## Introduction

The introduction of non-indigenous species into new regions can be directly related to natural dispersion or to anthropic actions. Humans have caused accidental and deliberate dispersal of pests for millennia. Ancient human migrations and trade led to the early spread of animals and domesticated plants, along with associated parasites1,2,3. However, the introduction of non-indigenous species has intensified due to increasing global trade, with pest species rapidly spreading around the world, markedly as events post-Columbus3. The establishment of nonindigenous species in new regions depends on biotic and abiotic factors. Biotic factors include the presence or absence of natural enemies and the presence of suitable host plants. Abiotic factors are related to suitable environmental conditions and the effectiveness of protective measures adopted by man4. In addition, the spread of pest species is correlated with the same ecological or evolutionary phenomena that are responsible for the arrival of nonindigenous species in new regions5.

The tomato pinworm Tuta absoluta Meyrick (Lepidoptera: Gelechiidae) is currently considered as the most devastating invasive lepidopteran pest of tomato crops worldwide5,6,7,8,9. Tuta absoluta is native to South America and is a specialist predator of Solanaceae5,8,10. The destructive effect of these insects is due to the formation of “mines” within the leaves and the formation of galleries in the stems, fruits and flowers, which can result in losses of up to 100% in tomato crops11. Tuta absoluta was first reported attacking tomato crops in South America in the 1960s. In Europe, the first detection of T. absoluta occurred in Spain in 2006 and during the following years this pest was found in North Africa and sub Saharan Africa. More recently T. absoluta was detected in India, causing large losses in tomato crops grown in open fields and in greenhouses5,7,12,13,14. Tuta absoluta is now successfully established in Europe, Africa and Asia, becoming the main invasive pest of tomatoes in these regions. Favourable conditions for the establishment of this pest are due of the absence of natural enemies in these recently invaded regions15,16 and the abundance of susceptible host plant species5,10,17,18.

Tomato, Solanum lycopersicum L., is the preferred host when considering T. absoluta oviposition and development, but other solanaceous species, cultivated or non-cultivated (wild), have been shown to serve as alternative hosts, especially potato Solanum tuberosum L., eggplant S. melongena L., tobacco Nicotiana tabacum L., wild tomato Solanum spp., and thorn apple Datura ferox L.10,19,20,21,22,23,24,25,26. The Solanaceae are one of the most diverse plant families among the angiosperms, with approximately 9000–10,000 species and approximately 2000 of those species belonging to the genus Solanum. Members of this family occur in all the continents, except Antactica. The highest numbers of species are found in tropical and temperate climes, with a significant predominance in the Neotropics. Plants within this family vary from ephemeral herbs to large forest trees and many species are found colonizing areas modified by agricultural practices27.

The high diversity and extensive geographic distribution of solanaceous plants may have played an important role in the rapid and continuous spread of T. absoluta into new regions. Thus, in this study we seeked to answer the following fundamental question: how is the biological performance of the tomato pinworm influenced by host plant species? To answer this question the oviposition preference of T. absoluta on a range of solanaceous species was investigated and biological life tables for the tomato pinworm on each of these plant species were constructed.

## Materials and methods

### Rearing of Tuta absoluta

The tomato pinworms used in the bioassays were obtained from a colony which was established from insects originally collected from commercial tomato crops in Viçosa, Minas Gerias State, Brazil in 1994. The colony was initiated with approximately 200 field-collected individuals. The laboratory colony was maintained following the methodology of Silva et al.28, and kept under controlled conditions: temperature (25 ± 0.5 °C), photoperiod (12 h L: 12 h D) and relative humidity (75 ± 1%). The insects were reared in wooden cages (40 × 40 × 40 cm) covered with fine mesh netting. One cage was used for oviposition; one for maintaining the leaves with eggs and first instar larvae; one containing second, third and fourth instar larvae and one cage for pupae and adult emergence. The larvae were fed on tomato leaves of the “Santa Clara” variety, cultivated in a greenhouse without application of insecticides. The tomato pinworm population used in the bioassays was approximately the hundredth generation of this insect.

### Plant species used in bioassays

For bioassays, the Solanaceae used here included six cultivated species and four non-cultivated plant species, which occur naturally in tomato growing regions in Brazil. The cultivated solanaceous species were tomato Solanum lycopersicum L. (variety: Santa Clara), potato Solanum tuberosum L. (variety Ágata), green pepper Capsicum annuum L. (variety All Big), pepper Capsicum chinense Jacquin (variety Cumari do Pará), gilo Solanum gilo Raddi (variety Verde Claro), and eggplant Solanum melongena L. (variety Comprida Roxa). The seeds were obtained from a local market. The non-cultivated species were wild tomato Solanum habrochaites, American black nightshade Solanum americanum Mill., “jurubeba” Solanum paniculatum L. and “jua” Solanum aculeatissimum Jacq. (1787). The seeds of the above listed species were obtained from a germplasm bank (Vegetable Gene Bank of the Federal University of Viçosa, Minas Gerais, Brazil) or from collecting in vegetable cultivation areas. The plant seedlings were grown in styrofoam cell trays (68 × 34 × 5 cm) with 125 plugs (filled with vermiculite). After 30 days, the seedlings were transplanted to 6 L plastic pots containing substrate composed of 1/3 cattle manure and 2/3 soil. The soil fertility was corrected in accordance with agronomical recommendations. Plants were watered according to the daily necessity of each species. The temperature of the greenhouse was maintained at approximately 27 °C.

### Ethical approval

All international, national and institutional guidelines applicable to the care and use of animals were carefully followed in this study.

## Experiments

### Bioassay of oviposition preference on Solanaceae

Oviposition preference experiments were performed in the greenhouse (27 ± 3 °C; 60 ± 5% RH and 14L:10D) using 12 wooden cages (100 × 100 × 80 cm) covered in netting. One plant of each species was randomly placed in a circle within each cage. Fifty pairs of adult T. absoluta (50 two-day-old males and 50 females) were released into each cage and provided with a honey solution (10%) as a food source. Twenty-four hours later the adults were removed and the number of eggs laid on each plant species was counted. Each cage was considered as one replicate. In this experiment 1200 T. absoluta adults (600 females and 600 males) and 12 plants of each species were used.

### Mortality and survival of different Tuta absoluta developmental stages

This experiment was performed by constructing biological life tables for the tomato pinworm. In the greenhouse, under the same conditions as described above, thirty T. absoluta neonate larvae were transferred individually to leaves of each plant species. Prior to starting the experiments all plants were inspected and leaves with any visible symptoms of fungal disease were removed. The larval instars were determined from the width of the head capsule or exuvia from molted larvae. At the end of the fourth instar (prepupae), the larvae were removed and placed in test tubes and fed with pieces of plant leaves from each species until they pupated. The pupae were then sexed and weighed. Mortality of the larvae and pupae were recorded on a daily basis, as was the duration of the larval and pupal stages.

As the adults emerged from the pupae, one male and one female of the same age were transferred to a plastic oviposition cage (11.3 cm in diameter and 23.4 cm in depth). The oviposition cage was closed at the top with fine mesh netting to aid ventilation. In each cage one leaf disc from each host plant (6.2 cm in diameter) was offered for oviposition. The discs were placed in Petri dishes containing a layer of cottonwool moistened with distilled water. The number of eggs laid by the females was recorded daily and the leaf discs were replaced at that time. The adults were fed with a honey solution (10%).

### Data analysis

The oviposition preference data were subjected to analysis of variance (one-way ANOVA; PROC ANOVA; SAS System 2002), and the means were compared using the Scott-Knott test (p < 0.05).

The fertility life tables for the tomato pinworm were constructed based on Southwood & Hendersen29 and Galdino et al.8. The life table parameters were estimated and analysed from the Jack-knife procedure using a program developed by Maia et al.30 for SAS (SAS Institute 2000). The Maia method allows for the use of the log-rank test to analyse the different survival curves. The parameters calculated were; net reproductive rate (R0), intrinsic rate of natural increase (rm), and finite rate of increase (λ, individuals/females/day), mean generation interval (T), generation doubling time (Dt, day).

The net reproductive rate (R0), that is the cumulative summation of lxmx, is age-specific net maternity which is the product of lx (age-specific survival rate) and mx (age-specific fecundity), and was calculated as:

$$R_{0} = \sum\limits_{x = 0}^{\infty } {l_{x} m_{x} }$$

The intrinsic rate of natural increase (rm), where T is generation time:

$$r_{m} = \frac{{\ln \left( {R_{0} } \right)}}{T}$$

The finite rate of increase (λ):

$$\lambda = e^{{r_{m} }}$$

The mean generation interval (T):

$$T = \frac{{\log \left( {r_{0} } \right)}}{{r_{m} }}$$

The generation doubling time (Dt):

$$Dt = log_{2} \left( {r_{m} } \right)$$

The oviposition preference was checked for homogeneity of variance (Bartlett test) and normality of errors (Shapiro–Wilk test). As the data was homogenous and normally distributed, no transformation was necessary.

A cluster analysis using the Euclidian distance method was performed on the oviposition preference data, larval mortality, pupal weight, and adult fecundity of T. absoluta on the range of Solanaceae used here. The statistical software for this analysis was MINITAB version 17. The figures were created using SigmaPlot 12.5 (Systat Software Inc., San Jose, California, USA) and Inkscape 0.91 (www.inkscape.org).

## Results

Tuta absoluta oviposition preference was significantly influenced by the plant species offered to this insect (F = 144.91; df = 9, 110; p < 0.001). The highest percentage of eggs was laid on tomato S. lycopersicum leaves (51.15%), followed by potato S. tuberosum (17.64%) and wild tomato S. habrochaites (10.40%) (Fig. 1). Lower percentages of eggs were observed on S. gilo (2.32%), S. paniculatum (1.27%), C. chinense (0.97%) and C. annuum (0.79%) (Fig. 1A). There was a significant difference between the percentage of eggs laid on tomato plants when comparing plants grouped into cultivated and non-cultivated solanaceous varieties (F = 48.35; df = 2, 33; p < 0.001). Tomato plants were preferred when compared to all other cultivated or non-cultivated plants (Fig. 1B) and the total number of eggs laid on cultivated plant species was similar to the number laid on non-cultivated plants.

The mortality rates of the immature stages of T. absoluta were significantly affected by the plant species on which they developed (F = 67.80; df = 9, 50; p < 0.001). Mortality rates of the larval and pupal stages, when neonate insects were placed on the leaves of different plant species, are shown in Fig. 2. The Scott Knott test separated the plant species into five groups when considering mortality rates (larvae and pupae). The first group consisted of C. annuum, C. chinense and S. paniculatum (100% mortality). The second group consisted of S. aculeatissimum (83.33% mortality) and S. americanum (72.65% mortality). The third group was S. gilo (59.2% mortality) and S. melogena (60.1% mortality). The fourth group consisted of S. habrochaites (37.55% mortality) and S. lycopersicum (35.48% mortality). The lowest rates of larval and pupal mortality occurred on potato plants S. tuberosum (16.35%), which was significantly different to the other four groupings (Fig. 2). For C. annuum, C. chinense and S. paniculatum, larvae were observed to continue crawling on the leaf surfaces for long periods without creating mines.

In the analysis to determine the effect of the host plant on the biology and fecundity of T. absoluta, data for S. aculeatissimum was not included due to the low number of adults developing from the immature stages. For T. absoluta developmental factors on different host plants, the larval period (F = 50.99; df = 5, 533; p < 0.001), pupal period (F = 2.44; df = 5, 533; p = 0.034), adult longevity (F = 32.08; df = 5, 152; p = 0.009) and lifespan (F = 5.23; df = 5, 152; p < 0.001) were all affected by the host plant species. The longest larval period was observed on S. americanum (20.47 days) and the shortest on S. tuberosum and S. lycopersicum (13.86 and 14.52 days respectively; Table 1). S. americanum was unique in that development of T. absoluta on this plant resulted in an increase in the pupal period (11.58 days). The pupal period was 1.43 days longer than that of insects developing on S. lycopersicum. Adult longevity was highest on S. melogena and S. tuberosum (35.77 and 31.97 days respectively) and lowest on S. lycopersum (26.6 days). The shortest lifespan (larval, pupal and adult longevity) was observed on S. lycopersicum (53.62 days) but this result did not differ from the lifespan on S. habrochaites (51.27 days) and S. gilo (54.08 days). On the other hand, the lifespan was highest on S. melogena (63.34 days) and S. americanum (61.70 days) (Table 1 and Fig. 3).

Pupal weight was significantly affected by plant species (F = 32.08; df = 5, 533; p < 0.001). The highest weight was observed on S. americanum and S. tuberrosum (3.63 mg and 3.89 mg respectively). The lowest pupal weight was observed on S. habrochaites and S. melogena (2.71 mg and 2.56 mg respectively). The pre-oviposition period was also significantly affected by plant species (F = 2.64; df = 5, 72; p = 0.03), the longest pre-oviposition period occurred on S. gilo (3.50 days), whilst insects offered other plant species had similar pre-oviposition periods. The oviposition periods of adult T. absoluta were not affected by host plant species (F = 1.68; df = 5, 71; p = 0.15). However, the daily (F = 13.52; df = 5, 72; p < 0.001) and total fecundity (F = 12.50; df = 5, 72; p < 0.001) were highest when the larvae were reared on S. lycopersicum, with a daily mean fecundity of 13.41 eggs/female and a total mean fecundity of 211.88 eggs/female. For potato S. tuberosum, daily mean fecundity was 10 eggs/female and total mean fecundity was 188 eggs/female (Table 2). The lowest daily and total fecundities were observed for adults reared from larvae that had fed on S. americanum (3.33 eggs/female and 51 eggs/female respectively) (Table 2).

The highest net reproductive rate (R0) occurred when insects were reared on potato S. tuberosum (81.01) or tomato plants S. lycopersicum (72.72). The lower R0 was recorded on S. americanum (6.72). The shortest interval mean generation (T) and doubling time (Dt) on tomato S. lycopersicum were 30.60 and 4.95 respectively, and the highest T and Dt on S. americanum were 37.85 and 13.56. The highest intrinsic rates of natural increase (rm) and highest finite rates of increase (λ) were seen on tomato S. lycopersicum (0.14 and 1.15 respectively) or for wild tomato S. tuberossum (0.13 and 1.14 respectively) (Table 3).

The cluster analysis (Fig. 4) assembled the plant species into three groups (A, B and C). Cluster A (99% similarity) included C. annum, C. chinense and S. paniculatum, which were classified as unsuitable for T. absoluta. Cluster B denominated as the intermediate group, was divided into two subclusters (B1 and B2). Subcluster B1 (92% similarity) grouped S. aculeatissimum and S. americanum together. Subcluster B2 (80.76% similarity) consisted of S. gilo, S. melongena and S. habrochaites. Cluster C, considered as the most susceptible host plants group (78.40% similarity), consisting of S. lycopersicum and S. tuberosum.

## Discussion

This is the first study describing oviposition preference and biological performance of the tomato pinworm on cultivated and non-cultivated solanaceous plant species. It was not surprising that adult T. absoluta preferred to lay their eggs on domesticated tomato plants (S. lycopersicum), with more than fifty percent of the eggs being laid on this species when insects were offered a range of solenaceous plant species at the same time (Fig. 1). This preference for tomato is a result of evolutionary adaptation of T. absoluta to tomato toxins and the capacity of females to select sites that are most favourable for the survival of their offspring8. Interestingly, the other 50% of the eggs were laid on plants for which the tomato pinworm would not have had any previous contact, especially potato, wild tomato, jua, American black nightshade and eggplant. This demonstrates that T. absoluta females can recognize these plants as potential alternative hosts for the development of their offspring. Volatile compounds released by plants play an essential role in the location of suitable host species. These compounds can serve as airborne semiochemicals, promoting or deterring interactions between host plants and insect herbivores31. Volatiles from host plants enhance the effect of contact stimulants, increasing landing and oviposition rates when compared to non-host plants, with females capable of detecting small variations in host plant volatile signatures32,33.

The high levels of mortality that occurred on C. annuum, C. chinense and S. paniculatum could be due to the presence of deterrent compounds or physical barriers that prevented larval feeding, causing abandonment of the plants and subsequent mortality due to starvation. Interestingly, the survival rate of the immature stages was higher on potato than on tomato plants. Tomato plants are rich in trichomes, anti-feedant proteins, toxins and protease inhibitors that protect them against pest attack20,34,35. The movement of early larval instar T. absoluta on the surface of tomato plants is made more difficult because of their small size and the presence of protective barriers (trichomes and hairs) on the leaf surface36,37. To overcome these defences, T. absoluta larvae bore into the leaves, constructing mines close to the oviposition sites and they remain within the mine, feeding until the third or fourth instar, at which time they have developed morphological and physiological adaptations which enable them to tolerate the tomato defences8. This behaviour may be interpreted as an adaptive strategy of the tomato pinworm to avoid host plant defences.

The survival rates and development time of herbivorous insects on different plant species can be considered in the context of life history and adaptation of the insects to host plants38. High rates of survival and short development times are indicators of nutritional quality10,17,18,21,39. Thus, the higher development times that occurred on American black nightshade could be a reflection of the poor nutritional quality of this host. Potato plants could be considered as “better hosts” than tomato plants, as indicated by the higher survival rates, shorter development times and high female fecundity rates. These results are similar to those reported by Megido et al.17 when testing four potato varieties, where the development of T. absoluta larvae on potato was faster than on tomato plants. Whereas, previous studies10,18,21,23,25 found shorter development times on tomato than on potato. One possible explanation for these conflicting results is the range of potato and tomato varieties tested and differences in the genetic make-up of tomato pinworm populations used in each study.

The fecundity of T. absoluta was substantially influenced by the solaneceous host plant species. The daily fecundity rate was highest on tomato and lowest on jua, however the total fecundity rate on tomato was similar to that of potato. Fecundity is a good parameter to measure the effect of the host plant on lepidopteran herbivores as the reproductive potential is defined during the larval stage39,40. Nutritional quality of the host plant may influence the nutritional reserves and the allocation of resources in larvae, resulting in changes in fecundity, morphology and life history of the insect41,42. Previous studies with lepidopterans found a high correlation between body size and fertility, thus a higher immature body mass can result in larger adults, providing greater reproductive success39,40,43. This phenomenon occurred in potato, eggplant, tomato and wild tomato, although in gilo and American black nightshade, the higher pupal weight was not reflected in greater fertility of T. absoluta adults. Some insect species may compensate for poor nutritional status of the host plant by increasing the consumption rate and protracting larval development. However, these strategies result in a higher risk of juvenile mortality and environmental asynchrony, which may have strong effects on insect population dynamics40,43,44.

The net reproductive rate (R0) represents the potential of the female offspring to replace the mother throughout her life. This parameter is a good predictor of population growth and is strongly influenced by immature survival rates and female fecundity44. The high values of R0 calculated for S. americanum, S. gilo, S. habrochaites, S. lycopersicum, S. melongena and S. tuberosum indicated that T. absoluta populations would be predicated to increase when larvae developed on these plants. These experiments were performed under controlled conditions, with T. absoluta larvae isolated from natural environmental factors (natural enemies and climatic elements). Under field conditions, the natural mortality of T. absoluta is around of 92–99%, as verified in ecological life tables by Bacci et al.16 and Miranda et al.36. The colonization of plants with low nutritional quality may jeopardize the persistence of non-indigenous insects in the environment due the high mortality, decreased fecundity and extended life cycle, increasing exposure to natural mortality factors15,40,43. However, environment conditions suitable for survival and the absence of natural enemies may reduce the impact of food quality on the establishment and spread of invasive species into new regions. Cherif et al.26 demonstrated that temperature and relative humidity can impact the biological performance of T. absoluta on host plants such as tomato, potato and eggplant. On average, 81 females replace the mother when insects were reared on potato, and this number was similar to that on tomato (72.72). A possible explanation for this result could be the low mortality of the immature stages (larvae and pupae) (24%) and high female fecundity (211.88 eggs) following development on tomato and potato plants (13% and 188 eggs). These results are different to those obtained by Pereyra and Sanchez20 and Kanle Satishchandra et al.45, who documented a higher R0 on tomato (49 and 52.15) in relation to potato (14 and 27.60). We attribute these differences to the plant varieties and genetic differences between T. absoluta populations used by each research group.

The intrinsic rate of increase (rm) denotes the rate at which a population increases in size without resource limitations46. The rm is regularly used in entomological research to assess plant antibiosis effects41,47,48,49,50. A small delay in the reproduction of an organism with a high rm can reduce net reproduction more than proportionally. When the rm is low, fecundity becomes a critical factor in altering the rate of population growth. The rm values for T. absoluta were influenced by larval mortality, development time and adult fecundity of insects reared on the range of solanaceous species used here. The highest rm value was recorded for tomato due to higher fecundity and shorter development times (30.6 days) during the immature stages. However, the low rm value recorded on American black nightshade (S. americanum) was a result of lower fecundity (50.7 eggs/female) and lower immature stage survival rates (72.65%), as well as longer development times (37.85 days).

The dendrogram clusters clearly demonstrated the performance of T. absoluta on the host plants tested here. Cluster A plants were considered as unsuitable for oviposition and inhibitory for larval development. Cluster B (both subclusters B1 and B2) was formed of plants that showed moderate resistance to this pest and cluster C was formed of plants that were highly susceptible to T. absoluta. Subcluster B2 brought together cultivated plants (gilo and eggplant) and a non-cultivated plant (wild tomato). This cluster received around 18% of the eggs in the oviposition preference test. The immature survival rate was 40–60% and adult fecundity was 89–127 eggs, whilst R0 was 18–38, rm 0.08–0.11, giving a biological performance similar or higher than that seen on tomato17,21,51,52,53 and potato17,21 in others studies. This finding indicates that these plants may be used as alternative hosts by T. absoluta. Cluster C contained only tomato and potato, with a similarity of around 78%. This high similarity was due to the moderate rate of adult oviposition on potato (18% of the eggs during the oviposition preference test), lower larval mortality (20%) and high adult fecundity (188 eggs/female) of insects reared on potato plants. Although potato plants were suitable for T. absoluta larval development, there appears to be a barrier that reduces oviposition rates on potato leaves. In wild potato the presence of glandular trichomes (type A and B) are responsible for resistance to oviposition of potato tuber moth Phthorimaea operculella (Zeller) (Lepidoptera: Gelechiidae), and this has been related to larval mortality54,55. We hypothesize that the presence of trichomes could also be a barrier against T. absoluta oviposition on potato, however more studies are necessary to clarify this possibility.

In this study we used a T. absoluta population that has been maintained in the laboratory for more than one hundred generations without introducing new individuals. This extended time in the laboratory induces genetic drift due to consanguineous mating, resulting in the production of homozygotic progeny, with a loss of fitness or production of ecotypes unsuitable for field conditions54,55. However, when T. absoluta adults and larvae were confronted with a range of solanaceous species, they maintained a high capacity to colonize these plants. We believe that field populations of this pest species also possess a very high capacity to colonize a range of solanaceous plants, because of greater genetic diversity resulting from gene flow between populations and highly selective environment pressure56,57.

The results clearly demonstrate that T. absoluta is capable of laying eggs and developing on a range of cultivated and non-cultivated solanaceous plants under laboratory conditions. In the field and in greenhouses, T. absoluta has been reported attacking S. muricatum L., Nicotiana tabacum L, S. nigrum L., S. eleagnifolium L., S. bonariense L., S. sisymbriifolium Lam., S. saponaceum, Lycopersicum puberulum Ph., Physalis peruviana, Phaseolus vulgaris, Lycium sp. and Malva sp6,19,24,58,59,60,61. However, more studies are necessary to investigate the biological performance of T. absoluta on these host plants. It is therefore very important for the effective management of this pest, to eliminate alternative host plants from areas close to susceptible crops and to avoid simultaneous sowing of hosts preferred by T. absoluta in adjoining plots.