Breeding for Resistance to Basal Rot in Narcissus

Breeding for Resistance to Basal Rot in Narcissus J.H. Carder and C.L. Grant Horticulture Research International Wellesbourne Warwick, CV35 9EF U K Tel: +44(0) 1789 470382 Fax: +44(0) 1789 470552 E-mail: john.carder@hri.ac.uk Keywords: Basal rot, breeding, Fusarium, Narcissus, resistance Abstract Basal rot of Narcissus caused by Fusarium oxysporum f. sp. narcissi frequently results in major losses in UK commercial bulb crops. Until recently, resistance breeding programmes have used as parents only tetraploid varieties, none of which have complete resistance to basal rot. The diploid species N. jonquilla, which has been reported to be completely resistant, was used as the pollen parent in crosses with two varieties ( Golden Harvest and St. Keverne ) that display opposite extremes of field resistance and the seedling progenies were clonally propagated. The hybrids were screened for resistance by planting bulbs in compost infested with chlamydospores of the pathogen and recording bulb survival one year later. Progeny from self-pollinated N. jonquilla were also propagated and tested to study the inheritance of resistance in an F 1 generation from the diploid species. All of the N. jonquilla selfs displayed similar very high survival rates indicating homozygous resistance in this species. A continuous distribution of survival rates from 0 to 100% was observed in both the hybrid progenies but more lines with 100% survival were found when St. Keverne, a variety with high field resistance, was the maternal parent than when Golden Harvest, a more susceptible variety was used. A detached-scale puncture-inoculation technique has been developed as a rapid in vitro assay for basal rot resistance in Narcissus. Varying amounts of rot were recorded for eight varieties and five species that correlated well with their known resistance to basal rot. INTRODUCTION The daffodil (Narcissus L.) is the most important ornamental bulb crop grown in the United Kingdom. Basal rot, caused by the soil-borne fungus Fusarium oxysporum f. sp. narcissi can result in high levels of bulb rots both in the soil and during storage. Many of the most popular UK commercial varieties, e.g. Carlton and Golden Harvest, are highly susceptible to the disease. During the period from 1963-1990, crosses between named Division 1 and Division 2 (Kington, 1998) varieties generated large numbers of hybrid seedlings, some of which displayed improved resistance (Tompsett, 1986; Bowes et al., 1992). Resistance to basal rot has also been identified in certain Narcissus species (Linfield, 1992) although N. pseudonarcissus, a species from which large-flowered trumpet daffodils (Division 1) are believed to have originated (Wylie, 1952), is highly susceptible. N. x odorus L. var rugulosus Hort. is an F 1 hybrid between N. pseudonarcissus and N. jonquilla and is resistant, as is the male parent N. jonquilla. It was hoped that this species resistance would also be inherited by the F 1 progeny resulting from crosses using N. jonquilla as pollen parent and two commercial varieties, Golden Harvest and St. Keverne as seed parents. This paper describes the production, clonal propagation and resistance screening of hybrid lines generated from these crosses. A novel rapid in vitro assay for basal rot resistance has also been devised and tested. MATERIALS AND METHODS Proc. 8th Int. Symp. on Flowerbulbs Eds. G. Littlejohn et al. Acta Hort. 570, ISHS 2002 255

Production, Propagation and Screening for Resistance In the spring of 1991 flowers of the two seed parents, Golden Harvest (GH) and St. Keverne (SK) were emasculated before being fertilised with pollen from N. jonquilla (NJ). Forty different pollen donor plants were used (Table 1). Seeds from these crosses and from self-pollinated N. jonquilla were collected and planted during early summer in trays of multipurpose compost plus 10% v/v grit and placed in a glasshouse with supplementary heating to provide protection from frost. In 1996 all the surviving hybrid seedling bulbs were clonally propagated by chipping (Flint, 1982) and the resultant bulbils grown for two years under the same conditions as the seedlings. Bulbs of the two varieties were also chipped to provide two sets of control bulbs (GH Con and SK Con) of identical age and origin as the hybrids. In September 1998 all the small bulbs were replanted singly in 7cm square pots of compost/grit mixture, some of which also contained chlamydospores of two isolates of Fusarium oxysporum f. sp. narcissi mixed with talc powder as an inert carrier (Price, 1977). The talc inoculum was incorporated at the rate of 1.5 g/l and was equivalent to 4000 spores/g compost. Approximately one third of all bulbs were planted in non-infested compost to serve as controls. A randomised layout of pots within the glasshouse was used to minimise environmental influences. Leaf and flower (where present) morphology of the hybrids was recorded during the spring. In July 1999 all bulbs were removed from their pots and inspected for rots by squeezing the sides and the base of the bulbs. Surviving bulbs were replanted in their original pots and compost. The presence or absence of shoot growth during the spring of 2000 was recorded and absence of a shoot in pots containing infested compost was presumed to be due to bulb death from basal rot during the intervening period. Data from these two assessments are combined in Table 1. These data were also subjected to statistical analysis using Genstat version 5 Release 4.1 to produce regression analyses using accumulated analysis of deviance. Determination of Ploidy Leaf tissue samples from 75 of the hybrid lines and from each of the three parents were collected in March 1998 and ploidy levels were measured by a commercial laboratory using flow cytometry. Screening for Resistance using an in Vitro Assay The six (occasionally five from small bulbs) outermost fleshy white scales were carefully excised from three bulbs of each of five species and eleven varieties of Narcissus. Scales were placed on moistened tissues in covered polythene boxes with the abaxial surfaces uppermost. The abaxial epidermis of each scale was pierced with a sterile needle and 10 µl of water or spore suspension (1 X 10 5 spores.ml -1 ) of a highly virulent isolate of F. oxysporum f. sp. narcissi was placed on the wound site. The areas of any necrotic lesions that developed were measured at 7 and 14 days after the initial wounding. The test was repeated on at least two separate occasions. RESULTS AND DISCUSSION Production, Propagation and Screening for Resistance Some crosses only generated a single seed but in most crosses between two and six seeds were produced. Almost all seeds germinated and produced seedling bulbs that were grown on for another four years. Four chips were cut from each seedling bulb and one or two bulbils formed on each after 12 weeks incubation at 20 0 C. A few of these juvenile bulbs died during the next two years but most grew to between 1 and 2 cm diameter. Bulbs from all the seeds in each pod resulting from pollination by each male pollen parent were bulked to generate large enough sets to allow sufficient replication and permit appropriate statistical analysis. The bulbs from crosses yielding only single seeds are clonal. 256

A total of 49 hybrid lines (19 with GH as seed parent and 30 with SK) and 25 lines arising from self-pollinated NJ were screened for resistance to basal rot. All of the self-pollinated NJ progeny had typical jonquil-type leaves, narrow, dark green and strongly incurved. The leaves of the two varieties were much broader, grey-green and almost flat. All of the hybrids leaves were intermediate in appearance between the varieties and the species. Few flowers formed on any of the plants but those that were seen on the varietyspecies hybrids were identical to the maternal parent i.e. the trumpet flower of the varieties. The rot assessment data are presented in Table 1 and can be summarised as follows: All Crosses: High deviance ratios were noted for infested vs. non-infested compost when numbers of rots in the progeny from the crosses involving either of the two varieties were analysed. Conversely, a very low deviance ratio resulted from a similar analysis of rots in progeny from self-pollinated NJ. This means that the presence of fungal inoculum in the compost significantly influenced the numbers of rotted bulbs found amongst the progeny of NJ crossed with either of the two varieties but did not affect the number of rotted bulbs when NJ was selfed. Some NJ plants were used as pollen donors in crosses with both GH and SK but the statistical analysis did not reveal any correlation between resistance of a hybrid line and the particular male plant used (data not presented). GH x NJ: Three crosses (J022, J014 and J005) generated bulbs that, when challenged, resulted in greater numbers of survivors than would be predicted by chance (the binomial distribution probability function for each of these lines is <0.05). One cross (J002) resulted in significantly fewer survivors than expected. SK x NJ: Six crosses (J088, J092, J095, J101, J105 J110 and J91) generated between 4 and 12 bulbs per cross and none of these bulbs rotted in infested compost (between 2 and 8 bulbs from each were challenged). Bulbs in lines J088 and J095 were clones i.e. derived from a single seedling, and gave consistent results in the test. Probability values for all but two (J105 and J110) of these lines were greater than 0.05 (Table 1b) and so were considered not significant in this test. Low numbers of bulbs tested coupled with low levels of rot in the variety control (37% in SK Con) are possible reasons for high p values in these other lines. Six of the remaining 23 crosses displayed significantly higher levels of rot (>73%) than the control bulbs (J100, J114, J085, J107, J108 and J116). NJ Selfs Although four lines (J51, J74, J58 and J49) displayed significantly more or less rots than the mean value for all lines the remaining 21 lines showed no segregation for resistance. There are at least two possible explanations for this observation. This species may be homozygous for this character. The number of genes involved in resistance of NJ is unknown at present. Alternatively, the test used may be unable to detect variation among this progeny if even the lowest level of expressed resistance is sufficient to resist the challenge of the applied amounts of pathogen. The range of resistance observed in both the progenies where a variety was used as maternal parent could have arisen in different ways. Clearly, even if NJ resistance is single genebased it is not being expressed in these offspring at the same level seen in the NJ self-pollinated progeny. This may be because species resistance is recessive to variety susceptibility or possibly because of gene dosage effects resulting from the combination of gametes from a diploid species and a tetraploid variety (see Determination of ploidy below). Alternatively, it may have resulted from different combinations of resistance genes from two parents with polygenic resistance. It could also simply be the result of assortment of variety resistance genes during meiosis with no contribution from NJ. Earlier work (Bowes et al., 1992) has demonstrated a range of survival rates (0 to 45%) in seedling progenies from variety crosses. These authors concluded that their data indicates that resistance to basal rot in varieties is polygenically determined. The ten hybrid lines that displayed high levels of resistance will be tested with greater replication to verify their resistance status and will be grown to flowering size to assess their 257

commercial potential. Some of the progeny from the variety crosses made earlier (Bowes et al., 1992) will be re-screened for resistance now that the single seedling bulbs with potentially unique genotypes have multiplied naturally. The proportion of surviving lines from several previously non-screened crosses will be compared with those from the variety x NJ crosses and provide additional information about the relative contribution of NJ genes to resistance of these hybrids. Reciprocal crosses between N. pseudonarcissus and N. jonquilla will be made and the F 1 generations examined for rot resistance in order to investigate the inheritance of NJ resistance where both parents are diploid. These crosses should provide information about the heterozygosity of NJ resistance and the recessive or dominant nature of its inheritance. Determination of Ploidy All leaf samples taken from the hybrids had similar nuclear DNA content and this amount most closely resembled the value expected from triploid individuals. This was predicted since the species parent is diploid and both the variety parents are tetraploid. No aneuploidy was detected. Screening for Resistance using an in Vitro Assay The results of the in vitro tests are shown in Table 2. No necrosis was seen in any scales from species or varieties determined as absolutely resistant (Linfield, 1992) with the exception of the 14-day, second occasion result for the variety Larkelly. Highest values for areas of necrosis at both 7 and 14 days after inoculation were recorded for moderately and highly susceptible varieties. Intermediate values resulted from tests of varieties and species described as partially resistant by several authorities (Linfield, 1992, Linfield and Price, 1986 and Tompsett, 1986). Golden Dawn is a Division 8 narcissus variety with N. tazetta ancestry and basal rot is rarely found in tazetta varieties (Linfield, 1992). The reaction of this variety in the scale inoculation test predicts that it would be resistant in whole bulb tests. Sweetness, the other variety of unknown resistance used in the test, gave a test response indicating that its resistance is likely to be partial. These results indicate a positive correlation between area of necrosis developing after inoculation of detached scales and the whole-bulb basal rot resistance of the species or variety. A larger number of genotypes will be tested using both methods and if the accuracy and reliability of the in vitro assay continues to be maintained, it will be used to replace the longer more laborious whole-bulb test. ACKNOWLEDGEMENTS This work was funded by the Ministry of Agriculture, Fisheries and Food, UK. We thank J. Jones and A. Mead for assistance with experimental design and statistical analyses and D. Barbara and D. Pink for helpful discussions. Literature Cited Bowes, S.A., Edmondson, R.N., Linfield, C.A. and Langton, F.A. 1992. Screening immature bulbs of daffodil (Narcissus L.) crosses for resistance to basal rot disease caused by Fusarium oxysporum f. sp. narcissi. Euphytica 63:199-206. Flint, G. 1982. Narcissus propagation using the chipping technique. Annual Review of Kirton Experimental Horticulture Station for 1981:1-9. Kington, S. 1998 In: The International Daffodil Register and Classified List compiled by S. Kington. Royal Horticultural Society, London. Linfield, C.A. 1992. Wild Narcissus species as a source of resistance to Fusarium oxysporum f. sp. narcissi. Annals of Applied Biology 121:175-181. Linfield, C.A. and Price, D. 1986. Screening bulbils, twin scales and seedlings of several cultivars for resistance to Fusarium oxysporum f. sp. narcissi. Acta Hort. 177:71-75. Price, D. 1977. Effects of temperature and inoculum concentration on infection of narcissus bulbs by Fusarium oxysporum f. sp. narcissi. Annals of Applied Biology 86:433-436. Tompsett, A.A. 1986. Narcissus varietal susceptibility to Fusarium oxysporum (basal rot). Acta Hort. 177: 77-83. 258

Wylie, A.P. 1952. A history of the garden narcissi. Heredity 6:137-156. Tables Table 1.a. Progeny bulbs from N. jonquilla x N. jonquilla cross Hybrid line Total number of % rots in infested % rots in clean Significance 4 code 1 bulbs planted 3 compost compost J044 10 0 0 J045 19 0 0 J052 8 0 0 J055 5 0 0 J060 12 0 0 J061 10 0 0 J062 11 0 0 J065 5 0 0 J066 8 0 0 J071 20 0 0 J073 7 0 0 J075 11 0 0 J049 51 0 11.8 L* J047 15 0 20 J067 9 0 33.3 J068 32 4.5 0 J070 23 6.3 0 J048 17 8.3 0 J077 20 7.1 16.7 J050 70 14.9 0 J063 9 16.7 0 J051 13 33.3 50 M* J046 7 40 50 J074 5 50 0 M* J058 4 66.7 0 M* Means 16.0 9.9 7.3 259

Table 1.b. Progeny bulbs from Golden Harvest x N. jonquilla cross Hybrid line Total number of % rots in infested % rots in clean Significance 4 code bulbs planted 3 compost compost J022 5 0.0 0.0 L* J014 12 12.5 0.0 L** J005 18 16.7 0.0 L** J009 6 25.0 0.0 J030 14 40.0 25.0 J008* 2 4 50.0 0.0 J007* 4 50.0 0.0 J031 11 50.0 0.0 J010 16 54.5 0.0 GH Con 5 116 63.4 14.3 J032 8 66.7 0.0 J013 6 75.0 50.0 J006* 7 75.0 33.3 J023 13 77.8 0.0 J001* 8 100.0 0.0 J029 5 100.0 0.0 J016* 4 100.0 0.0 J002 13 100.0 0.0 M* J019* 5 100.0 0.0 J028 6 100.0 50.0 Means 14.1 62.8 8.6 260

Table 1.c. Progeny from St. Keverne x N. jonquilla cross Hybrid line Total number of % rots in infested % rots in clean Significance 4 code 2 bulbs planted 3 compost compost J088* 5 0.0 0.0 J092 6 0.0 0.0 J095* 4 0.0 0.0 J101 9 0.0 0.0 J105 10 0.0 0.0 L* J110 12 0.0 0.0 L* J091 9 0.0 33.3 J111 12 12.5 0.0 J109 14 20.0 25.0 J083 7 20.0 0.0 J112 18 33.3 0.0 J103 20 35.7 16.7 SK Con 5 116 36.6 18.4 J106 7 40.0 0.0 J087 5 50.0 0.0 J117 21 50.0 0.0 J097* 4 50.0 0.0 J115 10 57.1 0.0 J113 17 58.3 0.0 J102 7 60.0 0.0 J098 8 66.7 0.0 J080 10 71.4 33.3 J100 16 72.7 0.0 M* J114 20 78.6 0.0 M** J085 8 83.3 0.0 M* J107 12 87.5 0.0 M** J108 11 87.5 66.7 M** J089* 4 100.0 0.0 J099* 4 100.0 0.0 J116 19 100.0 33.3 M*** J104* 4 100.0 50.0 Means 13.8 47.5 8.9 1 six different N. jonquilla plants were used as pollen donors. 2 * all the bulbs in each of these lines are clones i.e. from a single seed; all other bulbs from other lines in Tables 1b and 1c have a unique male (N. jonquilla) parent but originate from several seeds within a single pod. 3 the total number of bulbs from each cross were divided into 2 unequal batches. Between 50 and 80% of bulbs from each cross were planted in infested compost; the remainder into clean compost. 4 *, ** and *** indicate probabilities of <0.05, <0.01 and <0.001 respectively. L and M refer to less and more than the expected number of rotted bulbs respectively. Values for individual hybrid lines in Table 1a are compared with the mean value (9.9) but in Tables 1b and 1c are compared with the % rots in infested compost value for the control bulbs (GHCon and SKCon) 5 sets of control bulbs were clonally produced by chipping (Flint, 1982) each variety: GHCon and SKCon. 261

Table 2 Response of detached scales to wounding and inoculation with Fusarium oxysporum f. sp. narcissi 2 Species or variety Resistance/susceptibility a Necrotic area (mm 2 ) after 7 days b Necrotic area (mm ) after 14 days b N. broussonetti Absolutely resistant 0 0 N. canaliculatus Absolutely resistant 0 0 N. jonquilla Absolutely resistant 0 0 Larkelly Absolutely resistant 0 * Thalia Absolutely resistant 0 0 N. lobularis Partially resistant 4 84 St. Keverne Partially resistant c 18 128 N. obvallaris Partially resistant 20 118 Ice Follies Partially resistant d 41 112 Scarlett O Hara Moderately susceptible d 44 125 Dutch Master Moderately susceptible d 48 179 Carlton Highly susceptible c 90 231 Golden Harvest Highly susceptible c 104 254 Golden Dawn unknown 0 0 Sweetness unknown 10 115 Areas of necrotic tissue were measured 7 and 14 days after the abaxial epidermis of each scale was pierced with a sterile needle and 10 µl of water or spore suspension (1 X 10 5 spores.ml -1 ) of a highly virulent isolate of Fusarium oxysporum f. sp. narcissi was placed on the wound site. a Comparative estimates of basal rot resistance were taken from Linfield (1992) except as c and d below. b Values presented are the means for 6 scales from three individual bulbs determined on two or more occasions (* - a value of 0 was recorded the first time that this variety was tested but on the second occasion the necrotic area after 14 days was 180mm 2 ). c Basal rot resistance characteristics of these varieties were taken from Linfield and Price (1986). d Basal rot resistance characteristics of these varieties were taken from Tompsett (1986) 262