This study comprehensively investigated host plants utilised by B. tabaci species especially for growth and development in order to understand the role that alternative host plants play in their population dynamics.
Host-plants utilised by cassava-colonising B. tabaci
The findings indicate that in Uganda, cassava is only colonised by cassava B. tabaci, SSA1 (-SG1 and -SG2), and SSA2, which was consistent with previous studies on cassava-colonising B. tabaci (Legg et al., 2002; Sseruwagi et al., 2005, 2006; Legg et al., 2014). The very closely related SSA1-Hoslundia (a sub-group of SSA1 at mtCOI level), however, was not detected on cassava but colonised A. africana, M. glaziovii, H. opposita, A. conyzoides, S. acuta and I. batatas. A. africana, H. opposita, A. conyzoides and S. acuta are wild hosts and are considered to be medicinal, except S. acuta. The medicinal plants are often found on borders that have been left by farmers to fallow or hedge barriers in cases where there no nearby shrubs or forests. Medicinal plants are of great importance to women who pick, process and sell them. A. africana is used to stop bleeding and to treat wounds (Komakech et al., 2019). H. opposita treats skin blisters, worms, diarrhoea and yellow fever (Namukobe et al., 2011). It is also used as a stomach cleanser and to treat postnatal pain and wounds. A. conyzoides is used to treat wounds and provides strength to pregnant women (Namukobe et al., 2011; Tugume et al., 2016). M. glaziovii was introduced into East Africa in the 19th century as a source of rubber (Munro, 1983) but is currently used as an ornamental and hedge-barrier plant in Uganda. I. batatas is an important food crop in Uganda (Mwanga & Ssemakula, 2011) that was introduced in East Africa in the 16th century (Grigg, 1974). It is commonly intercropped with cassava by small scale farmers. It was hypothesized that the inability of some B. tabaci to colonise cassava could be due to the poisonous hydrogen cyanide (Legg, 1996).
Hydrogen cyanide (HCN) is stored in an intact form as cyanogenic glucosides in the cell vacuoles of cassava but tissue damage such as insect feeding exposes cyanogenic glucosides to β-glucosidases, leading to the formation of HCN (Calatayud et al., 1994). However, SSA1-Hoslundia colonised M. glaziovii, a wild relative of cassava which also contains cyanogenic glucosides (Joseph et al., 2000), so this is unlikely to be the reason it did not colonise cassava. Studies have shown that wild relatives of cassava contain a higher cyanogenic content than cassava (Wang et al., 2014). SSA1-Hoslundia does not appear to have evolved the ability to colonise cassava. This may be due to the high sugar levels in the cassava phloem-sap (Li et al., 2016; Yan et al., 2019), which may create a high osmotic potential that is unsuitable for SSA1-Hoslundia. Transcriptome studies revealed that a cultivated cassava variety (KU50) from South East-Asia had a significantly higher number of genes involved in sucrose transport than in a wild sub-species of cassava (W14) (Wang et al., 2014).
In addition to cassava, SSA1-SG1 colonised 11 plant species from seven families and these included: A. africana, E. heterophylla, P. niruri, M. glaziovii, P. urens, L. camara, H. opposita, O. gratissimum, B. pilosa, V. amygdalina and I. batatas while, SSA1-SG2 colonised E. heterophylla and M. glaziovii. These results complement findings by Sseruwagi et al. (2006) who reported that SSA1 colonised five alternative hosts, M. glaziovii, J. gossypifolia, E. heterophylla, A. africana, and A. esculentus. However, SSA1 species were not detected on A. esculentus in our study area. We also did not identify J. gossypifolia in the area where the study was carried out. MEAM1 and MED-ASL (recently called Uganda ‘‘ASL’’) developed on A. esculentus in the field (Table 3). MEAM1 has been reported to displace indigenous species in the Americas, Australia and China, particularly in areas where insecticides are used, due to its resistance to several groups of insecticides (Costa and Brown, 1991; De Barro et al., 2006; Zang et al., 2006; Hu et al., 2011). Insecticide usage in our study area was limited and thus it could be that MEAM1 and MED-ASL are more specialised in utilising A. esculentus than SSA1. It has been hypothesized that insect herbivores may shift from one host-plant to another as result of competition (Gassmann et al., 2006). Therefore, SSA1-SG1 could have shifted from A. esculentus due to competition from MEAM1 and MED-ASL.
The alternative host-plants that were colonised by SSA1-SG1 were mainly wild hosts, of which E. heterophylla and P. urens are common weeds. L. camara, A. africana, P. niruri, H. opposita, O. gratissimum, B. pilosa and V. amygdalina are medicinal plants (Namukobe et al., 2011). Most of them are commonly found in the fallow borders, hedge-barriers, shrubs, or forests except B. pilosa and Phyllanthus niruri, which are common weeds in the gardens. L. camara treats ring worms and tuberculosis (Kirimuhuzya et al., 2009; Tugume et al., 2016). B. pilosa stops bleeding and treats wounds. P. niruri is used to treat measles, wounds, fever and stomach problems (Tugume et al., 2016). O. gratissimum is a remedy for stomach pain (Tugume et al., 2016). V. amygdalina cures malaria, stomach-ache and convulsions (Namukobe et al., 2011; Tugume et al., 2016). Although SSA1-SG1 was polyphagous, it did not colonise A. conyzoides and S. acuta, which were colonised by SSA1-Hoslundia. These clear differences in host-plant utilisation between the two populations suggest that these could be separate species. This however, requires further investigation through reciprocal crosses.
SSA1-SG1 was more abundant on cassava than alternative host-plants commonly found in proximity to cassava, suggesting that it has a preference for cassava. Omondi et al. (2005) examined the host preference of cassava whiteflies in Ghana and revealed that they significantly chose cassava for oviposition and development than A. esculentus, S. melongena, S. lycopersicum, S. aethiopicum and V. unguiculata. In addition, SSA1-SG1 colonised 11 alternative host-plants found in close proximity to cassava compared to > 300 m from cassava, where it colonised only six hosts. This implies that the high populations on cassava spread to nearby hosts including poor hosts such as L. camara (Namuddu et al., Unpublished data), whereas in a distance of > 300 m, SSA1-SG1 selects the most suitable hosts.
Cassava was introduced in East Africa in 18th century (Guthrie, 1987). Recent studies reported that Africa is the centre of origin of B. tabaci (Mugerwa et al., 2018). In this study, 75% (9/12) of the host-plants colonised by SSA1-SG1 were wild hosts, implying that these might be the native plants on which this species evolved. The strong preference for cassava may be the result of much more recent evolutionary change due to the intensive and continuous cultivation of cassava. Furthermore, six and one of the host-plants colonised by SSA1-SG1 and SSA1-SG2, respectively, were found at least 300 m away from cassava, showing that these populations are capable of surviving on other plant species. These alternative hosts allow these whitefly species to survive in the absence of cassava and colonisation, whenever it is planted. The high whitefly populations on cassava in Uganda could be due to the wide host-plant range of SSA1-SG1. Studies have shown that the presence of alternative host-plants provides a continuous breeding habitat for insect pests leading to increased populations (Joyce, 1983; Sétamou et al., 2000; Johannesen & Riedle-Bauer, 2014). Sétamou et al. (2000) reported that the abundant wild hosts for a snout moth, Mussidia nigrivenella in Benin were responsible for its high numbers on maize during the cropping season. Similarly, Johannesen & Riedle-Bauer (2014) observed that a new host-plant, Urtica dioica led to increased populations of a plant-hopper, Hyalesthes obsoletus on grapes in Austria.
The ability of SSA1-SG1 to utilise several alternative hosts increases the risk of transmitting viruses between these hosts and cassava. Findings by Amisse et al. (2019) showed that Cassava Brown Streak Virus (CBSV) infected Zanha africana and Trichodesma zeylanicum that are commonly found near cassava fields in Mozambique. They also detected CBSV and Ugandan CBSV in Manihot carthaginensis subsp. glaziovii. In India, Rajinimala et al. (2009) identified the bitter gourd yellow mosaic geminivirus in cassava, A. indica, C. sparsiflorus and M. coromandelianum. The availability of various host-plants for SSA1-SG1 may result in the transmission of unknown viruses to cassava that could be epidemiologically important. These host-plants may also serve as reservoirs for cassava viruses.
Host-plants colonised by non-cassava B. tabaci
This study found non-cassava B. tabaci species that colonised only non-cultivated plants and those that developed on both cultivated and non-cultivated plants. IO, SSA6, SSA10, SSA12 and SSA13 colonised only non-cultivated plant species. IO colonised P. urens, S. acuta and C. benghalensis. Contrary to our findings, IO (Ug7) was previously reported as a polyphagous species that colonised cultivated plant species namely, P. vulgaris and G. hirsutum and a weed species, C. benghalensis in Uganda (Sseruwagi et al., 2005). SSA6, SSA10 and SSA12 colonised only O. gratissimum, V. amygdalina and C. benghalensis, respectively. This study is in agreement with the findings of Sseruwagi et al. (2005) that SSA6 (Ug3) colonised only O. gratissimum in Uganda. EA1 colonised H. opposita, P. urens, L. camara, B. pilosa and a cultivated plant species, P. vulgaris. EA1 was first identified from a single adult in Tanzania (Legg et al., 2014), and here we report that it is polyphagous, colonising both cultivated and non-cultivated plants. However, its role as a vector requires further investigation. MEAM1 utilised P. vulgaris, A. esculentus, S. lycopersicum and a weed species, S. acuta, while MED-ASL colonised non-cultivated plants including H. opposita, P. urens, S. acuta, A. africana and cultivated plants such as A. esculentus, V. unguiculata, I. batatas and G. hirsutum. MEAM1 predominantly colonised cultivated plants, probably due to competition from the indigenous B. tabaci. Our findings supplement previous studies in Uganda, which reported that MED-ASL (Ug4) was polyphagous and colonised Cucurbita sativus, C. pepo, L. nepetifolia and P. urens, whereas MEAM1 (Ug6) was oligophagous and developed on A. esculentus and C. benghalensis (Sseruwagi et al., 2005). Vyskočilová et al. (2019) also observed that MED-ASL (Uganda ‘‘ASL’’) developed well on G. hirsutum, A. esculentus, I. batatas, S. melongena and C. pepo.
Development of B. tabaci species on cassava
Our findings showed that EA1, MEAM1, MED-ASL, SSA1-Hoslundia, SSA6 and SSA12 cannot develop on cassava. Adults of these cryptic species of B. tabaci did not survive on cassava longer than three days. The immatures did not develop beyond first-instar stage and the eggs that did not hatch were shrivelled or dehydrated. In a host transfer experiment, Legg (1996) observed that B. tabaci adults collected from G. hirsutum and I. batatas in the field (mtCO1 species identity was not determined) and established on these host-plants, did not develop on cassava and none of the adults lived longer than two days. The first-instar nymphs also died soon after hatching. The shrivelled and dehydrated eggs were probably laid by dying females that could not make a slit in the leaf to insert the pedicel, which absorbs water and solutes from the plant for embryo development and eclosion (Buckner et al., 2002). Mortality of adults and first-instars was probably as result of failure to feed successfully due to the inability to penetrate the leaf cuticle or presence of toxic compounds in cassava phloem-sap. Feeding studies conducted by Milenovic et al. (2019) using an electrical penetration graph (EPG) system showed that feeding by non-cassava B. tabaci MED on cassava was characterised by short total phloem ingestion periods (< 1 h) compared to that of the cassava B. tabaci SSA1-SG3, which was > 4 h. There was longer total xylem feeding for MED, suggesting that the whiteflies were not ingesting enough phloem-sap and thus became dehydrated. Douglas (2006) reported that phloem-feeding insects encounter a high osmotic pressure created by sucrose in the phloem-sap, which can lead to dehydration. The high sucrose levels in cassava (Wang et al., 2014) probably led to dehydration of the non-cassava B. tabaci adults and first-instar nymphs. In addition, cassava is a well-defended plant species that contains cyanogenic glucosides (McMahon et al., 1995) and flavonoids (Prawat et al., 1995) and possibly suitable for only specialised whitefly species.
Some adults of EA1, MEAM1, MED-ASL, SSA1-Hoslundia, SSA6 and SSA12 survived on cassava for at least one day, which increases the possibility of transmission of non-persistent plant viruses that require only insect probing (Polston et al., 2014). When searching for suitable hosts, phloem-feeders make shallow probes to detect the carbohydrate content and presence or absence of secondary metabolites (Walling, 2008). Milenovic et al. (2019) observed that during a 12 h EPG recording period, MED whiteflies were able to ingest cassava phloem-sap for 41 min.
Implications for control of B. tabaci and whitefly-vectored viruses
The polyphagy of SSA1-SG1 and other B. tabaci like MED-ASL in this study contributes to the high populations on cassava and other crops, respectively. Polyphagy makes the survival of polyphagous B. tabaci more flexible and their control difficult.
García and Zerbini (2019) proposed that uncultivated wild hosts were the original hosts for majority of the causal viruses from which they were transferred to crops by polyphagous whitefly vector populations. The ability of some whitefly species (SSA1-SG1, MED-ASL and B. Uganda1) to colonise both crops and weeds could potentially increase the transmission of viruses between plants resulting in mixed infections. Recombinant begomoviruses have been detected in cassava (Zhou et al. 1997; Berrie et al. 2001) and the viruses often originate from other hosts, suggesting the ability of polyphagous vector populations to facilitate formation of novel viruses.