecia of ICMP 21513 on PDA

Fig. 3
figure 3

Best scoring RAxML tree of Fusaria obtained from concatenate sequence alignment of ITS and tef1 region. The scale bar shows 0.02 substitutions per site and bootstrap replicate values from 1000 replicates are shown next to the branches. Isolates obtained in this study are in bold

The AFC includes symbiont fusaria associated or not with ambrosia beetles (Kasson et al. 2013). Fusarium ambrosium is a symbiont of the Tea shot-hole bore (TSHB), Euwallacea fronicatus (= Xyleborous fornicatus), a serious pest of tea in India and Sri Lanka (Gadd and Loos 1947). Fusarium pseudensiforme and F. cf. ensiforme are not known to be associated with ambrosia beetle (Nalim et al. 2011; Kasson et al. 2013). These two species and the isolates associated with collar canker and dieback of tea differ from F. ambrosium by having typical fusiform macroconidia, the latter produces a clavate macroconidia (Gadd and Loos 1947). Members of N. haematococca are also economically important plant pathogens causing root rot, fruit rot and dieback in important plant species (O’Donnell 2000).

The occurrence of isolates belonging to the AFC and N. haematococca clade indicate the genetic variability of the casual agents of collar canker and dieback of tea. Artificial inoculation of the selected isolates from each group resulted in prominent cankers in 50–80% of the nursery plants and 70–90% of the field plants, confirming its pathogenicity (Fig. 1d). This study identified causal agents of collar canker and dieback of tea in Sri Lanka to be members of FSSC clade 3 and established the existence of diversity among the isolates. The information generated will help to develop intervention strategies and efforts to reduce the economic impact of this recently emerged disease of tea.