Quantifying the impacts of symbiotic interactions between two invasive species: the tawny crazy ant (Nylanderia fulva) tending the sorghum aphid (Melanaphis sorghi)

Introduction

When invasive species arrive to a new location, they may establish symbiotic interactions with resident organisms (Cleland & Mooney, 2001). New symbiotic interactions have allowed some invasive species to increase population size, expand their geographic range, or disrupt the presence of native species and natural enemies, among other possible scenarios (Simberloff & Holle, 1999; Wilder et al., 2011; Traveset & Richardson, 2014; Prior et al., 2015; Meijer et al., 2016; Hulcr et al., 2021). In particular, mutualisms between native and invasive species or among invasive species may ecologically facilitate biological invasions (Helms & Vinson, 2003; Daane et al., 2007; Green et al., 2011; Rassati, Marini & Malacrino, 2019). Ants can adjust foraging preferences to incorporate carbohydrate rich resources for nutrition (Csata & Dussutour, 2019), facilitating interactions with honeydew-producing hemipterans. For instance, in field experiments it was reported that invasive red imported fire ant (Solenopsis invicta Buren, RIFA) tended invasive mealybugs (Phenacoccus solenopsis Tinsley), which increased ant populations, likely due to the provision of honeydew to ants (Zhou et al., 2012). In addition, this study found that RIFA was more numerous at this honeydew resource, crowded out the presumed native ants (Tapinoma melanocephalum Fabricius), and that mealybug populations increased, potentially due to decreased parasitism in the presence of ants (Zhou et al., 2012). While ant-hemipteran interactions can often increase plant damage, ant presence on plants and tending behavior can decrease the presence of other non-hemipteran herbivores and subsequently reduce plant damage (Styrsky & Eubanks, 2007). For example, the presence of RIFA on cotton plants was reported to increase cotton aphid numbers while decreasing chewing herbivores (Diaz, Knutson & Bernal, 2004). Similarly, in other systems the presence of ants decreased chewing herbivores and increased overall plant health or fruit set (Styrsky & Eubanks, 2007; Singh et al., 2016).

Invasive ants are often reported to have traits that confer advantages over native competitors, including higher aggression to other ant species, greater reproductive output, a more general diet, or better ability to tolerate disturbance (Porter & Savignano, 1990; LeBrun, Abbott & Gilbert, 2013; Calcaterra, Cabrera & Briano, 2016). Beneficial symbiotic interactions or mutualisms can also promote invasive species growth and establishment (Traveset & Richardson, 2014). The interaction between ants (Hymenoptera:Formicidae) and aphids (Hemiptera:Aphididae) is an excellent system to explore potential mutualisms, due to these taxonomic groups’ variability in symbiotic interactions over evolutionary history (Stadler & Dixon, 2005; Pringle, 2021). Although there are many beneficial instances of ants tending aphids, the level of ant attendance towards different aphid species varies (Stadler & Dixon, 2001; Yao, 2014); with some aphids rarely being tended by ants while others rely on ant tending for increased survival via reduction in natural enemies or removal of excess honeydew. In addition, there are instances in which ants may eat aphids, regulate aphid population size, or reduce dispersal (Offenberg, 2001). In contrast to symbiotic interactions among co-evolved organisms (e.g., leafcutter ants and their symbiotic microbes, Lycaenid butterflies and ants, ambrosia beetles and their fungi, etc.,) (Mueller et al., 2005), invasive species may engage in new symbiotic interactions in new locations. The potential for invasive species symbiotic interactions to enhance invasion processes requires further investigation (Ness & Bronstein, 2004; Holway et al., 2002).

To further understand how a potential mutualism may influence two recently invasive species, we quantified the symbiotic interaction between sorghum aphid (Melanaphis sorghi Theobald), which are distributed across the continental US (Nibouche et al., 2021) and tawny crazy ant (Nylanderia fulva Mayr), which occupy some portions of the southern US. While there are instances of different ant and aphid species engaging in symbiotic interactions across evolutionary history, the interaction of these invasive insect species is relatively recent. Tawny crazy ant (TCA) are known to be opportunistic tenders of a variety of hemipterans, and sorghum aphid are tended by a variety of ant species in Texas (Wright, 2021; Holt, 2022). Multiple ant species including RIFA, acrobat ants (Crematogaster spp. Lund), and TCA have been observed tending sorghum aphids (Wright, 2021); Holt personal observations). While sorghum aphids have already resulted in significant economic damage to grain sorghum crops (Villanueva et al., 2014; Zapata et al., 2016), a potential mutualism between ants and these invasive aphids could increase pest population sizes, potentially resulting in greater damages to sorghum through reduced availability of nutrients, yield reduction, buildup of honeydew which results in sooty mold growth and reduced photosynthesis, and potential pathogen transmission (Zapata et al., 2016; Zapata et al., 2018). TCA opportunistic tending behavior was recently reported to cause greater plant damage in sugarcane crops (Pazmino-Palomino, Mendoza & Brito-Vera, 2020). In Florida, TCA was reported as a pest around the 1990s, and has opportunistically tended both native (e.g., willow aphids, juniper aphids, psyllids, cottony maple scale, kermes scale) and invasive (e.g., citrus whitefly, cowpea aphids, mealybugs) hemipterans (Sharma, Oi & Buss, 2013). Both TCA and sorghum aphid have continued to expand their geographic range (MacGown, 2015; Nibouche et al., 2021), and now these insects occur in many of the same geographic regions.

In addition, we examined whether aphid-host plant adaptation or, more broadly, local adaptation, influenced ant-aphid interactions or aphid biomass. In some instances, host plants may play a role in attracting ant tenders. For example, black bean aphids (Aphis fabae Scopoli) preferred genotypes of tansy host plants (Tanacetum vulgare L.) that emitted chemical volatile blends including eucalyptol and terpineol, and black garden ants (Lasius niger L.) were attracted to these volatile blends, which facilitated ant-aphid interactions by bringing ants in contact with aphids (Zytynska et al., 2019). Local adaptation may also influence ant tending rates. Such was the case with cowpea aphid (Aphis craccivora Koch), where aphids collected from one location were tended more frequently by ants than those collected from a different location further away (Katayama et al., 2013). Thus, aphids and ants adapted to different geographic locations or host plants may generate different mutualistic interactions. Aphids from different host plants were used in our study since sorghum aphids commonly occur on grain sorghum and Johnson grass. More recently sorghum aphid was reported on sugarcane (Nibouche et al., 2021), potentially due to aphids blowing over onto this host plant.

Given the opportunistic tending behavior of TCA, we hypothesized that ants would readily tend sorghum aphid in a greenhouse setting. We also hypothesized that this symbiotic interaction would result in increased aphid biomass, suggesting that in some instances this may be a mutualistic interaction. In order for tending to occur ants need to discover aphids. We hypothesized that aphids located lower on the sorghum stem would be more easily discovered by foraging ants and receive greater tending activity than those located higher up on leaves. To accomplish this, we quantified the tending behavior duration of TCA workers to sorghum aphids. In addition, we analyzed whether aphid biomass was influenced by ant presence or the host plant from which an aphid was collected (i.e., sorghum, Johnson grass, sugarcane). Overall, our study found that symbiotic interactions can establish relatively quickly between recently introduced invasive species, and that this interaction should continue to be monitored.

Methods

Ant-Aphid interaction experimental design

Plant cages built to enclose each experimental unit were placed inside a Fluon-coated (Insect-a-Slip Barrier by BioQuip, Rancho Dominguez, CA, USA) container (Sterlite 58 Quart Storage Box dimensions 59.69 L × 42.88 W × 31.11 H cm). Plant cages were built using PVC pipe (1.27 cm diameter) assembled into a rectangular shape (dimensions 25.4 L × 43.18 W × 76.2 H cm), and mesh bags composed of 100% polyester white voile (86.36 cm L × 182.88 cm H) were used to enclose the PVC frame. As an additional barrier to prevent ants from escaping, the table legs of the greenhouse benches were placed into a container filled with unscented mineral oil (Equate Mineral Oil, Al Ahmandi, Kuwait).

A total of 20 individual aphids from a clone line were transferred onto individual sorghum plants with a damp paintbrush. Ten aphids were placed on the stem and ten were placed on the 3rd leaf from the bottom of each sorghum plant used in experiments. The same aphid clone was used for a pair of experimental units (i.e., a sorghum plant with aphids and ants and a sorghum plant with aphids and without ants). Aphids used in the experiments were from eight clonal lines: four clones from Johnson grass, two from sorghum, and two from sugarcane (Table S1). Each aphid clone was used at least once for an observation period (Table S1).

In the case where aphids were placed with ants, aphids were allowed to acclimate between two to four hours before adding an ant sub-colony (500 workers + 1 queen). Each new ant sub-colony was placed inside a 1-gallon bucket (Lowes, Mooresville, NC, USA) filled approximately 90% with soil (Jolly Gardener Pro Line C/25, Atlanta, GA, USA) centered around a 6-inch round pot (McConkey, Sumner, WA, USA) that contained a plant with 20 aphids. The pot nested inside the bucket was placed inside of a PVC frame with a mesh cage that was tied at the top, and then this was placed inside a large plastic container coated with Fluon at the top rim (Fig. 1). This experiment included aphids collected from sugarcane (n = 6), from grain sorghum (n = 3), and from Johnson grass (n = 4), which were in the presence or absence of ants (n = 2), for a total of 26 experimental units. Within each unit, aphids were either placed on a grain sorghum plant stem or leaf (n = 2), yielding a total of n = 52 observations (n = 26 with ants and n = 26 without ants).

Results

Sorghum aphid honeydew excretion behavior, ant tending, and aphid biomass

Sorghum aphids generate copious amounts of honeydew, which was observed being flicked away from the insect when it was not tended by ants (https://youtu.be/VJcEpk5jXJw). However, after being antennated by TCA workers, sorghum aphids were observed modifying this behavior to excrete honeydew for ant consumption (https://youtu.be/H8jD5YkBkpM).

We analyzed whether ant tending duration of aphids was significantly different from no tending events. When tending was analyzed by aphid position on plants using a t-test, we found that the tending duration was significantly higher than zero when aphids were feeding on stems (t = 4.68, df = 12, p = 0.001, n = 13) and leaves (t = 2.97, df = 12, p = 0.012, n = 13). After fitting a LMM, we found that only aphid position on plant significantly influenced the TCA tending duration (χ² = 10.400, df = 1, p = 0.001, n = 26), with a higher value when aphids were located on stems (estimate = −3.29, p = 0.014, Fig. 2A), and neither host plant origin (χ² = 1.905, df = 2, p = 0.551), or the interaction between host plant origin and aphid position (χ² = 1.337, df = 2, p = 0.513), influenced TCA workers tending duration towards sorghum aphids.

We also analyzed whether aphid biomass was affected by the host plant aphids were collected from, aphid position, or the presence of TCA using a LMM. We found a significant effect driven by the interaction between host plant aphids originated from, aphid position, and ant presence (χ² = 6.93, df = 2, p = 0.031, n = 52, Fig. 2B). Post-hoc contrasts showed that aphids had a higher biomass when feeding on sorghum plant stems compared to those feeding on sorghum leaves, when aphids originated from Johnson grass, and when ants were present in the system (estimate = −3.243, p = 0.007; Fig. 2B). No differences in biomass were observed for aphids in the presence of ants when aphids originated from sorghum (estimate = −1.057, p = 0.431) or originated from sugarcane (estimate = 0.747, p = 0.431), nor in the absence of ants for aphids from sorghum (estimate = 0.733, p = 0.584), sugarcane (estimate = −1.860, p = 0.053), or Johnson grass (estimate = −1.825, p = 0.119).

Discussion

We found that although tawny crazy ant and sorghum aphid are relatively recent invasive pests in the US and are not previously reported to have coevolved with each other, both organisms engaged in a potential mutualism. This invasive ant antennated sorghum aphids for honeydew while sorghum aphids often modified their behavior from flicking honeydew to secreting honeydew after being antennated by TCA. Similar to other reports of TCA being opportunistic tenders of other hemipterans (Sharma, Oi & Buss, 2013), this was also the case with sorghum aphids.

In this study, the position of aphids on a sorghum plant was found to influence TCA worker tending rate. Higher tending occurred when aphids were located on the stem of a sorghum plant than on the leaves. This difference in tending could have been influenced by ants scouting and finding aphids more often when they were located on the stem compared with when aphids were higher up on the leaf. This could also reflect the quality of the honeydew resource when aphids were located on stems versus leaves or the proximity of the resource to the ant nest. While sorghum aphids in field patches of Johnson grass and sorghum were tended by ants on leaves of different heights (personal observation) this was potentially due to increased foraging by larger ant colonies, that honeydew flicked by aphids may attract ants, and/or that aphids walked to different parts of the plant or were moved by ants to facilitate tending. Interestingly, aphids collected from Johnson grass showed increased biomass when positioned on stems versus leaves, while this was not the case for aphids collected from sugarcane or sorghum. In addition, this increase in biomass was recorded only when ants were present. It is possible that ants preferentially tended aphids on stems due to this honeydew containing a greater sugar and nutrient concentration than that exuded by aphids on leaves. For instance, soybean aphid (Aphis glycines Matsumura) feeding on stems were found to increase in numbers more than those feeding on leaves, likely due to the greater content of amino acids and sugars when compared with phloem contents of leaves (Nalam et al., 2021). Since all aphids were reared on sorghum plants in our study, it is possible that the aphids collected from Johnson grass had greater reproductive outputs in the presence of ants due to the more extended availability of Johnson grass throughout the year and increased likelihood that aphids are tended by ants. This is in contrast with aphids collected from sorghum or sugarcane, crops that are grown for a shorter time, which may proportionally reduce aphid interactions with ants.

Understanding the symbiotic interactions of invasive species can aid with pest management efforts. For example, after the invasive mealybug (P. solenopsis) arrived in China, it established a new interaction with the native ghost ant (T. melanocephalum) in less than five years (Feng et al., 2015). This symbiotic interaction benefited ghost ants that received mealybug honeydew, while mealybugs gained ant protection against native parasitoid wasps (Feng et al., 2015), increasing population growth of both species (Zhou et al., 2014). As a result, invasive mealybugs were likely to cause greater injury (e.g., loss of plant nutrients, leaf chlorosis, decreased seed set/yield, pathogen transmission) to cotton and hibiscus plants in the future. Similarly, the spread of invasive Argentine ant (Linepithema humile Mayr) across the US was likely facilitated by the consumption of honeydew produced by native membracids (Harvey & Wheeler, 2015). Despite the recent formation of this symbiotic interaction between the aforementioned insects, this tending interaction resulted in damage to both urban and wildland systems (Menke, Ward & Holway, 2018). Symbiotic interactions, many of which are beneficial to invasive insects, can therefore facilitate ecosystem disruption and, as a consequence, can result in future potential invasions.

In the scenario of TCA workers and sorghum aphids, a symbiotic interaction may be beneficial to both these insect pests and increase numbers in sorghum fields. In our experiment, visual inspection of TCA colonies at the end of the experiment revealed the production of brood, suggesting a trajectory of colony growth in the presence of sorghum aphids that requires further investigation. This beneficial dynamic has been reported for TCA in other scenarios. In Colombia, opportunistic tending of mealybug (Antonina sp.) by TCA resulted in damaged grasslands, which reduced the forage quality and value for cattle (Zenner De Polania, 1990). Similarly, higher sugarcane aphid (Melanaphis sacchari Zehntner) presence on sugarcane was reported in Ecuador, where large TCA populations tended these aphids (Pazmino-Palomino, Mendoza & Brito-Vera, 2020). Previously, aphid injury towards sugarcane was considered low and aphid population sizes were controlled by parasitoid wasps and other natural enemies. However, TCA tending reduced natural enemy presence (Pazmino-Palomino, Mendoza & Brito-Vera, 2020), and resulted in increased sugarcane aphid populations. In these instances, interactions between two invasive pests ecologically facilitated aphid fitness and reduced crop health. Currently in the US, the composition of sorghum aphid natural enemies and ability to control aphids has been reported (Hewlett, Szczepaniec & Eubanks, 2019; Maxson et al., 2019; Elliott et al., 2021; Faris, Brewer & Elliott, 2022; Faris, Elliott & Brewer, 2022) along with the finding that RIFA did not negatively impact the presence of natural enemies that parasitize sorghum aphids feeding on sorghum (Wright, 2021). However, additional assessments of ant-sorghum aphid dynamics is needed, which could include identifying other ant species that are likely to increase aphid biomass or assessing the impact of ant presence on natural enemies in different ecosystems. Potential factors influencing the symbiotic interaction between invasive TCA and sorghum aphid are complex. For instance, conditions reflective of climate change (i.e., increased CO₂ and increased temperature) were reported to cause increased honeydew production by aphids (Kremer et al., 2018; Blanchard et al., 2019), which could result in a greater attraction of ants and reduction in plant nutrients. Factors such as honeydew concentration or chemical composition (Wright, 2021), or honeydew bacterial composition (Holt et al., in preparation) could also influence parasitoid or ant attractiveness to honeydew exudates. The intensity of ant-aphid interactions therefore may vary by aphid population, host plant consumption, or geographic location.

In our study, ant tending was observed both at the initial placement of twenty aphids and as aphid numbers increased; although this tending only significantly increased sorghum aphid biomass when those aphids were collected from Johnson grass. This is similar to the finding that sorghum aphid population growth can be variable. In one study, sorghum aphid populations tended by RIFA were reported to have significant population increases when aphid numbers were low compared with when aphid numbers were high (Wright, 2021). The number of aphids on a plant may differ from the overall biomass of the aphids (i.e., small aphids versus large aphids), which is likely to influence the amount of honeydew that is secreted, and the potential attractiveness of honeydew to ant tenders. Further investigation into how aphid numbers affect ant tending or the long-term impact ant tending has on aphid biomass on a host plant is needed to understand ant and sorghum aphid dynamics. The impacts of sorghum aphid honeydew on the presence of ants and/or on colony growth still need to be quantified. For instance, invasive RIFA were reported to have increased colony growth in the presence of cotton aphid (Aphis gossypii Glover) feeding on cotton plants (Wilder et al., 2011). An increase in TCA colony growth from consuming honeydew could result in further displacement of native arthropod species or an exacerbation of invasive hemipteran damage.

Conclusions

While our study finds that the presence of TCA can increase sorghum aphid biomass in greenhouse settings, there remain unanswered questions about the potential ecological impacts of this interaction. Since the symbiotic interaction between these two insects is recent, with overlapping distributions and potential interactions as early as 2013 (Meyers & Gold, 2008; MacGown, 2015), it is possible that the intensity of this interaction will increase over time. Symbiotic interactions between these two invasive insects could cause population increases for aphids as well as ants. Focusing on agro-ecosystems, it is still unknown what the potential long-term consequences of ants tending sorghum aphids are on aphid biomass over a growing season or to plant health and grain production. When we shift to natural ecosystems or wildland/green space areas we should consider what roles this symbiotic interaction may play in increasing invasive ant population growth, excluding native insects, or facilitating further invasive organisms. These symbiotic interactions may alter the effectiveness of management practices such as biological control or further disrupt ecosystems. Overall we recommend continued monitoring of the interaction between sorghum aphids and invasive ants.

Supplemental Information