The response of resistant kiwifruit (Actinidia chinensis) to armoured scale insect (Diaspididae) feeding
- First Online:
- Cite this article as:
- Hill, M.G., Mauchline, N.A., Jones, M.K. et al. Arthropod-Plant Interactions (2011) 5: 149. doi:10.1007/s11829-011-9124-9
- 147 Views
The responses of five experimental genotypes and one commercial variety of kiwifruit (Actinidia chinensis) to attack by two polyphagous, congeneric armoured scale insect pests (Hemiberlesia rapax and H. lataniae) are described. H. lataniae feeding elicits a response in the bark and fruit of all but one of the experimental genotypes, leading to the development of wound periderm over a 4–5 week period, and death of the insect. The response, which differs slightly between tissue types and genotypes, consists of wound periderm formation in a bowl shape beneath and around the insect, preventing its stylet from reaching normal unmodified parenchyma tissue. Wound periderm cell walls become suberised and cells beneath the insect become filled with phenolic compounds. In some cases, cells beneath the insect become hypertrophic or undergo lysis, exhibiting characteristics of a hypersensitive-like response. The remaining genotype showed no physical change in tissue structure in response to H. lataniae feeding, and the insects survived but were substantially reduced in size. These results suggest that both physical and chemical plant resistance responses are involved. In contrast, H. rapax elicited no observable histological response from any of the genotypes and the insects developed normally on bark and fruit. Both insect species developed normally on leaf petioles and these exhibit only slight cell wall thickening in response to their feeding. This unusual plant defensive response to a sucking insect has similarities to simple types of gall formation in response to insect and pathogen attack and has characteristics of resistance gene-mediated models of plant defence.
KeywordsPlant resistanceWound peridermHypersensitive responseDiaspididaeArmoured scale insectsGalls
In spite of their economic importance as pests of perennial horticultural crops, the interactions between armoured scale insects (Homoptera: Diaspididae) and their hosts have received very little study (Miller and Davidson 2005). By contrast, at the next level of trophic interaction, armoured scale insect host-parasitoid interactions and biological control have been the subjects of a large body of basic and applied research (Rosen 1990; Miller and Davidson 2005). The spectacular success of classical and augmentative biological control in regulating populations of several key diaspidid pest species on horticultural crops in the middle and late decades of the twentieth century has led to many being successfully controlled without recourse to studies of host plant resistance (Rosen 1990).
This study of the interaction between armoured scale insects and their hosts arises from the need for economic control of two armoured scale insect pests, greedy scale (Hemiberlesia rapax Comstock) and latania scale (Hemiberlesia lataniae Signoret) on kiwifruit in New Zealand, and the absence of effective biological control options. These insect species are both highly polyphagous and occur throughout the world in tropical to temperate zones (Miller and Davidson 2005; Normark and Johnson 2010; CABI 2009). Both species exist in New Zealand as uniparental, parthenogenetic populations, having two generations per year and are resident on the wood of kiwifruit vines throughout the year, settling on leaves and fruit during the summer (Edwards et al. 2008).
A research programme to investigate kiwifruit resistance to armoured scale insects has found large differences in the levels of armoured scale insect infestation on canes and fruit of germplasm from the species Actinidia chinensis Planch. with some genotypes showing complete resistance to H. lataniae, but not to H. rapax (Hill et al. 2009). During the course of these studies, swelling was noticed around the insects on the canes of some resistant genotypes, followed by death of the insects. This study investigates this unusual response of several A. chinensis genotypes to feeding by H. lataniae and H. rapax. The response of the cane bark, fruit and petiole to feeding by H. lataniae and H. rapax is described. We suggest that this is an example of a gene-for-gene mediated plant resistance response. We hypothesise, with reference to the insect gall-formation literature, that the histological response of the A. chinensis varieties to H. lataniae is characteristic of a simple insect gall and may represent a precursor to the evolution of true galls, a hypothesis proposed 80 years ago by Zweigelt (1931). An understanding of the interaction between these pest species and kiwifruit germplasm will assist the breeding of future commercial varieties with enhanced levels of resistance.
The plant material consisted of canes, leaves and fruit on 10-year-old vines of the commercial kiwifruit variety A. chinensis ‘Hort16A’ and 12-year-old vines from four experimental A. chinensis genotypes (2C, 3C, 8A, 9A) from a breeding experiment (details in Cheng et al. (2004)). All vines were diploid. Previous research (Hill et al. 2007) had shown that these genotypes exhibited resistance to H. lataniae but not to H. rapax. All experiments were carried out on vines growing in orchard blocks in the field, at the Plant & Food Research Te Puke Research Centre, New Zealand.
Insect settlement assay
Twenty Hemiberlesia lataniae and H. rapax crawlers (the newly eclosed, mobile stage of the first instar) were settled separately using a fine brush, onto five leaves of four vines of each of the five genotypes on 23–24 November 2006. Twenty crawlers of each species were also settled in the same manner onto sites on new season (<1 year old (yo)) and previous season (>1 yo) canes on five vines of each genotype on 10 and 11 January (H. lataniae) and 15 and 17 January 2007 (H. rapax). Wool was wrapped around the cane to aid crawler settlement. The wool was removed on 18 January 2007. Crawlers of both scale insect species were settled onto fruit in the same manner on 8–10 March 2007. Observations of scale insect settlement and plant response began as soon as the insects were settled on the leaves or fruit and as soon as the wool was removed from the bark. The insects were photographed using a Canon Coolpix 995 camera with a frame attached to the front allowing the size of the insect to be measured in the image (Hill et al. 2005). Observations of the scale insect nymphs and the plant tissue around them were made three times a week for 5–7 weeks until no further changes were observed. Any physical response by the plant in the form of blackening or swelling of tissue around or beneath the scale was recorded.
Tissue samples for histology, including the insect, were taken on days 14, 21, 28 and 35 after crawler settlement for cane samples, day 41 for leaf (petiole) and days 42 and 49 for fruit samples in 2007/2008. A second series of cane (<1 yo) and fruit samples with settled H. lataniae nymphs was collected at weekly intervals for 5 weeks after crawler settlement from genotypes ‘Hort16A’ and 2C in November/December 2007 (cane samples) and March/April 2008 (fruit) for microscopic examination. Finally bark samples were taken 5 weeks after H. lataniae crawler settlement on >1 yo canes of a partially resistant experimental A. chinensis variety ‘Hort22D’ in January 2010 to allow comparison of the histological response between completely and partially resistant germplasm.
Pieces of tissue (fruit, canes and leaves) 5mm × 5 mm square were fixed in 2% (v/v) formaldehyde with 2.5% (v/v) glutaraldehyde in 0.1 M phosphate buffer pH 7.2 under vacuum for 1 h, washed in buffer, dehydrated with an alcohol series, and embedded in LR White resin (London Resin Company Ltd, Reading, England) (Sutherland et al. 2004). Embedded material was sectioned with a diamond knife on a Leica UCT ultramicrotome at a thickness of 1 μm and stained with 0.05% (w/v) toluidine blue in borate buffer (pH 4.4). Light microscopy was carried out on an Olympus Vanox AHT3 compound microscope (Olympus, Tokyo) with image recording by a CoolSnap digital camera (Roper Scientific, Tucson, AZ). Forty-two insects and their attached substrate were sectioned, stained and mounted (details in Online resource 1) and over 500 slides were prepared and examined.
To visualise autofluorescence of suberin/lignin and condensed phenolics, the UV filter (excitation 330–385 nm, emission 420 nm) and the G (green) filter (excitation 545 nm, emission 590 nm) were used respectively on 1-μm thick sections (Ruzin 1999). In our experience glutaraldehyde-induced autofluorescence is minimal at this thickness. For immunolabelling with JIM5 (anti-homogalacturonan; unesterified or partially esterified) (Clausen et al. 2003), 200-nm thick sections were mounted on poly-l-lysine coated slides and dried overnight. They were then blocked with 0.1% bovine serum albumin (BSA-c, Aurion, Wageningen, The Netherlands) in PBS-T (phosphate-buffered saline plus 0.1% Tween® 20) for 15 min and then incubated overnight in JIM5 diluted 1:20 in blocking buffer. After incubation the slides were washed in PBS-T and incubated in goat anti-rat Alexa 488 diluted 1:600 in PBS. Sections were finally washed in 2–3 ml PBS-T and mounted in Citifluor AF1 antifadent solution (Citifluor, Leicester, UK). For details of staining techniques for Actinidia tissue and a description of Actinidia fruit skin histology, see Hallett and Sutherland (2005).
Measuring scale insect size
The size of the scale insect cap on a minimum of 10 scale insects per substrate was estimated using the image analysis software ImageJ (version 1.37v; http://rbs.info.nih.gov/ij/) from the photographic images taken in the field (Hill et al. 2005) at a range of dates up to 6 weeks after crawler settlement.
Armoured scale insect settlement, survival and growth
Total number of armoured scale insects on fruit 8–11 days after crawler settlement and the number of scale insects and proportion that were alive on the fruit 43 days after crawler settlement
No. scale settling after 8–11 days
Survival to day 43
Total scale on fruit
% Live scale
H. lataniae (%)
H. rapax (%)
The distribution of life stages of live and dead H. lataniae and H. rapax after 43 days on fruit from all vines
Externally visible response of the plant to scale insect settlement and feeding
Hemiberlesia rapax on the fruit elicited no externally visible response from any of the vines. H. lataniae elicited no externally visible response from the fruit of any of the genotypes except for one vine of genotype2C which showed cell death (hypersensitive-like reaction) of the tissue around the scale 6 weeks after scale insect settlement.
Crawlers of both species settled predominantly (95%) on the upper surface of the petiole at its junction with the leaf lamina (the remainder settling on the adaxial leaf lamina). After 6 weeks, the aggregations of both H. rapax and H. lataniae elicited a slight darkening of the petiole cells on genotype 2C only. No other visible leaf response to either species was observed.
Stylet penetration and feeding
Histology of the plant response to armoured scale insect feeding
One week after H. lataniae settlement on ‘Hort16A’ cane bark, there was no evidence of a change in the plant cell structure (Fig. 4d). Two weeks after scale insect settlement (Fig. 5a), cell division had been initiated in collenchyma and parenchyma cells beneath and around the insect extending up to the periderm. The periderm layer beneath the insect (dislodged during preparation—its position is indicated by an arrow) also appeared to be undergoing an enlargement in cell size, becoming more like collenchyma cells in appearance. Three weeks (Fig. 5b) after scale insect settlement, an organised layer of cells characteristic of wound periderm (Evert 2006) had formed in a bowl-shape, encircling the insect. The wound periderm layer, about 8–12 cells thick, extended from the original (‘native’) periderm, beneath and at some distance from the insect, down through the parenchyma as far as the layer of fibre cells. Four weeks after scale insect settlement (Fig. 5c), the new cells making up the wound periderm (Wp) formed a continuous and morphologically indistinguishable layer with the original native periderm (Np) layer.The periderm layer beneath the insect had become indistinguishable from collenchyma, although the epidermis remained unmodified. Some of the cells in the collenchyma layer above the wound periderm were full of dark-staining phenolic material, while other cells had become hypertrophic or were showing signs of lysis and collapse. After 5 weeks (Fig. 5d), the bowl of wound periderm had undergone further cell division, forming a layer of cells up to 20 cells thick, completely encircling the scale insect, which was dead. Dark-staining phenolic material filled many of the cells between the wound periderm and the insect, and most of the remaining cells showed signs of lysis or collapse.
As far as we are aware, this is the first detailed description of a plant defensive response to a diaspidid scale, and the first record of wound periderm formation and hypersensitive–like reaction as a defensive response to diaspidid feeding. The response of cane and fruit tissue of the ‘resistant’ A. chinensis selections to H. lataniae settlement and feeding shows a consistent pattern of wound periderm formation in a bowl of suberised/lignified cells filled with phenolic material, effectively walling-off the insect from unmodified parenchyma tissue on which it can feed. There was some variability in the degree of cell hypertrophy and lysis (hypersensitive-like cell death) within the ‘bowl’ which may have been related to tissue age or genotype. The fruit tissue response was more variable than the cane bark tissue, and this may reflect physiological changes taking place in the rapidly-growing fruit. We did not set out to characterise fully the effect of tissue age on its response to insect feeding and the limited number of insects sectioned in this study means that we may not have observed the full range of plant responses to H. lataniae. The mode of insect death cannot be determined with certainty from the observations, however there is evidence to suggest that both physical (lignin/suberin deposition, cell lysis and collapse (e.g. Figs. 3, 6e, and 7a)) and chemical (phenolic deposition (e.g. Figs. 5d, 6c and 11b)) defensive mechanisms are involved. The observation of substantially reduced insect growth on the partially resistant genotype ‘Hort22D’ without a visible physical response from the plant (Fig. 9) would also suggest in this instance a predominantly chemical basis to the plant response.
The early initiation (within 2 weeks of H. lataniae settlement) of wound periderm formation through parenchyma cell division (Fig. 5a) in the absence of any other visible plant response or tissue change suggests that this is a primary response from the plant to the detection of the insect’s stylet. The later development (within 4–6 weeks) of a hypersensitive-like response in tissues between the wound periderm and the insect might indicate that this is a secondary response from the plant, possibly resulting from the wound periderm formation. However, there is ample evidence from studies of hypersensitive-like responses to fungal and bacterial invasion in plants to suggest that hypersensitivity is a major defensive response in its own right, mediated in many cases by specific resistance genes (Stuible and Kombrink 2004; Greenberg and Yao 2004). Thus, these two responses—hypersensitivity and wound periderm production—are most likely to be direct and distinct plant responses to insect attack. We hypothesise that both may be mediated by resistance genes in the manner of the gene-for-gene model of plant defence (Kaloshian 2004).
Wound periderm formation, cell hypertrophy and lysis (hypersensitive-like response), cell-wall lignin/suberin deposition and within-cell phenolic deposition, have been observed in previous studies of conifers, fruit and nut trees in response to invasion by fungi (Biggs 1989; Solla et al. 2002; Hoff and McDonald 1972), nematodes (Yamada and Ito 1993; Kaplan and O’Bannon 1981), a bark beetle (Franceschi et al. 2000) and a margarodid (Liphschitz and Mendel 1989). In most cases, these studies have noted a differential response between ‘resistant’ and ‘susceptible’ varieties, and the speed of wound periderm formation appears to be related to the degree of resistance in pathogen attacks. Most wound periderm responses occurred over a 3–5 week time frame, similar to those observed here. Also similar to the finding in this study that petiole, fruit and cane bark tissues mount different responses to H.lataniae feeding, Liphschitz and Mendel (1989) noted that wound periderm and hypersensitive-like responses occurred in the stem but not in the shoot cortex of resistant Pinus spp. challenged by Matsucoccus josephi Bodenheimer and Harpaz. The well known resistance of north American native Vitis rootstock to phylloxera (Daktulosphaira vitifoliae Fitch) in grapes is, as far as we are aware, the only other well documented example of hypersensitive-like response to feeding in the Coccoidea (Roush et al. 2007; Blank et al. 2009; Dietrich et al. 2009). The resistance of specific parts of the plant to H. lataniae fruit and bark, but not leaves, may not be unusual, as many diaspidids are known to feed on only one plant part (Beardsley and Gonzales 1975; Miller and Davidson 2005).
While the speed of wound periderm formation may be important in determining the degree of resistance in pathogen-tree interactions (Solla et al. 2002), for a slow-developing sessile insect like H. lataniae (2 generations per year, 8–12 weeks from eclosion to the onset of reproduction on kiwifruit in New Zealand (Hill et al. 2009)), wound periderm formation and a hypersensitive reaction would appear to be an ideal plant defensive response from which, unlike a pathogen, the insect cannot escape. Although this may be the first time that such a response has been noted against a diaspidid, there are over 30 published papers on host plant resistance in at least 16 economic plant species to 19 diaspidid species (MG Hill unpublished). In only one other instance, the tea scale (Pseudaulacaspis pentagona) (Mizuta 2003, 2005; Tanaka and Taniguchi 2007), has the underlying defence mechanism been studied in detail and a resistance gene located. Relative to other economically important homopteran families (e.g. Aphididae, Aleurodidae, Cicadellidae), host plant resistance to the Diaspididae (indeed to the Coccoidea) has received scant attention. The wound periderm defence responses and hypersensitive-type reactions in A. chinensis discovered in this study would suggest that closer examination of insect neonate -host interactions may provide a fruitful avenue for investigating plant defensive responses to Coccoidea. We suggest that the plant response observed here may be relatively common as a defence against the Diaspididae, and possibly more broadly, the Coccoidea, but because it acts upon first instars and can confer complete resistance to the insect species, may have been overlooked.
Actinidia response to scale insect feeding and its similarity to gall formation
Thirty-nine diaspidid species out of 2,370 (1.6%) have been recorded as gall-forming, with at least 12 showing stem or twig swellings (Gullan et al. 2005). The superficial similarity of the plant defensive responses observed in this study to those involved in gall formation by diaspidid (Larew 1990), asterolecaniid (Gullan et al. 2005) and cynipid (Harper et al. 2004) gall formers and to a gall-forming maize smut (Callow and Ling 1973), namely hypertrophy, hyperplasia, dense-staining cytoplasm, cell collapse and lignification, supports an old hypothesis that gall formation is a plant defensive reaction against herbivory that has come under the control and manipulation of the insect (Hilker and Meiners 2006; Zweigelt 1931).
A hypersensitive reaction to gall-forming insects is widespread across many plant families (Fernandes 1990; Fernandes and Negreiros 2001; Cornelissen et al. 2005). Gall formation can be host-dependent, tissue-dependent and within some genera, scale insect species-dependent (Larew 1990; Gullan et al. 2005), indicative of an insect-host interaction that is similar in principle to a gene-for-gene mediated plant defensive response.
Stylet penetration was intra-cellular and feeding appeared to occur primarily from the parenchyma. Diaspidid feeding has not received much study, but intra-cellular stylet penetration terminating in parenchyma has been observed previously in Pseudaulacaspis pentagona (Targ. Tozz.) on mulberry bark (Yasuda 1979) and by Aonidiella aurantii (Maskell) on citrus leaves and bark (Washington and Walker 1990). Stylet penetration into vascular tissues has also been observed in at least two instances, although it is probably not a main source of food (Heriot 1934; Washington and Walker 1990). This phenomenon was not observed in the current study, where the fibre layer of sclerotised cells beneath the parenchyma is likely to act as a barrier to stylet penetration into the vascular bundle.
The levels of host plant resistance observed in this study indicate that the development of future commercial kiwifruit varieties with enhanced levels of diaspidid scale insect resistance is a realistic prospect. Further work will seek to identify the genetic basis of kiwifruit resistance to armoured scale insects, and to understand the nature of the wound periderm response to the highly polyphagous H. lataniae but not to the congeneric and equally polyphagous H. rapax.
We are indebted to Margaret McCully, CSIRO, for advice on interpretation of the histological responses observed. Erik Rikkerink, Graham Walker and an anonymous reviewer provided comments on earlier drafts. The New Zealand Foundation for Research Science and Technology (C06X0301) provided funding support.