The groundnut leaf miner (GLM) was initially reported in Africa in 1998 (
Epieru, 2004) and presumed to be an incursion of the Asian
Aproaerema modicella (Deventer) (Lepidoptera: Gelechiidae), a main pest of several legume crops in that continent
(Murugasridevi et al., 2022; Kenis and Cugala, 2006). This notion was based on the morphological similarities of the moth, larval feeding patterns on groundnut (
Arachis hypogaea) and soybean (
Glycine max) reported in Asia, Africa and Australia
(Buthelezi et al., 2021; Kenis and Cugala, 2006). Furthermore, anatomic similarity in groundnut leaf miner male genitalia (
Van der Walt et al., 2008), successful capture of the GLM moths in Australia using sex pheromone blends of
A. modicella (
Buthelezi et al., 2016; 2012) reinforced this assumption. Consequently,
Buthelezi (2015) tentatively synonymized GLM populations from Africa, Australia and India as
Bilobata subsecivella.
Subsequent mitochondrial DNA analyses using cytochrome oxidase I (COI), cytochrome oxidase II (COII) and cytochrome b (cytb) gene regions later revealed that GLM populations from Africa, Australia and India are closely related, with the Australian population showing greater genetic diversity
(Buthelezi et al., 2024; Buthelezi, 2015). To extend this work, two nuclear DNA regions, 28S ribosomal DNA (28S rDNA) and elongation factor-1 alpha (EF-1 ALPHA) were analysed using 41 specimens obtained from four sites in South Africa, three from one site each in India and Australia and three from two sites in Mozambique (
Buthelezi, 2015).
The ambiguity surrounding the taxonomy of GLM poses a significant challenge for the development of sustainable management strategies, especially in Africa where the pest threatens the production of groundnut and soybean (
Buthelezi and Zharare, 2025;
Buthelezi et al., 2021). A major research gap is the limited molecular evidence required to accurately determine the species level genomic connection among GLM populations in Australia, Asia and Africa. This calls for extensive molecular studies, including whole-genome sequencing of GLM populations from Africa, Australia and Asia, complemented by detailed morphological analyses. Therefore, this study aimed to investigate phylogenetic relationship of the GLM populations from Africa, Australia and India by analysing nuclear DNA regions, 28S rDNA and EF-1 ALPHA, thereby evaluating whether these populations are conspecific or represent the distinct cryptic species.
Specimen collection
In South Africa, GLM specimens were collected from four sites from groundnut, soybean, lucerne (
Medicago sativa) and pigeon pea (
Cajanus cajan) at Manguzi (27
o35’5.32” S, 33
o22’3.56” E), Nelspruit (25
o45’4.52” S, 31
o37’1.54” E), Vaalharts (22
o34’33” S, 35
o10’06” E) and Brits (25
o59’ 1.35” S, 28
o16’8.75” E). In Mozambique, they were collected from groundnuts at Pambara Produtor (22
o40’24” S, 35
o15’13” E) and Pambara EP1 (22
o34’33” S, 35
o10’ 06” E). In India, they were collected from groundnut at Hyderabad- ICRISAT (17
o21’57" N 78
o28’33" E). Australian specimens constituted moths which were caught using pheromone-baited traps placed in natural vegetation containing wild soybean in Brisbane (27
o32’08” S 152
o51’35” E) (
Buthelezi, 2015). All samples were preserved in 10-ml polycarbonate vials with absolute ethanol. The vials were securely closed with plastic snap-on lids sealed with press-on plastic lids and stored in a -80
oC freezer preceding to sequencing.
DNA amplification and analyses
Amplification of DNA by Polymerase Chain Reaction (PCR) was conducted following the procedure described by
Buthelezi (2015). Two primers, Ron and Nancy, designed by Inqaba Biotech Industries were used for DNA amplification and are detailed in Table 1. Each PCR condition contained 1x KAPA Robust Ready Mix (KAPA Biotech), 1x Enhancer A, 0.4 μM of each primer and 20ng of genomic DNA. Amplification was carried out using a Verity PCR thermal cycler (Applied Biosystems). The thermal cycling programme comprised an initial denaturation at 95
oC for 5 min, followed by 40 cycles of 95
oC at 30s, 55
oC at 60s and 72
oC for 90s. A final extension step was performed at 72
oC for 10 min. Following amplification, products were purified using the NucleoFast Purification System (Separations). Sequencing was performed using the respective primers and BigDye Terminator V1.3 (Applied Biosystems). The sequencing products were separated by electrophoresis using a 3730xl DNA Analyser (Applied Biosystems). Sequence data were subsequently analysed with Sequencing Analysis Version 5.3.1 software (Applied Biosystems).
DNA sequence editing
DNA sequences were inspected and corrected manually for base calling and errors, followed by pruning and alignment using the ClustalW algorithm implemented in the BioEdit Sequence Alignment Editor (
Hall, 1999). Consensus sequences were subsequently generated and exported in FASTS format for further analysis in MEGA5 (
Hall, 1999).
Sequence analyses and phylogenetic reconstruction
Maximum Parsimony (MP) and Neighbour Joining (NJ) using PAUP* 4.0b10 were employed to reconstruct phylogenetic analyses (
Swofford, 2003). The tree-bisection- reconnection (TBR) algorithm and bootstrap re-sampling (1000 iterations) were applied. The nucleotide substitution models were GTR+I* for 28S and HKY+I for EF-1 ALPHA (
Posada and Crandall, 1998). Genetic relatedness was calculated using PAUP* 4.0b10 (
Swofford, 2003).
Nuclear ribosomal 28S gene
The phylogenetic tree (Fig 1) was inferred from 180 nucleotides of the 28S rDNA gene, with
Trichoplusia ni (Hübner) (Lepidoptera: Noctuidae) and
Hemileucua sp. L6 (Walker) (Lepidoptera: Saturniidae), obtained from the NCBI gene bank, serving as outgroups. All 41 sequences from the South African samples, together with three sequences each from Mozambique (MOZA G B, MOZA B C and MOZA G D), Australia (GLM 1, GLM 2 and GLM 3) and India (SAT 1, SAT 2 and SAT 3), clustered within a major single clade. The clade was strongly supported by bootstrap value of 100%. The finding furthermore displays no genetic variety between GLM populations from Australia, India, Mozambique and South Africa. These hints restricted genetic variation among GLM populations from these regions based on this marker. The rDNA 28S gene is recognized as a robust tool for solving deep phylogenetic relationships within the Acari (
Cruickshank, 2002). These findings align with those reported by
Buthelezi et al. (2012), which demonstrated 100% similarity between sequences from Mozambique and South Africa and those of
A. simplexella retrieved from the BOLD GenBank.
Elongation factor-1 alpha (EF-1 ALPHA) gene
The phylogenetic tree was inferred from 209 nucleotides of the EF-1 ALPHA region (Fig 2). No suitable reference sequences were available in the gene bank for use as out-groups; therefore, the phylogenetic analysis was conducted without an outgroup. The phylogenetic tree demonstrates three distinct clusters. The largest incorporates 37 sequences from South Africa, together with all three sequences from Mozambique (MOZA G B, MOZA G C and MOZA G D) and India (SAT 1, SAT 2 and SAT 3). This cluster accounted for 93% of all sequences and received bootstrap support of 63-64%. The second cluster contained a single sequence from Australia (GLM 1), whereas the third comprised the two other Australian sequences (GLM 2 and GLM 3). Genetic pairwise distances amongst the clusters varied from 0.00% to 6.99% (Table 2). The tree reveals underlying differences, particularly within Australian population. These results are constant with the results of
Buthelezi et al. (2024), which revealed that mtDNA regions (COI) sequences from some Australian populations grouped together with GLM sequences from India, South Africa and Mozambique, whereas in COII and cytb, the Australian sequences formed separate groups from the other populations. This pattern underscores the need for multi-locus approaches, combining mitochondrial and nuclear markers, to better resolve population invasion histories
(Sario et al., 2024). Moreover, integrating morphological and molecular approaches enhances the understanding of genetic diversity, as shown in
Phaseolus vulgaris L. (
Del Rosario-Arellano et al., 2025;
Quat et al., 2025). Furthermore, genome-wide sequencing could clarify whether the observed divergence in Australian populations reflects long-term isolation, multiple introductions, or cryptic speciation information critical for both taxonomy and the development of effective pest management strategies.
In addition, the observed genetic diversity in Australian GLM populations, coupled with their restriction to soybean (whereas African and Indian populations infest both groundnut and soybean), highlights the need to investigate the genetic basis of host preference. Understanding these differences could provide valuable insights into host–pest interactions and inform breeding programs aimed at developing resistant groundnut and soybean varieties.