Alterations and Interchange of Morphometric Characters in Different Life Cycle Stages with Reference to Genomic Variations of Anopheles subpictus (Diptera; Culicidae) Sibling Species Complex in Sri Lanka

The species complex of the mosquito Anopheles subpictus is designated by the sibling species A–D, depending on morphological characters of life cycle stages and variations in polytene chromosomes. However, morphological aberrations in the life cycle stages make the identification of sibling species uncertain and imprecise. The objective of the present study is to determine the suitability of morphological variations of sibling species and their genomic variations to identify the sibling species status of an An. subpictus population in Sri Lanka. Life cycle stages of larvae, pupal exuviae, and adults were examined for previously reported distinctive morphological features. Five nuclear and mitochondrial genome regions, including the Internal transcribed spacer 2 (ITS2) region, D3 region, white gene, cytochrome c oxidase I (COI), and Cytochrome b (Cyt-b), were sequenced and analyzed for variations. The eggs changed their distinct sibling morphological characters during metamorphosis (89.33%). The larvae, pupal exuviae, and adult stages showed deviation from their sibling characters by 26.10%, 19.71%, and 15.87%, respectively. However, all the species from the analysis shared two distinct sequence types for all regions, regardless of the morphological variations. In conclusion, the An. subpictus sibling species complex in Sri Lanka is not identifiable using morphological characters due to variations, and the genomic variations are independent from the morphological variations.


Introduction
The Anopheles (Cellia) subpictus sensu lato (s.l.) Grassi 1899 species complex is the most abundant anopheline mosquito in the Indian subcontinent [1,2]. It acts as a vector for malaria and Japanese encephalitis in many parts of Asia [2]. An. subpictus is the major secondary vector of malaria in Sri Lanka [1,3].
In India, An. subpictus was first deemed to be a species complex based on differences in larval morphology [4]. Further, the occurrence of two distinct types of eggs and cytological evidence has temporarily designated two forms of An. subpictus sibling species as A and B in India [5]. This taxon has further been categorized as a complex of four sibling species-designated as A, B, C, and D-based Insects 2018, 9,89 2 of 11 on stage-specific morphometric characters [6]. The presence of four sibling species (A-D) has been confirmed through fixed inversions in the X-arm of polytene chromosomes viz. A = X+ a ,+ b ; species B = Xa,b; species C = Xa,+ b ; species D = X + a, b [6]. In Sri Lanka, the existence of sibling species A and B was first reported by Abhayawardana et al. [7], based on the single inversion (X+ a /X a ) on the X chromosome. No other cytotaxonomic studies have been carried out yet to identify all the members of the An. subpictus complex in Sri Lanka. However, all four sibling species (A-D) have been reported to occur in Sri Lanka [8,9] based on identification using the morphometric characteristics described by Suguna et al. [6]. The method of Suguna et al. [6] was confirmed by Singh et al. [10], except for the egg morphology, and it was further concluded that a single identification character in a randomly picked individual from a population of any life stage could be used for identification of the An. subpictus sibling species status.
Although An. subpictus was described and designated as A-D, intraspecific variations within the taxon were first reported in India, among urban, pre-urban, and rural populations [11]. Kirti and Kaur [12] observed morphological differences in wings and palpi of An. subpictus in India. Further, a different set of morphological variations (not reported in Suguna et al. [6]), mainly in proboscis and palpi of An. subpictus, was reported in Sri Lanka [13][14][15]. Furthermore, the An. subpictus complex has been designated as only A and B in Sri Lanka based on ITS2 and cytochrome c oxidase I (COI) sequence polymorphism [16], showing the complexity of sibling species identification of the An. subpictus population in Sri Lanka.
Therefore, the current study was carried out to investigate the suitability of morphological character variations and genomic variations for the identification of the An. subpictus population in Sri Lanka.

Morphological Characterization of An. subpictus Life Cycle Stages
The eggs laid by the collected adult mosquitoes were separately transferred to rearing basins and reared to obtain F1 progeny. Life cycle stages of eggs, fourth instar larvae, pupal exuviae, and F1 adults were examined from 10 randomly selected individuals in each rearing basin for the characteristic features to designate the specimen as sibling species A, B, C, or D. The observed morphological characters were mesothoracic seta 4 of larvae, seta 7-I of pupal exuviae, and

Morphological Characterization of An. subpictus Life Cycle Stages
The eggs laid by the collected adult mosquitoes were separately transferred to rearing basins and reared to obtain F 1 progeny. Life cycle stages of eggs, fourth instar larvae, pupal exuviae, and F 1 adults were examined from 10 randomly selected individuals in each rearing basin for the characteristic features to designate the specimen as sibling species A, B, C, or D. The observed morphological characters were mesothoracic seta 4 of larvae, seta 7-I of pupal exuviae, and proportions of apical pale and pre-apical dark bands in palpi of adults. The randomly collected larvae, pupae, and adults from the wild were examined for the morphological variations specified by Suguna et al. [6]. Individual eggs were separated into rearing basins and examined for the respective features in larvae, pupae, and adult stages.

Results
The wild-caught adult mosquitoes consisted of all sibling species A, B, C, and D in the studied population in Sri Lanka, based on the palpi polymorphism. Additionally, 66 individuals (23.0%) of the studied population showed aberrant palpi formations which could not be categorized as belonging to any of the sibling species. Percentages of each sibling species and aberrant types are shown in Table 2. Among all the observed larvae (n = 567), those with doubly branched mesothoracic seta 4 with a long stem were categorized as sibling species A (n = 218), while those with doubly branched mesothoracic seta 4 with a short stem were identified as sibling B (n = 67). The larvae which had triply branched mesothoracic seta 4 with the branching occurring at the same point and at two different points were classified as sibling C (n = 66) and D, respectively (n = 68). Larval mesothoracic seta IV of sibling A, B, C, and D are shown in Figure 2. One hundred and thirty-nine (139) individuals showed two different types of mesothoracic seta 4 in the left and right sides of the mesothorax. These variations have given rise to two different sibling species statuses for both the left and right sides of

Results
The wild-caught adult mosquitoes consisted of all sibling species A, B, C, and D in the studied population in Sri Lanka, based on the palpi polymorphism. Additionally, 66 individuals (23.0%) of the studied population showed aberrant palpi formations which could not be categorized as belonging to any of the sibling species. Percentages of each sibling species and aberrant types are shown in Table 2.

Results
The wild-caught adult mosquitoes consisted of all sibling species A, B, C, and D in the studied population in Sri Lanka, based on the palpi polymorphism. Additionally, 66 individuals (23.0%) of the studied population showed aberrant palpi formations which could not be categorized as belonging to any of the sibling species. Percentages of each sibling species and aberrant types are shown in Table 2.

Results
The wild-caught adult mosquitoes consisted of all sibling species A, B, C, and D in the studied population in Sri Lanka, based on the palpi polymorphism. Additionally, 66 individuals (23.0%) of the studied population showed aberrant palpi formations which could not be categorized as belonging to any of the sibling species. Percentages of each sibling species and aberrant types are shown in Table 2.

DNA Sequence Analysis
The sequences were assembled using the program DNA Baser v3.5.3 and were blasted over the NCBI GenBank to confirm the mosquito origin of the sequences. Open reading frames were found for the white gene and mitochondrial gene coding sequences and confirmed for the absence of stop codons. Sequences of each genomic region were aligned using ClustalW multiple alignment to observe the sequence polymorphism. Haplotype frequency was analyzed using DNASP 5.10.01 software. All the sequences were deposited in the NCBI GenBank.

Results
The wild-caught adult mosquitoes consisted of all sibling species A, B, C, and D in the studied population in Sri Lanka, based on the palpi polymorphism. Additionally, 66 individuals (23.0%) of the Insects 2018, 9, 89 4 of 11 studied population showed aberrant palpi formations which could not be categorized as belonging to any of the sibling species. Percentages of each sibling species and aberrant types are shown in Table 2.

Pupal Exuviae
From a total of 542 pupal exuviae observed, in 142, seta 7-I was as simple and long as hairs 6 and 9 (sibling A; n = 142). It was 4-5 branched and shorter than hairs 6 and 9 in 75 of the pupal exuviae (sibling B; n = 75). Seta 7-I was medium-sized and doubly branched in 123 (sibling C; n = 123), while it appeared medium-sized and triply branched in 92 of the pupal exuviae (sibling D; n = 92). Sixtysix pupal exuviae showed two different setal morphologies in the left and right sides. Seta 7-I was indistinguishable or absent from one of the sides in 44 pupal exuviae.

Adults
Sibling species designations in adults (n = 523) were based on morphological differences in female palpi. Sibling A was identified for those having an apical pale band longer than the pre-apical dark band in palpi (n = 195), while the apical pale band was shorter than the pre-apical dark band in palpi of sibling B (n = 77). In sibling species C, the proboscis was longer than the palpi, with the apical pale band length equal to the pre-apical dark band length of palpi (n = 75). For sibling D, the proboscis and palpi were of equal length, and the apical pale band length was equal to the pre-apical dark band length in palpi (n = 93). Eighty-three (83) adults showed variations to the above standard features (Table 3; Figure 3). The percentage deviation from existing morphological characteristics in each life cycle stage is given in Table 4.

Pupal Exuviae
From a total of 542 pupal exuviae observed, in 142, seta 7-I was as simple and long as hairs 6 and 9 (sibling A; n = 142). It was 4-5 branched and shorter than hairs 6 and 9 in 75 of the pupal exuviae (sibling B; n = 75). Seta 7-I was medium-sized and doubly branched in 123 (sibling C; n = 123), while it appeared medium-sized and triply branched in 92 of the pupal exuviae (sibling D; n = 92). Sixty-six pupal exuviae showed two different setal morphologies in the left and right sides. Seta 7-I was indistinguishable or absent from one of the sides in 44 pupal exuviae.

Adults
Sibling species designations in adults (n = 523) were based on morphological differences in female palpi. Sibling A was identified for those having an apical pale band longer than the pre-apical dark band in palpi (n = 195), while the apical pale band was shorter than the pre-apical dark band in palpi of sibling B (n = 77). In sibling species C, the proboscis was longer than the palpi, with the apical pale band length equal to the pre-apical dark band length of palpi (n = 75). For sibling D, the proboscis and palpi were of equal length, and the apical pale band length was equal to the pre-apical dark band length in palpi (n = 93). Eighty-three (83) adults showed variations to the above standard features (Table 3; Figure 3). The percentage deviation from existing morphological characteristics in each life cycle stage is given in Table 4.  sibling species A, B, C, and D (×200×2). Position of the setae is shown with an arrow. The shape of the setae is schematically shown [6] in the top-right corner of each figure.

Pupal Exuviae
From a total of 542 pupal exuviae observed, in 142, seta 7-I was as simple and long as hairs 6 and 9 (sibling A; n = 142). It was 4-5 branched and shorter than hairs 6 and 9 in 75 of the pupal exuviae (sibling B; n = 75). Seta 7-I was medium-sized and doubly branched in 123 (sibling C; n = 123), while it appeared medium-sized and triply branched in 92 of the pupal exuviae (sibling D; n = 92). Sixtysix pupal exuviae showed two different setal morphologies in the left and right sides. Seta 7-I was indistinguishable or absent from one of the sides in 44 pupal exuviae.

Adults
Sibling species designations in adults (n = 523) were based on morphological differences in female palpi. Sibling A was identified for those having an apical pale band longer than the pre-apical dark band in palpi (n = 195), while the apical pale band was shorter than the pre-apical dark band in palpi of sibling B (n = 77). In sibling species C, the proboscis was longer than the palpi, with the apical pale band length equal to the pre-apical dark band length of palpi (n = 75). For sibling D, the proboscis and palpi were of equal length, and the apical pale band length was equal to the pre-apical dark band length in palpi (n = 93). Eighty-three (83) adults showed variations to the above standard features (Table 3; Figure 3). The percentage deviation from existing morphological characteristics in each life cycle stage is given in Table 4.

Isofemale Progeny Observation for Sibling Species Status
Among the observed 46 separate egg clutches (10 individuals per egg clutch), none of the parental females produced a single type of sibling species. Twenty individuals (43.47%) from the egg clutches produced 100% of the parental sibling status for the observed 10 individuals. However, all

Isofemale Progeny Observation for Sibling Species Status
Among the observed 46 separate egg clutches (10 individuals per egg clutch), none of the parental females produced a single type of sibling species. Twenty individuals (43.47%) from the egg clutches produced 100% of the parental sibling status for the observed 10 individuals. However, all of these clusters showed different or a mixture of sibling characteristics in the larvae and pupal exuviae. The sibling species composition at each life stage for the observed 10 individuals from the egg clutches is shown in Table 5. Table 5. Morphological identification of sibling species status of parental female, larvae, pupal exuviae, and F 1 adults.

Sibling Species Status of Parental Female
Sibling

Examining Individual Eggs for Life Cycle Stages until Emergence
From the 98 individual eggs observed until emerging adults, only 8 individuals (10.67%) were consistent in the sibling status in consecutive larvae, pupae, and emerged adult stages. A percentage of 89.33% progeny showed different sibling characteristics in each life cycle stage. In the larvae and pupal exuviae, the occurrence of two different sibling features on the left and right sides of the body amounted to 33.33% and 26.66%, respectively (Table 6). Table 6. Morphological identification of sibling species status of larvae, pupae, and F 1 adults of individual eggs in five isofemale progenies.

Individual No. Larval Sibling Species Designation
Pupal Sibling Species Designation F 1 Adult Sibling Species Designation

DNA Polymorphism of the An. subpictus Population in Sri Lanka
Two length variations of the ITS2 region, 480 bp and 575 bp, were identified among 29 sequences. All shorter sequences (n = 15) were identical, while the longer sequences (n = 14) were polymorphic in nine positions. All sequences were identified as having three haplotypes, in which the longer sequences encompassed two haplotypes.
All other analyzed nuclear DNA regions (D3 and white gene) and mitochondrial DNA regions (COI, Cyt-b) also showed two types of nucleotide sequences for all observed morphological forms. The variations were independent from the type of morphological characteristics. Further, the same sequences that shared the shorter ITS2 sequence had identical D3, white gene, COI, and Cyt-b sequences. Similarly, the individuals that shared the longer ITS2 sequence were in congruence with the sequence similarity in all other analyzed regions. The number of haplotypes of each genomic region and the NCBI accession numbers of the sequences are shown in Table 7. Table 7. Number of haplotypes of each genomic region and the accession numbers received from the NCBI GenBank.

Genomic Region
No. of Haplotypes NCBI GenBank Accession Numbers

Discussion
The results suggest that the An. subpictus sibling species [6] are distributed in five study locations, which are dry (Puttalam, Batticaloa) and intermediate (Chilaw, Kurunegala, Monaragala) climatic zones in Sri Lanka. The initial screening of wild-caught adult mosquitoes from the five selected locations for their morphological characteristics to determine sibling species status shows that the distribution of morphological variations is independent of the study locations. Morphological variations previously designated in the An. subpictus sibling species [13][14][15] and additional variations observed in the current study were commonly found from all locations. Hence, further analysis was carried out by treating the samples collected from all locations as a single pooled An. subpictus population from Sri Lanka.
In the morphological examination of individuals, the egg morphology was excluded due to the time restriction until hatching. The eggs were hatching on the microscopic slide, prior to completing observation of the whole egg clutch. Further, it has been shown that the eggs do not serve as a good taxonomic feature due to phenotypic plasticity [24]. Therefore, only larvae, pupal exuviae, and adult morphology were considered for further analysis.
The current study shows that morphological variations of life cycle stages previously used for sibling species characterization are unreliable and polymorphic for the population found in Sri Lanka. The percentage of variation reported from the morphological identification keys of Suguna et al. [6] was 15.87-26.10% for all life cycle stages. The extended pre-apical dark bands and the additional dark spots at the tips of palpi were first-time reports of An. subpictus variations, which have not been reported in India or in Sri Lanka. Furthermore, deviations of the features on either the left or right side of the larvae and pupal exuviae were observed in the present study. However, no such one-sided variations were seen in the palpus of adults; instead, all variations were indicated in the pair of palpi. Parental sibling species characteristics were not consistent for the respective F 1 adults. Further, F 1 adults from a single egg clutch gave rise to a mixture of sibling species, as well as individuals with features deviating from A-D sibling species. Therefore, these findings justify the unreliability of the An. subpictus sibling species discrimination characters [6] for the identification of the An. subpictus population found in Sri Lanka. Moreover, this study does not agree with the findings of Singh et al. (2010) [10], who stated the possibility of discriminating a species by a single identification character in randomly picked individuals in a population of any life stage.
In the genomic region analysis, using nuclear and mitochondrial DNA regions, two types of sequences were reported for all analyzed DNA regions. A remarkable length variation (90 bp) was observed in the ITS2 region after annotation of the sequence using the ITS2 database. The occurrence of two types of sequences was independent from the morphological variations at any life cycle stage. This finding further confirms the unsuitability of existing morphological characterization methods [6] to discriminate the An. subpictus species complex found in Sri Lanka. Further, the haplotype analysis showed that the shorter and longer sequence forms of the ITS2 region are two distinct sequences which were similarly separated in all studied genomic regions.
Although India is closely related and shares a biodiversity in Western Ghats that is similar to Sri Lanka, there may have been independent and distinct evolutionary forces after the main land separation during the marine introgression in the Pleistocene period. Therefore, despite the close proximity of Sri Lanka and India, the evolutionary roots and relationships of Sri Lanka to other surrounding land masses-Madagascar, Africa, Australia, and the islands of Andaman-have to be considered. In the cases of the major malaria vector of Sri Lanka, the An. culicifacies species complex [25], and the major vector of leishmaniasis, the Ph. argentipus species complex [26,27], the initial reports on the similarity of their composition to Indian species were later shown to have variations and deviations. Therefore, it is essential to consider the taxonomy and evolutionary studies for such species found in Sri Lanka without considering India as a basis. Unbiased analysis of species in Sri Lanka may reveal more biologically and evolutionarily important information on species. Further, the evolutionary patterns exerted on the two countries may have variations based on the different geographic and environmental influences. Since the geography and topology of the two land masses are different, it can be concluded that the two countries have different independent evolutionary processes for these mosquito species.
Finally, although the variations in vectorial capacity have been reported for the sibling species of the major vector species An. culicifacies, such variations could not be observed among the sibling species of the secondary or minor vector species of malaria. Therefore, sibling species status and the relationship of morphological variations to sibling status of An. subpictus should be reassessed to delineate the composition and the diversity of the An. subpictus species complex found in Sri Lanka.

Conclusions
The An. subpictus population in Sri Lanka could not be identified as a sibling species using their previously reported morphological characteristics. In addition, the genomic variations found in the study are independent from the morphological variations of the An. subpictus population found in Sri Lanka.