First record of an African grass-feeding wasp ( Tetramesa ; Eurytomidae) on the invasive grass Eragrostis curvula (African lovegrass; Poaceae) in Australia

An undescribed phytophagous wasp belonging to the Tetramesa genus (Hymenoptera: Eurytomidae), that is native to South Africa, is currently being investigated as a potential weed biological control agent for the invasive grass Eragrostis curvula (Poaceae) in Australia. Host-specificity testing is underway in South Africa, but the wasp has not been exported into quarantine in Australia and further research is required before it could be considered for release. Here, we used DNA barcoding to demonstrate that Tetramesa specimens collected on invasive E. curvula populations in Australia represent the same wasp species currently being investigated in South Africa. We discuss our findings in the context of developing a biological control programme against E. curvula in Australia and the potential risk posed to native Australian grasses.


Introduction
The stem-galling wasp genus Tetramesa Walker (Hymenoptera: Eurytomidae) is a cosmopolitan genus of 204 described species, and numerous presently undescribed species . The larvae of Tetramesa are phytophagous, endophagous borers of cereals and other grasses (Poaceae) (Claridge 1961). Most Tetramesa have a narrow host range, with most species being recorded from a single genus of host plants (Claridge 1961;Goolsby and Moran 2009;Sutton et al. 2021). Tetramesa spp. can also have significant impacts on plant productivity, with several species considered important cereal crop pests (Spears and Barr 1985). Due to their high degree of host specificity and potential to reduce plant fitness, Tetramesa have been identified as high priority biological control agents for grasses that have become invasive in non-native geographic regions across the globe (Goolsby and Moran 2009;Sutton et al. 2019;Olckers et al. 2021).
Since 2017, the Centre for Biological Control (CBC) at Rhodes University in South Africa has performed extensive field surveys and laboratory trials in South Africa investigating numerous Tetramesa as biological control agents for African grasses that have become invasive in Australia or the USA, namely: Sporobolus pyramidalis Beauv., Sporobolus natalensis (Steud.) Durand & Schinz, Eragrostis curvula (Schrad.) Nees, Megathyrsus maximus (Jacq.) B.K. Simon & S.W.L. Jacobs and Andropogon gayanus Kunth (Sutton et al. 2019Olckers et al. 2021). Little is known about the Tetramesa occurring in Africa, with only four described species from the continent: T. aristidae Risbec from Senegal, T. decaryi Risbec from Madagascar, T. tananarivense Risbec from Madagascar, and T. macalusoi De Stefani from Somalia (van Noort 2022). No formally described Tetramesa are known from South Africa other than an adventive population of Tetramesa romana Walker found on the invasive reed Arundo donax L. (Poaceae) (Canavan et al. 2014). As such, all the Tetramesa species being investigated for biological control in South Africa remain undescribed. However, DNA barcoding has allowed us to distinguish among different putative novel species of Tetramesa in South Africa and assess their host range and potential as biological control agents. These DNA phylogenies will also assist with the formal description of these species by relevant taxonomic experts, which is currently underway.
Eragrostis curvula is considered an invasive weed in Australia and a biological control programme has been initiated. Surveys for potential agents have been conducted in South Africa resulting in the discovery of an undescribed Tetramesa species associated with the target plant (Olckers et al. 2021). The same Tetramesa species, confirmed with molecular barcoding, has been reared from at least six native non-target Eragrostis species under field conditions in South Africa, namely: E. capensis (Thunb.) Trin., E. plana Nees, E. gummiflua Nees, E. biflora Hack., E. cylindriflora Hochst. (=E. rigidior Pilg.) and E. trichophora Coss. & Durieu (Sutton et al. unpublished). Laboratory no-choice testing has also demonstrated that this Tetramesa species is able to complete development and produce F 1 progeny on at least two native African Eragrostis species, namely: Eragrostis planiculmis Nees and Eragrostis plana Nees (Sutton et al. unpublished). It is therefore questionable whether this particular Tetramesa species is suitably host specific for release in Australia, and additional host-specificity testing against several native Australian Eragrostis species is currently being performed under quarantine conditions in South Africa to assess this.
Ad hoc surveys by two of the authors (GFS, AM) across New South Wales, Victoria and South Australia, constituting a large portion of the invaded distribution of E. curvula in Australia ( Figure 1) were conducted in September 2022, and resulted in the discovery of a Tetramesa sp. attacking invasive populations of E. curvula ( Figure 2). The wasp was recorded at six out of eight sites visited, being absent from the two field sites in South Australia ( Figure 1). Where present, the wasp was found in high numbers Eragrostis curvula records in Australia were downloaded from the Global Biodiversity Information Facility using the R package 'rgbif' (Chamberlain et al. 2022; GBIF Occurrence Download 2022).

Figure 2. Tetramesa A)
Adult female (note the characteristic "golden" pronotal spot), B) larva collected on Eragrostis curvula in Australia, and C) characteristic emergence holes located above the 1 st and/or 2 nd tiller node Scale bars represent 1 mm.
across all different life-stages (larvae, pupae and adults). Field observations indicated that the wasps were causing damage to the host plants, with infested grass tillers showing severe stem deformation, and in many cases, wasps appear to have stopped stem elongation and inflorescence formation entirely, which prevents seed set. The presence of Tetramesa can be rapidly assessed by characteristic emergence holes located above the 1 st and 2 nd tiller nodes (Figure 2). To date, only one Tetramesa species is known from Australia, namely Tetramesa australiensis Girault (=Isosoma australiensis) (Girault 1913). It was described from a single male specimen collected using a sweep net in Kuranda in northern Queensland, and as such, no host-plant records are available (Girault 1913). It is however likely that other Tetramesa species may be present and indigenous in Australia that are undescribed. Given the high degree of host specificity exhibited by most Tetramesa species, and the non-native status of E. curvula in Australia, it is plausible that these wasps collected in Australia may be the same species of Tetramesa that is recorded on the plant in South Africa.
The aim of this study was to document the occurrence of the undescribed Tetramesa species on E. curvula in Australia, and evaluate whether the wasps may represent a previously unknown introduction of the South African Tetramesa sp. currently being assessed as a biological control agent or if it is a novel association between the invasive grass and a native Australian Tetramesa. The morphology of both adults and larvae of Tetramesa is extremely cryptic, with little morphological variation present to allow for distinction of different species. Moreover, the lack of a taxonomic backbone for Tetramesa in both Africa and Australia largely negates the use of traditional morphological taxonomy for this group. As such, we compared cytochrome oxidase I (COI) gene sequences for South African Tetramesa associated with native African Eragrostis species, with sequences generated for representative Tetramesa samples collected on E. curvula in Australia. We discuss our findings in the context of developing a biological control programme against E. curvula in Australia.

Methods
Grasses showing signs of possible Tetramesa infestation were collected from field sites across southern Africa and Australia. Wasps were extracted from these grasses either by dissection using a fine scalpel or were placed in standard insect emergence chambers. At least one representative specimen was included in the current study from all 16 grass species from which Tetramesa specimens have been collected in South Africa, to date (see Supplementary material Table S1). Five representative Tetramesa specimens were also included that were collected from E. curvula plants sampled in Australia (Table S1). The Tetramesa COI dataset (n = 115 sequences) included five wasp sequences collected from E. curvula in Australia, 37 sequences collected from four populations of E. curvula collected across South Africa, 31 sequences collected from seven congeneric Eragrostis grasses across South Africa, and 38 sequences from non-Eragrostis grasses sampled across southern Africa.
Genomic DNA extractions were performed using a PureLink™ Genomic DNA Mini Kit (ThermoFisher Scientific, catalogue number K182002) according to the manufacturer's protocols. Mitochondrial cytochrome c oxidase I (COI) sequences were obtained using the forward FWPTF1 (5' CCT GGTTCTTTRATTGGTAATGATC 3') and reverse LepR1 (5' TAAACTTC TGGATGTCCAAAAA 3') primers (Li et al. 2010). Successfully amplified PCR products were sent to Macrogen Inc. in the Netherlands for post-PCR purification and sequencing in the reverse direction. The reverse direction was chosen to avoid the poly-T region at the beginning of the forward sequence that interrupts downstream sequencing. Sequences were subsequently reverse complemented and aligned online using the default parameters in MAFFT v7 (https://mafft.cbrc.jp/alignment/server/) (Katoh et al. 2019).

Results
The Bayesian phylogeny indicated that the Tetramesa collected from all the Eragrostis grasses, in both South Africa and Australia, form a well-supported clade that is sister to undescribed Tetramesa specimens collected from other African grass genera such as Sporobolus, Eustachys, Hyparrhenia and Andropogon, as well as the only two described Tetramesa species with COI sequences available online, namely: Tetramesa romana and Tetramesa bambusae Philips (Figure 3). All the Australian Tetramesa specimens collected from E. curvula formed a well-supported clade (Clade E) with wasps collected from E. curvula in the Free State Province of South Africa (Figure 3).

Discussion
The CO1 phylogeny clearly indicated that the Tetramesa on E. curvula in Australia is of African origin and was most likely introduced along with its host plant. This introduction could have occurred as early as the 1920's (E. curvula was recorded as naturalized in New South Wales, Victoria and South Australia by 1920; Australasian Virtual Herbarium 2023). It is possible that undiscovered Tetramesa are associated with Australian Eragrostis spp., but these would most likely be genetically distinct from African species and would certainly have fallen outside of any E. curvula clade of wasps from South Africa. This would not be the first time a Tetramesa species has been introduced with its host plant into a non-native geographic region. The Arundo wasp T. romana has been recorded on its host plant Arundo donax L. in several countries around the world where A. donax has been introduced, such as the USA and South Africa (Goolsby and Moran 2009;Canavan et al. 2014). Although damaging genotypes of T. romana were introduced as a biological control agent to the USA, an adventive genotype of the same wasp species was already present prior to these introductions (Dudley et al. 2006). Tetramesa species typically overwinter inside the grass tiller, and can emerge from tillers that have been excised from the grass tussocks for at least four months (Sutton pers. obs.), which would allow them to survive being transported to Australia from South Africa as live plants, packing material, fodder or hay bales, as a possible modes of introduction.
The Tetramesa sp. found in Australia has been reared from at least six native Eragrostis species under field conditions in South Africa and produce F 1 progeny on at least two native African Eragrostis (E. plana and E. planiculmis) under controlled greenhouse no-choice conditions (Sutton et al. unpublished). The host specificity of this Tetramesa sp. therefore appears to be confined to a number of closely-related Eragrostis species, which is in contrast to the majority of Tetramesa species that are typically monophagous (Goolsby and Moran 2009;Sutton et al. 2019Sutton et al. , 2021. There are more than 70 native Eragrostis species in Australia, several of which, such as E. pilosa (L.) P. Beauv., E. leptocarpa Benth., and E. leptostachya (R. Br.) Steud, amongst others, are as closely (or more closely) phylogenetically related to E. curvula than either E. plana or E. planiculmis (Barrett et al. 2020). Given that insect herbivore host range is typically strongly linked to phylogenetic distance between the host plants (Wapshere 1974), it is possible that several Australian native Eragrostis species may be at risk of attack by this Tetramesa sp. Field surveys of the native Eragrostis species that are closely related to E. curvula should be conducted across Australia to determine how widespread the wasp is and if it is impacting native Eragrostis populations. In the meantime, the Tetramesa sp. recorded on E. curvula in Australia has been reported to the Australian Chief Plant Protection Officer and the Consultative Committee on Emergency Plant Pests.
The Tetramesa sp. is also likely to be providing benefits in Australia by damaging the invasive alien populations of E. curvula. These impacts need to be quantified but, given that the most promising candidate agent has been discovered in the invaded distribution, the need to release this agent may have been negated. The relative impacts of the South African Tetramesa populations that were being considered for biological control could be compared to the impacts of the Australian Tetramesa populations to determine whether more damaging genotypes should be considered for release, as was the case for the biological control of A. donax using T. romana in the USA (Goolsby and Moran 2009). It would, however, be important to complete host-specificity testing of the South African Tetramesa populations before they are considered for release because the preliminary results do suggest that the host range of the insect may be too broad for consideration for release in Australia.