The greenhouse whitefly, Trialeurodes vaporariorum Westwood (Hemiptera: Aleyrodidae) is a globally serious pest in vegetable and ornamental crop production systems, attacking crops within the families Crucifareae, Legumimoseae, Malvaceae, Solanaceae, and Asteraceae (Kim et al., 2014; Wang et al., 2017). The damage from T. vaporariorum is caused by both the adults and the nymphs through phloem-feeding, resulting in the loss of nutrients and subsequent low plant productivity (Arnó i Pujol et al., 2009; Gao et al., 2017). Both the adults and nymphs have a sap-sucking feeding behaviour and use a specialised stylet that passes through the epidermis and mesophyll layers allowing them to feed exclusively from the phloem bundle (Wang et al., 2017). During feeding, whiteflies also excrete honeydew, which is a substrate for sooty mould (Capnodium sp.). The sooty mould reduces the photosynthetic capacity which directly affects the growth and productivity of the plant. It also reduces the aesthetic value of the crop, and this is important in crops like ornamentals, leafy vegetables, and fruits because their marketability depends on their appearance. In addition to their direct infestation to the crops, they also cause indirect important economic damage through their transmission of several plant viruses. T. vaporariorum is a vector of several criniviruses such as Tomato chlorosis virus and Tomato infectious chlorosis virus, and torradoviruses like Tomato torrado virus, although some of these viruses can also be transmitted by other whitefly species, such as T. abutiloneus (Haldeman) and Bemisia tabaci (Gennadius) (Hemiptera: Aleyrodidae) (Navas-Castillo et al., 2011). Losses from T. vaporariorum transmitted plant viruses depend on the nature of the virus, the crop, and its developmental stage, as well as the disease incidence (Lapidot et al., 2014). For example, losses resulting from Lettuce infectious yellows virus resulted in a yield reduction ranging from 50 to 75% in lettuce and 20 to 30% in sugar beets (Wisler et al., 1998), while strawberry yield losses ranged between 15 and 20% caused by Strawberry pallidosis associated virus (Wintermantel, 2004). In addition, losses due to Tomato infectious chlorosis virus amounted to US$ 2 million in tomatoes (Wisler et al., 1998; Wintermantel, 2004).

The proliferation of greenhouses in Kenya has provided optimum climatic conditions for the high reproduction rates and proliferation of T. vaporariorum (Jones, 2003; Wainaina et al., 2018). From the greenhouses, the pest subsequently escapes into open fields, hence, its presence in both open and protected agricultural systems (Lapidot et al., 2014; Perring et al., 2018). French bean, Phaseolus vulgaris L., and tomato, Solanum lycopersicum L. are two of the most important vegetables in the horticultural sector of Kenya. French bean is an important export vegetable (Okello and Swinton, 2007; Gogo et al., 2014b) which is exported as fresh or canned, contributing to around 21.3% of the total value of vegetable exports, second to mixed vegetables (Horticultural Crops Directorate, 2019). In terms of production and value, tomato is the second leading vegetable after potato, contributing up to 20.1% of the total value of vegetables (Horticultural Crops Directorate, 2019) and mainly grown for the domestic market (Minot and Ngigi, 2004; Mithöfer et al., 2008; Gogo et al., 2014a). Since the cultivation of these crops by smallholder farmers in Kenya is mainly carried out in open fields, it makes them highly predisposed to whiteflies infestations. Together with other pests such as leafminer (Liriomyza spp.), thrips (Frankliniella spp.), red spider mites (Tetranychus spp.), and aphids (Aphis spp.), whiteflies are also listed among the major pests of tomatoes and French beans in Kenya (Nyasani et al., 2012; Gogo et al., 2014a). Tomatoes also serve as a propagative host for some of the criniviruses vectored by the pest (Wintermantel, 2004).

Farmers rely heavily on synthetic pesticides to control whiteflies (Nderitu et al., 1997) and this has been the basis to suppress the pest populations (Kim et al., 2014; Lapidot et al., 2014). However, factors such as high fecundity rates, a short life cycle which enables quick population build-up within a short time, a waxy cuticle layer that limits penetration by contact pesticides, and polyphagy which ensures abundant alternative hosts have led to the low success of the chemical control (Hirano et al., 1993; Gilbertson et al., 2011; Abd-Rabou and Simmons, 2012). This has thus placed them among the few pest species that have driven intensive pesticide use. The intensive application of pesticides, especially in areas of high infestations, as is often the case with T. vaporariorum (Palumbo et al., 2001) and in high-value crops like vegetables where pest thresholds should remain low, has led to the development of pesticide-resistant whitefly populations (Denholm et al., 1998). Therefore, whiteflies are now reported as major vegetable pests with high resistance to all classes of pesticides (Capinera, 2001; De Bon et al., 2014; Coffey et al., 2015), and with resurgence often seen even after spraying (Legg et al., 2014). Because of the negative impacts that high pesticide use poses to the environment, biodiversity, and public health, the adoption of sustainable crop protection methods has increasingly become key to sustainable agriculture.

The management of T. vaporariorum especially in protected agriculture has been successful using biological control agents such as parasitoids Encarsia formosa (Gahan) and Eretmocerus eremicus (Rose and Zolnerowich) (Hymenoptera: Aphelinidae) (Gonzalez et al., 2016); predators like Amblyseius swirskii (Athias-Henriot) (Acari: Phytoseiidae) and Delphastus catalinae (Horn) (Coleoptera: Coccinellidae) (Spence et al., 2020), and some species of entomopathogenic fungi, namely, Beauveria bassiana (Balsamo) Vuillemin, Metarhizium anisopliae (Metschnikoff) Sorokin (Hypocreales: Clavicipitaceae), and Isaria fumosorosea (Wize) Brown and Smith (Hypocreales: Cordycipitaceae) (Gökçe et al., 2005; Gonzalez et al., 2016).

Some entomopathogenic fungi, known as endophytes, live symbiotically within plant tissues for part of or their entire life cycle and are known to protect plants against abiotic and biotic stressors (Lacey et al., 2015; Vidal and Jaber, 2015; Jaber and Ownley, 2018). Endophytes stimulate the synthesis of secondary metabolites such as terpenoids, phenols, and phytoalexins volatile oils which confer resistance to several insects, resulting in the deterrence of feeding, oviposition, stem boring, sap sucking, and leaf mining by pests (Gao et al., 2011; Kambrekar, 2016; Agbessenou et al., 2020). The systemic effect of endophytes against insects has been shown in several other insects including diamondback moth [Plutella xylostella (L.)] (Lepidoptera: Plutellidae) (Batta, 2013; Sun et al., 2018), Silverleaf whitefly (B. tabaci) (Gennadius) (Hemiptera: Aleyrodidae) (Garrido-Jurado et al., 2017), pea leafminer (Liriomyza huidobrensis) (Blanchard) (Diptera: Agromyzidae) (Akutse et al., 2013), tomato leafminer (Tuta absoluta) (Meyrick) (Lepidoptera: Gelechiidae) (Klieber and Reineke, 2016; Agbessenou et al., 2020), cotton aphid (Aphis gossypii) (Glover) (Homoptera: Aphididae) (Lopez et al., 2014), mealybugs (Planococcus ficus) (Signoret) (Hemiptera: Pseudococcidae) (Rondot and Reineke, 2018), and spider mites (Tetranychus urticae) (Koch) (Acari: Tetranychidae) (Pappas et al., 2018). Another additional benefit from some endophytes is their potential to promote plant growth (Mayerhofer et al., 2013; Hassan, 2017; Bamisile et al., 2018a,b, 2020; Russo et al., 2019). Therefore, the study of plant–endophyte interactions is an important approach in continuing to build the knowledge on endophytes as candidates for the development of biopesticides against insects like T. vaporariorum, which have become resistant to synthetic pesticides. The objectives of the study were to assess the endophytic colonisation and persistence of some selected fungal isolates in P. vulgaris. and S. lycopersicum, evaluate their effects on plant growth, and assess the systemic effects of the endophytically colonised host seedlings on the development and survival of T. vaporariorum adults and progeny.

Endophytes play an important role in protecting plants against different pest species, pathogens, and environmental stresses (Mayerhofer et al., 2013; Gao et al., 2017; Gange et al., 2019). The success of this important association, however, depends on the compatibility between the host plant species and the endophyte. Compatibility determines whether the endophytes fail to establish or is able to proliferate within the different plant tissues. The results from the current study showed the successful colonisation of two plant species, S. lycopersicum and P. vulgaris, by endophytes from different genera and also demonstrated the influence of the host plant species, endophyte, plant part, and persistence time as key determinants of colonisation efficiency. This was exhibited by the pattern of colonisation which was different for the two host plants. Whereas in P. vulgaris, the full endophytic colonisation of all plant parts was achieved from the first week post-inoculation by all the isolates, the colonisation of the entire host plant in S. lycopersicum by all the endophytes occurred at the third-week post-inoculation. The colonisation rates within the successive host plant organs are an indication of the time an endophyte requires to passively move through interconnecting tissues to colonise the entire plant (Wagner and Lewis, 2000). The duration has been shown to be dependent on both the host and the endophyte species/strains (Qayyum et al., 2015). Our results showed a migration direction/speed of the endophyte inoculum from the root to the leaves, which was found to be faster in P. vulgaris than in S. lycopersicum. This observation concurs with what was highlighted by Akello et al. (2007) and Gange et al. (2019) that the differences in physiology and chemistry of different plant species influence colonisation efficiency. Wagner and Lewis (2000) also pointed out how the succulent nature of the corn pith proved ideal for the movement of B. bassiana as an endophyte. Since the same endophytes produced different effects in the hosts, it shows that the responses might be host-dependent. In addition, different plant species and different plant parts usually harbour diverse endophytes (Patil et al., 2016; Hassan, 2017), and therefore, successful colonisation also depends on the ability of the artificially inoculated endophyte to outcompete the other endophytes already present in the plant (Gao et al., 2011; Kambrekar, 2016). The host plant may also trigger a response that may be defensive, resulting in variations in the colonisation patterns between the two plant species and plant parts. As the plant ages, changes in the physiology, hormonal composition, and competition of the plant in the utilisation of nutrients with other endophytes might also result in lower colonisation (Gaiero et al., 2013; Russo et al., 2019). This was demonstrated in the S. lycopersicum endophytically colonised by T. atroviride F5S21, T. asperellum M2RT4, and T. harzianum KF2R41, which showed a significant decline in the root colonisation levels by the pre–inoculated endophytes in the final week, while a higher diversity of the naturally occurring endophytes was observed during reisolation, which were previously absent. A clearer trend in the decline of colonisation might probably have been observed if the assessment had gone beyond 5 weeks. Further studies are therefore warranted to elucidate these variations in both tomato and French bean plants.

There was very low colonisation by B. bassiana isolates in both hosts, 15% (ICIPE 676 and 281) for the roots in S. lycopersicum and 1.25% (ICIPE 609) in the P. vulgaris stems, which seems consistent with some studies that reported that B. bassiana is more competent as an above-ground endophyte than a root coloniser (Meyling et al., 2011; Behie et al., 2015) because of tissue specificity (Compant et al., 2016), with foliar sprays achieving higher colonisation than soil drenching (Donga et al., 2018; Saragih, 2019; Wei et al., 2020). However, since other methods of inoculation by B. bassiana such as P. vulgaris and Vicia faba seed inoculation (Akutse et al., 2013), S. lycopersicum root dip (Qayyum et al., 2015), P. vulgaris foliar sprays (Afandhi et al., 2019), soil drenching in Musa spp (Akello et al., 2007), and stem injection in Coffea spp (Posada et al., 2007) resulted in successful high colonisation rates in other studies, the low colonisation in the current study could have been due to host-endophyte incompatibility.

The weekly measurements of various growth parameters did not show significant growth promotions in the inoculated plants in some traits compared with the uninoculated controls. However, the effect of T. atroviride F5S21 in enhancing growth in S. lycopersicum was evident in the shoot biomass accumulated during the entire growing period. Trichoderma atroviride F5S21 gave the highest fresh and dry shoot weight compared with other treatments. Endophytes have been shown to enhance growth in other different plant species such as soya bean (Russo et al., 2019), banana (Akello et al., 2007), turfgrass (Gaggìa et al., 2013), red chilli (Saragih, 2019), neem (Verma et al., 2011), broad bean (Jaber and Enkerli, 2016), and cucumber (Khan et al., 2012). The increase of growth by endophytes is attributed to their role in stimulating the production of plant growth hormones such as auxins, indole acetic acid, and gibberellic acid which are all important for host growth regulation (Kambrekar, 2016; Hassan, 2017; Saragih, 2019). For instance, indole acetic acid is an important compound that integrates the symbiotic relationship between the host plant and the endophyte (Hassan, 2017). However, for T. asperellum M2RT4, it was observed that as the plants achieved more colonisation in the different parts, a lower plant height and leaf length were recorded compared with the other treatments in some weeks. The reason for this response could be explained by the fact that since the symbiotic relationship between the host plant and the endophyte is such that it derives its nutrients from the photosynthates of the plants and in turn provides the plant with nitrogen and phosphates (Behie et al., 2017), the increased nutrient demand and carbon drain may be the reason for the corresponding reduction in the speed of growth (Mack and Rudgers, 2008; Rodriguez et al., 2009). Furthermore, the plant also uses an abundant amount of energy to sustain the symbiotic relationship between itself and the endophyte, and therefore, slower growth, in this case, can be viewed as a compensatory effect (Wei et al., 2020). In contrast to S. lycopersicum, endophyte inoculation did not affect the P. vulgaris shoot biomass while the plant height and leaf width were generally lower for the T. asperellum M2RT4 inoculated and control plants only in the first week with no significant differences for the remaining weeks for these growth parameters. As highlighted by Mayerhofer et al. (2013), the contrasts in the results for the growth parameters in which different hosts respond differently to the same endophyte can be attributed to host specificity and may not be reproducible for all hosts. Other examples where endophyte inoculation did not result in enhanced plant growth are cacao (Hanada et al., 2010), tomato (Wei et al., 2020), radish (Sun et al., 2018), and sorghum (Tefera and Vidal, 2009).

The systemic effects of the endophytes on T. vaporariorum were also investigated in terms of oviposition, nymphal development, and adult and progeny survival. There were significant variations with the oviposition counts (number of eggs laid) among the different treatments. For instance, the oviposition on the S. lycopersicum inoculated with H. lixii F3ST1 and T. asperellum M2RT4 was the same as the control plants in contrast to P. vulgaris where the same treatments had oviposition counts significantly lower than the controls. Despite the high oviposition counts, the same endophytes, H. lixii F3ST1 and T. asperellum M2RT4, generally managed to suppress nymphal development and adult emergence on the endophytically colonised S. lycopersicum. The reason for the difference in the effect of endophytes on the oviposition and nymph development in the current study might possibly be a result of the endophyte-mediated oviposition preferences through volatile cues (Jallow et al., 2008). Feeding on inoculated plants, on the contrary, directly exposes the insect to secondary metabolites such as terpenoids, isoflavonoids, and isocoumarins that have toxic effects which inhibit insect performance (Gaggìa et al., 2013; Jaber and Ownley, 2018; Wei et al., 2020). Similar trends were also observed in other studies in which Helicoverpa armigera (Hübner) (Lepidoptera: Crambidae) moths (Jallow et al., 2008) and T. vaporariorum (Vidal, 1996) showed oviposition preference on tomato plants inoculated with the Acremonium strictum and laid more eggs compared with endophyte–free plants although the H. armigera larvae and T. vaporariorum nymphs which later developed on the inoculated plants had prolonged development time and a higher mortality rate. Gange et al. (2019) analysed several studies on plant-endophyte-insect interactions and showed that some parameters do not necessarily respond in the same pattern as also observed in our study for oviposition, nymphal development, and adult emergence. Similar studies have also shown the negative systemic effects of endophytes on different life parameters of several insects. H. lixii F3ST1 caused the lower pupation, emergence, and survival of L. huidobrensis in Vicia faba (Akutse et al., 2013), reduced the oviposition and mining activity of T. absoluta in both S. scabrum and S. lycopersicum (Agbessenou et al., 2020), and reduced the feeding and oviposition of Thrips tabaci (Lindeman) (Thysanoptera: Thripidae) in Allium cepa (Muvea et al., 2014). Trichoderma asperellum M2RT4 is also reported to have a negative effect on the development and reproduction of Acyrthosiphon pisum (Harris) (Homoptera: Aphididae) in V. faba (Akello and Sikora, 2012), while S. lycopersicum inoculated with B. bassiana caused the lower oviposition and feeding by B. tabaci (Wei et al., 2020). In this study, the dead insects which were exposed to the endophytically colonised plants did not show any mycosis. Even though the intercellular presence of endophytic fungi through artificial inoculation/colonisation and their systemic effects on herbivorous pests/diseases have been demonstrated by several authors (Hardoim et al., 2008; Yuan et al., 2011; Hiruma et al., 2016; Muthukumar et al., 2016), limited studies have provided evidence of fungal spores/conidia growth on infected cadavers exposed to the endophytically colonised host plants (Garrido-Jurado et al., 2017; Gange et al., 2019). Therefore, the reason for the registered mortality might probably be a result of the bioactive secondary metabolites rather than the fungus itself. Also, through the plant-endophyte interaction, some endophytes have been reported to induce the jasmonic acid signalling pathway which mediates plant defense responses against insects, both chewers and phloem feeders, such as whiteflies, reducing their overall fitness (Pappas et al., 2018).

Apart from the effect of endophytes on T. vaporariorum development and survival, there was generally higher oviposition on P. vulgaris than S. lycopersicum even on the control plants. Whiteflies are known to select the most suitable sites for oviposition and the external physical characteristics of the leaf surface, such as hairiness, influence oviposition behaviour (Van Lenteren and Noldus, 1990; Mansaray and Sundufu, 2009). In the present study, the whiteflies preferred P. vulgaris to S. lycopersicum, a result also supported by other studies where B. tabaci and T. vaporariorum showed oviposition preference on plants with fewer trichomes (Van Lenteren and Noldus, 1990; Avery et al., 2015). Leaf hairs can act as a physical barrier to oviposition by deterring female whiteflies from resting on the leaf surface and secreting defensive chemicals (Mansaray and Sundufu, 2009; Bar and Shtein, 2019). Notably, B. tabaci does not oviposit on very hairy Gossypium hirsutum varieties (Van Lenteren and Noldus, 1990). However, a contrasting result with the same whitefly species reported oviposition preference on hairy plants, Glycine max L. (Mansaray and Sundufu, 2009) and S. melongena L. (Shah and Liu, 2013), compared with glabrous ones. Since several other factors such as leaf orientation, colour, texture, and metabolites in the sap all contribute to the oviposition preference by whiteflies (Mansaray and Sundufu, 2009), it can only be postulated that leaf hairiness could be one of the reasons for the observed difference in this study in addition to the endophyte effects, and further research is needed to conclusively determine the cause of the observed variation in the two hosts species.

In conclusion, the use of endophytes is more beneficial, especially against sap-sucking insects such as T. vaporariorum. Our results showed the potential of endophytes H. lixii F3ST1 and T. asperellum M2RT4 as potential biocontrol tools in S. lycopersicum and P. vulgaris for the management of T. vaporariorum due to their suppressive effects on pest survival, oviposition, nymph development time, and adult emergence compared with the control and the other tested endophytes. However, further research needs to be undertaken to understand the mechanisms underlying variations observed concerning endophyte colonisation speed and their effect on oviposition preference by T. vaporariorum in the two hosts, S. lycopersicum and P. vulgaris. Further studies are also warranted to validate the underlined findings under field conditions to integrate these endophytes in the sustainable management of whiteflies in tomato and French bean cropping systems.

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

VP, FK, SS, and KA conceived and designed the experiment. VP performed the experiment and analysed the data. VP, KA, AY, SE, SS, and FK wrote the manuscript. All authors have read and agreed to the published version of the manuscript.

The views expressed in this article do not necessarily reflect the official opinion of the donors.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.