There has been an increase in newly-emerged insect pests in recent years (Roques 2010) that can have a negative ecological and economic impact (Colautti et al. 2006, Kenis et al. 2009, Pimentel et al. 2001, Pimentel et al. 2005). There is an urgent need for the development of advanced tools for the early detection and accurate identification of new or emerging insect pests. One major problem is that many interceptions of newly-emerging insect pests are immature stages that are difficult or impossible to rapidly identify to species level (Wu et al. 2017). DNA barcoding is a tool for species identification, based on the use of a fragment of the cytochrome C oxidase 1 (COI) gene (Hebert et al. 2003). The divergence rate of DNA barcodes makes it possible to discriminate species for the vast majority of insects, which provides effective support for identification of individuals at the species level (Hebert et al. 2003). Although it has been criticised for low precision in certain cases (Moritz and Cicero 2004, Meyer and Paulay 2005, Song et al. 2008), its broad use makes it possible to assign an individual to a species regardless of its phenotype or the developmental stage or the state of the specimen collected. It effectively supplements taxonomic studies, based only on morphological criteria (Hajibabaei et al. 2007), provided that the data deposited in libraries are reliable and well curated (Kelnarova et al. 2018). DNA barcoding is, therefore, often used as a supporting tool to identify species or stages which are morphologically difficult to distinguish, especially in a context of invertebrate pest management. It can contribute to more rapid identification of insect pests (Ball and Armstrong 2008, Hodgetts et al. 2016, Wu et al. 2017) and has successfully been used as a tool for their bio-monitoring (Frewin 2013, Ashfaq and Hebert 2016). The online data base Barcode of Life Data Systems (BOLD, www.boldsystems.org) allows the automatic identification of species, based on their DNA barcodes (Ratnasingham and Hebert 2007, Ratnasingham and Hebert 2013). The development of the Barcode Index Number System (BINs) in BOLD further allows the automatic assignment of barcode sequences to genetic clusters, generating a web page for each cluster. Since clusters show high concordance with species (Lopez-Vaamonde et al. 2021), this system can be used as a proxy for species when taxonomic information is missing.

Cerambycidae are forest insects that play a major role in the decomposition of dead wood. Some species in this family also cause damage to a wide range of economically-important tree species (Haack et al. 2010, Eyre and Haack 2017) and can be economic pests of commercially-grown crops (Wang 2017). Cerambycids are uncommon pests of sugarcane crops (Carnegie and Conlong 1994), but some species have been found to cause severe damage in fields worldwide (Oyafuso et al. 2002,Mukunthan and Nirmala 2002, Dolinski et al. 2006). For example, in Thailand, the cerambycid Dorysthenes buqueti (Guérin-Méneville, 1844) showed a ten-fold population increase in sugarcane crops within a year causing significant damage (Pliansinchai et al. 2007). The Prioninae species Cacosceles newmannii (Thomson, 1877) is another recent example of a sugarcane pest. This species is native to Mozambique, Eswatini and South Africa and its biology has been poorly studied (Ferreira 1980; but see recent focused research efforts in, for example, Javal et al. 2018, Javal et al. 2019a, Javal et al. 2019b, Smit et al. 2021a, Smit et al. 2021b, Lehmann et al. 2021). The larval stage is thought to last for two years, during which larvae actively feed on root material and stem tissue (Way et al. 2017). Adults, on the other hand, have a very short life span of a few months and do not feed (Javal & Conlong, personal observations). Its host range has not yet been fully determined, even though preliminary studies show that the species is likely to be polyphagous (Smit et al. 2021b). Unlike many other Prioninae species, however, C. newmannii larvae are able to feed on living tissue and were found for the first time in 2015 in commercially-grown sugarcane in the KwaZulu-Natal (KZN) Province of South Africa. The reasons and the mechanisms underlying the rapid emergence of C. newmannii from its indigenous host plants on to sugarcane remain unclear (Javal et al. 2018). Larvae dig galleries into the sugarcane stool, but are, most of the time, found in the below-ground section of the sugarcane stalks. They cause severe crop damage, resulting in ongoing significant economic loss for the growers (Way et al. 2017) and research on biocontrol has shown limited success to date (Javal et al. 2019a). Bio-monitoring of this species is currently done by field surveys, since no lure has yet been found to efficiently monitor adults. Identification of trapped insects can be complicated due to several factors. Firstly, surgarcane agrosystems are usually made of sugarcane fields and natural non-cultivated fields or native forest, which increases the probability to catch a wide variety of species, both from the surrounding natural environment and from the cultivated crops. Secondly, the stage causing the damage is the larva, whose identification, based on morphological features, is very difficult and, therefore, relies heavily on molecular analysis. Finally, the genus Cacosceles includes four species that are all distributed in the Afrotropical Region and very difficult to distinguish morphologically by non-specialists, even at the adult stage (Ferreira 1980).

DNA barcoding has been used to accurately identify Cerambycidae pest species (Hodgetts et al. 2016, Wu et al. 2017, Kelnarova et al. 2018), as well as for the recognition of new taxa (Bouyer 2010, Bouyer 2016, Drumont and Ripaille 2019), but the subfamily Prioninae has not received enough attention so far, especially in tropical areas (Jin et al. 2020). The subfamily Prioninae includes more than 1,000 species in about 300 genera (Svacha and Lawrence 2014). They occur mainly in tropical and subtropical areas, but can be found worldwide (Monné et al. 2016). The main aims of this study are: 1) to generate DNA barcodes of the emerging pest C. newmannii to facilitate its identification by plant protection officers; 2) to compile a DNA barcoding reference library of other Afrotropical Prioninae to assess the validity of the Barcode Index Number system as a tool to identify them reliably; 3) to assess the level of deep intraspecific lineages that could suggest the existence of cryptic species.

Specimen sampling

Morphological identification

DNA barcoding

Sequence analysis

The resulting sequences, along with the voucher data, images and trace files from Sanger and SMRT sequencing are deposited in the Barcode of Life Database (BOLD, http://www.boldsystems.org) (Ratnasingham and Hebert 2007) and the sequences subsequently deposited in GenBank. All data are available from BOLD at: dx.doi.org/10.5883/DS-AFROPRIO

We generated a total of 70 COI sequences representing 20 named species (55 sequences identified to species level and remaining sequences identified to genus level) from 16 genera (results on 7 April 2021; Suppl. materials 1, 4). Genetic variability within species ranged from 0 to 14.34% with a mean of 2% (Suppl. material 2). The highest value (14.34%) was observed between two Mallodon downesi (Hope, 1843; Suppl. material 3). For C. newmannii, genetic distances varied between 0% and 4.8% with a mean of 1.5% (Suppl. material 3).

Sequences were assigned to 42 different BINs, 28 of which were new to BOLD (Suppl. material 4). Out of the 20 named species barcoded, 11 (55%) had their own unique BIN (Suppl. material 5). Seven species (35%) showed deep COI splits with several BINs (Suppl. material 5). The multiple BINs per species ranged from two to a maximum of five within Mallodon downesi (Suppl. material 5). Anomotoma wilwerthi (Lameere, 1903) showed a "shallow" COI split of only 1.61% between two different BINS (Suppl. materials 5, 6). Cacosceles newmannii showed two distinct BINs (BOLD:AEF3555 & BOLD:AEF6074) that diverged between 4.46 and 4.80% (Suppl. materials 5, 6). Out of 21 specimens of C. newmannii collected, 20 were sampled in the KZN Province of South Africa and only one in the Western Cape (WCape) Province. Specimens belonging to both BINs have been found sympatrically at Eshowe (South Africa) attacking sugarcane plantations (Suppl. material 1). Four different haplotypes were identifed in C. newmannii (Fig. 1, Table 2).

Figure 1.  

Minimum spanning haplotype network for C. newmannii specimens. Mutational steps are symbolised by dashes and the diameter of the circles is proportional to the number of individuals that belong to each haplotype. Specimens belonging to each haplotypes are listed in Table 2.

One Cacosceles specimen (MJ0009), collected in Eastern Cape, is morphologically similar to C. oedipus, but forms its own BIN (BOLD:AEF3200) and shows a high genetic divergence of 12.52% from C. newmannii (Suppl. material 5). More specimens are needed to assess its status and it is, therefore, treated here as an unconfirmed candidate species (sensu Lopez-Vaamonde et al. 2021).

DNA barcoding is a major tool in the bio-surveillance of insect pests. It allows rapid identification of an unknown specimen regardless of its developmental stage or state of preservation. DNA barcoding has been extensively used with Cerambycidae for pest diagnostics (Kelnarova et al. 2018) and biosecurity (Hodgetts et al. 2016, Wu et al. 2017). However, reliable molecular identification relies on comprehensive and well-curated databases, that are generally lacking for Prioninae. This study and the resulting barcode database will allow biosecurity officers to easily distinguish any life stages of C. newmannii specimens from other common Prioninae species, especially since, based on our data, C. newmannii ’s two BINS (BOLD:AEF3555 & BOLD:AEF6074) form a clade.

Interspecific and intraspecific genetic distances observed on this mitochondrial fragment were generally consistent with the currently accepted taxonomy and classification for most species of this subfamily (Ferreira 1980). The genus Cacosceles was more intensely sampled. As currently accepted, the genus Cacosceles contains four species: C. newmannii Thomson, 1877; C. oedipus Newman, 1838, C. latus Waterhouse, 1881 and C. gracilis Lackerbeck, 2000. C. oedipus was clearly distinct from C. newmannii, based on barcode sequences which confirms their clear species' boundaries. The two species showed 18.35% divergence, based on the barcode fragment. We have sequenced a specimen of a Cacosceles from Eastern Cape (MJ0009) that is morphologically similar to C. oedipus. This specimen is, however, 20% divergent from the C. oedipus sequenced from the Western Cape (MJ0005 & MJ0008). More specimens are needed to assess the status of these divergent BINS. Two Cacosceles species, C. latus and C. gracilis, have not been barcoded yet and a formal taxonomic revision of this genus is needed.

Regarding C. newmannii, the distance observed between the two BINS described is above the commonly used threshold of interspecific distance in arthropods (3%; Hebert et al. 2004). Interestingly, the single specimen from the Western Cape Province fell into one of these two clusters. Morphological observations of the specimens, belonging to these clusters, revealed no apparent distinguishing characters. In addition, the specimens of the two clusters from KZN are found in sympatry in infested fields where the species was found feeding on sugarcane. A more detailed study, including more specimens and nuclear genes, is needed to establish if this corresponds to sympatric deep splits and despeciation, lineage fusion events (Hinojosa et al. 2019) or mitochondrial artifacts, such as Wolbachia infection.

An extreme case of deep splits within one species is Mallodon downesii with five BINs out of six specimens barcoded. M. downesii has a very broad distribution across the African continent and is highly polyphagous (Delahaye and Tavakilian 2009). It has also been recorded as an alien introduction to the neotropics, representing a potential danger as a pest of coffee plantations (Wappes et al. 2018). Our barcoding results suggest that it could represent a complex of species that deserves further investigation.

More generally, this study revealed that multiple species of southern African Prioninae show deep intraspecific barcode distances that could not yet be associated with morphological divergences. It should be noted that forests have experienced strong contraction events during past climate oscillations in this region, itself strongly affecting the genetic structure and the speciation of forest insect species (Fabrizi et al. 2019). It is hypothesised that such divergences could originate from a complex evolutionary history of Prioninae, resulting from contraction/extension of their habitat over geological time.

To complement existing identification tools based on morphology, this barcode database will facilitate clear identification of the most common species of southern African Prioninae, irrespective of taxonomic skills of the observer or developmental stage of the insect. Further analyses using nuclear markers should be carried out to assess the multiple BINs and substantial intraspecific genetic divergences observed within four of the named species of Prioninae found to have multiple BINs. Particular attention should be focused on the pest species C. newmannii in order to clarify the species delimitation of the two BINs identified.

MJ thanks Natasha Govender (Durban Natural Science Museum) for allowing access to museum collections. CLV thanks Brian Fisher, Balsama Rajemison and other staff at the Madagascar Biodiversity Center for their assistance during fieldwork in Madagascar. The authors are grateful to Evgeny Zakharov, Jeremy deWaard, Claudia Steinke, Miduna Rahulan and other staff at the Centre for Biodiversity Genomics for their assistance. Thanks to Claude Ripaille and Thierry Bouyer for their help identifying some species. Thanks to Thibaud Decaens for allowing us to use four unpublished barcodes of Gabonese species. Thanks to Philippe Jeanmart of Precious Woods - Compagnie Equatoriale des Bois for permission to CLV to carry out fieldwork in their logging concession at Bambidie Site (Lastourville, Gabon). Relevant collecting permits are as follows: 32/17/MEEF/SG/DGF/DSAP/SCB. Re du 24 Février 2017; 243/17-MEEF/SG/DGF/DSAP/SCB. Re du 6 Octobre 2017; 010/18/MEEF/SG/DGF/DSAP/SCB. Re du 26 Janvier 2018 and 049/18/MEEF/SG/DGF/DSAP/SCB. Re du 15 Février 2018 enabling collections in reserves and other protected areas in Madagascar. The director of MICET at Antananarivo and their staff, including drivers, are thanked for facilitating these permits, export and logistics. Permit N°AR0050 /18/MESRS/CENAREST/CG/CST/CSAR (to CLV) enabled collections at "Compagnie Equatoriale des Bois" at Lastourville (Province of l'Ogooué Lolo) and AR 0008/12/MENESRIC/CENAREST/CG/CST/CSAR and AR 0002/13/MENESTFRSJS/ CENAREST/CG/CST/CSAR to collect at Parc National de la Lopé (Gabon) during ECOTROP field school 2012 & 2013. Thanks to Pr Daniel Franck Idiata and his staff at CENAREST for facilitating these permits, export and logistics. Permit OP1382-2019 (Ezemvelo KZN Wildlife Permits Office, Collecting Permit) allowed collections in the KwaZulu-Natal Province of South Africa (to JH) and permit ZA/LP/83364 allowed collections in Limpopo (South Africa) in 2017. We thank the editor Chen-Bin-Wang and both reviewers Natalia Kirichenko and Francesco Vitali for their helpful comments on the manuscript.

DNA barcoding was funded by grants from the South African Sugarcane Research Institute. CLV thanks ANR project SPHINX (ANR-16-CE02-0011-01) for funding fieldwork in Gabon and South Africa and the Critical Ecosystem Partnership Fund project IPSIO (A Network of Interdisciplinary Researchers Committed to Training, Sharing Tools, and Advocating for an Insect Focused Approach in Conservation) for funding fieldwork in Madagascar.

Contributed to: • Study design: MJ, JST, JMH • Specimen sampling, databasing: MJ, CLV, DEC, ND, EG, JMH • Sequence analyses: MJ, CLV, JMH • Taxonomic expertise, result validation: JMH, CLV, ND, EG • Writing of manuscript: MJ, CLV, JMH • All authors edited and commented to the manuscript. • JMH and CLV contributed equally to the work.

The authors declare no conflicts of interest.