INTRODUCTION

Species of Anopheles nuneztovari complex are important malaria vectors in several countries in South America^1,2 and, according to studies, it is considered a local vector in some areas of the Brazilian Amazon^3-6. The geographic distribution of this species is wide and can be found in Bolivia, Brazil (Amazon region), Colombia, Ecuador, Peru, French Guiana, Suriname, Venezuela, and also part of Panama in Central America^7-10.

An. nuneztovari s.l. Gabaldón, 1940 comprises four cryptic species that are difficult to distinguish or even indistinguishable based on external morphological criteria^7,11-13. Some of these species exhibit biological, ecological, and behavioral differences that influence malaria transmission and vector control¹⁴.

The species belonging to the Anopheles (Nyssorhynchus) nuneztovari complex are Anopheles nuneztovari s.s Gabaldón, 1940, Anopheles goeldii Rozeboom & Gabaldón, 1941, Anopheles dunhami Causey, 1945^15-18 and Anopheles jamariensis¹⁹, formerly known as Anopheles nuneztovari A ^17,18,20-23.

An. nuneztovari s.s, occurs in Colombia, Venezuela, and Brazil; An. goeldii and An. jamariensis in the Brazilian Amazon, and An. dunhami, in the Brazilian Amazon, Colombia and Peru^13,17,23-27.

Adult mosquitoes of Anopheles nuneztovari complex species can be identified using molecular methods and by the morphology of the male genitalia^{6,13,17-19,23}. Geometric morphometry, a tool that has proven efficient for the identification of medically important mosquitoes, is fast and presents quality results^28-36. Unlike classical morphometrics, which uses linear measurements (lengths, widths, and angles) in structures or parts of organisms for multivariate statistical analysis, geometric morphometrics uses Cartesian coordinates obtained from points distributed in a two- or three-dimensional plane, called homologous anatomical landmarks^37-39.

Anatomical landmarks can be categorized according to the degree of reliability regarding the homology of the points. They are classified into three types: type I (highly reliability) is used in bone structures or wing venation, type II (medium reliability) is used at the extremities of processes and type III (low reliability) is used for example at the top of a concavity^39-41. The wing is the part of the mosquito's body most used in morphometric analysis studies because it has a rigid two-dimensional structure and ribs where anatomical landmarks can be easily positioned, reducing scanning errors^42,43.

Thus, this study aimed to use wing geometric morphometrics as an auxiliary tool for differentiating species within the An. nuneztovari complex, providing information that may contribute to control measures for these malaria-transmitting vectors, given that females of some species from this complex do not present proven vector competence^19,27.

MATERIAL AND METHODS

Adult female specimens were collected using protected human attraction capture methods in the municipality of Altamira-PA (S 03º12'12", W 52º12'23) and by animal attraction (cattle and buffalo) in the municipality of Peixe Boi - PA (S 01° 09' 56.6"; W 047° 17' 60.6") with the aid of an oral suction aspirator (Castro aspirator) and flashlight, and stored in containers.

The specimens were identified under a stereoscopic microscope using the identification keys proposed by Gorham et al⁴⁴, Consoli & Oliveira⁹, and Forattini⁴⁵. The wings of female mosquitoes identified as An. nuneztovari s.l. were removed and subjected to geometric morphometric analysis.

For the morphometric analysis, the wings of female mosquito from two populations of An. nuneztovari s.l were used: one collected in the municipality of Altamira and the other in the municipality of Peixe Boi. In addition, two male populations identified as An. nuneztovari s.s. and An. goeldii were obtained from larvae collected in the municipality of Altamira²³. The species An. dunhami and An. jamariensis were not included in the study because they were not found in the two study areas. Wings of female mosquitoes from two additional species were also analyzed, Aedes aegypti, created in the entomology laboratory of the Arbovirology Section of the Evandro Chagas Institute and An. darlingi collected in the municipality of Peixe Boi (Table 1).

The preparation and assembly of the slides followed the protocol used in the Leishmaniasis Laboratory of the Evandro Chagas Institute (adapted), in which the pair of wings of each specimen was removed using a stereoscopic microscope with the aid of fine-tip tweezers and a stylet. The wing in the best physical condition, either right or left, from each specimen was used. The wings were then subjected to a dehydration process by passing them twice in absolute ethanol for 30 minutes. Afterwards, the wings were mounted between the slide and coverslip with a drop of Canada balsam in a horizontal position with the proximal region facing the left side of the slide. The slides were then labeled, photographed, and stored in a slide cabinet.

Wing images were captured using an Axiostar Plus stereoscopic microscope (ZEISS), equipped with an AxioCam CR3 (ZEISS®) camera, at 50x magnification with the help of the AxioVision version 4.7 program.

To acquire morphometric data, all images were digitized and transferred to the tpsUtil and tpsDig software programs from the TPS (Thin Plate Spline) series developed by Rohlf³⁸. The tpsUtil program was used to organize, arrange, and manipulate the image lists and the tpsDig program was used to digitize the anatomical landmarks on each wing.

Thirteen type I anatomical landmarks were digitized on each wing for geometric morphometric analysis (Figure 1). These landmarks were selected based on the wing morphology described in studies on anophelines^33,46,47. The order in which the landmarks were introduced was the same for all wings analyzed to establish the spatial homology of the structures.

Figure 1 - Wing Anopheles nuneztovari s.l. with delimitation of 13 landmarks (white circles) and resultant shape (black lines) used for geometric morphometric analysis 

After digitization the anatomical landmarks on the wings, the superposition method, namely Generalized Procrustes Analysis (GPA), was used to generate size and shape variables. For this analysis, Canonical Variables analysis (CVA) was then performed to assess shape variables between a priori defined groups using a multivariate data set. The morphospaces of each group were determined using ellipses generated with a 90% confidence interval. The Mahalanobis distances derived from shape variables were used to assess similarity among populations, as the greater the distance, the greater the difference in shape between groups or species⁴⁸. A Mahalanobis distance matrix was constituted and used for the phylogenetic reconstruction of the groups, through the UPGMA method⁴⁹ (Unweighted Pair Group Method with Arithmetic Mean; available at http://genomes.urv.es /UPGMA/DendroUPGMA). Analyzes were performed using the MorphoJ version 1.04b program⁵⁰. To verify differences in the size of the wing centroid and the construction of the violin-plot graph, analysis of variance (ANOVA) was performed, and the Tukey test was applied in the PAST version 2.17 program using data generated in the MorphoJ program.

RESULTS

Canonical Variable analysis showed that the first (VC1) and second (VC2) variables were responsible for 99.1% of the total variation in wing shape, highlighting the morphometric distinction among the study populations in the two-dimensional scatterplot.

The scatterplot (Figure 2) of populations in morphospace demonstrated the separation of the Aedes aegypti population (outgroup) from the target populations. No evident separation was observed among An. darlingi, An. nuneztovari s.l. populations from Peixe Boi and Altamira, however male populations of An. nuneztovari s.s and An. goeldii were distinct from the populations of Aedes aegypti and other populations of anophelines but showed little intersection between them. Specimens located outside the confidence ellipses were also observed.

Figure 2 - Analysis of Canonical Variables using the shape of the wings of the studied populations. The colors represent populations and Ellipses, 90% confidence interval 

In the dendrogram derived from Mahalanobis distances obtained from discriminant analyses among populations (Figure 3), a greater distance between males and females was observed, where the highest dissimilarity value was 35.32.

Figure 3 - Dendrogram of morphological similarity derived from Mahalanobis distances between study populations 

Among females, A. aegypti population was more distant from the other populations in the study, with a dissimilarity value ranging from 8.54 to 30.75, while the smallest morphological distance (1.06) was observed between the populations of An. nuneztovari s.l. from Altamira and Peixe Boi. The An. darlingi population is closer to other anopheline populations, when compared to the A. aegypti population, showing dissimilarity value close to 2. Male population An. nuneztovari s.s and An. goeldii showed great phenotypic distance, with a dissimilarity value of 11.31.

The result of centroid size analysis according to ANOVA and Tukey's test (Figure 4) showed that the populations of males An. goeldii and An. nuneztovari s.s were significantly different (p<0.05) between themselves and with all populations of females (A. aegypti, An. darlingi, An. nuneztovari s.l from Altamira, An. nuneztovari s.l from Peixe Boi).

Figure 4 - Violin-plot/boxplot graph showing the comparison between the size of the centroids of the study populations (* = P < 0.05) 

Among the female populations, the size of the centroid of A. aegypti differed from that of An. darlingi and An. nuneztovari s.l from Peixe Boi, but not from that of An. nuneztovari s.l from Altamira.

When comparing female anopheline populations, no significant difference was observed in centroid wing size between An. nuneztovari s.l. populations from Altamira and Peixe Boi. However, both populations differed significantly from An. darlingi.

DISCUSSION

The analysis of interpopulation wing morphometry revealed clear differentiation among the studied populations at both genus and species levels, particularly in the canonical variable analysis. These variables distinguished the populations into distinct groups within the morphospace, whereas centroid size analysis indicated significant differences in size among populations. The latter result is similar to that of Wilke et al³² and Lorenz & Suesdek³⁶, who managed to distinguish three genera of mosquitoes, Aedes, Anopheles, and Culex, down to the taxonomic classification levels of subgenus and species. It was also observed on the analysis of canonical variables some specimens outside the confidence ellipses, which in the future could be elucidated using a molecular approach, as these specimens may represent other species.

Interpopulation morphometric analyses among female anophelines belonging to the subgenus Nyssorhynchus showed that the population of An. darlingi differed from the populations of An. nuneztovari s.l from Altamira and Peixe Boi, with the distance from Mahalanobis to An. darlingi very close to both the population of An. nuneztovari s.l from Altamira (2.51) and from Peixe Boi (2.22). Furthermore, a significant difference was observed (p<0.05) in centroid size, highlighting the differentiation of these anopheline species.

Other authors found that wing geometric morphometrics is efficient in differentiating species of mosquitoes from the subgenus Nyssorhynchus^47-49, populations of mosquitoes from the subgenus Kerteszia^50,51 and culicid species belonging the subgenera Stegomyia, Ochlerotatus, Kerteszia, Nyssorhynchus and Culex³².

The results of inter and intrapopulation analyses of An. nuneztovari s.l females from Altamira and Peixe Boi showed that showed no significant difference in size (p=1.16) and shape between these populations, as no segregation was observed in morphospace in the analysis of canonical variables and small Mahalanobis distance (1.06). However, between male populations of An. nuneztovari s.s and An. goeldii, a significant differences centroid size were detected (p<0.05) and difference in shape, evidenced in the segregation in the analysis of canonical variables and a Mahalanobis distance with a dissimilarity value of 11.31.

Female specimens of mosquitoes belonging to the nuneztovari complex are morphologically indistinguishable according to existing identification keys, however, males can be identified based on genital morphology^17,19,23. Similar distinguishing characteristics have been reported in some species of mosquitoes of the Culex subgenus that are considered difficult to identify, as they do not present evident morphological differences and with distinctive characteristics restricted to the male genitalia⁵².

Monteiro & Reis³⁹ proposed that the sample size should be at least four times greater than the number of anatomical landmarks, although this is not always feasible for all target populations because of the limited number of specimens collected in the field. However, the present study demonstrated that males of species from the An. nuneztovari complex can also be differentiated by wing geometric morphometrics despite sampling limitations. Among females of this complex, no difference in wing shape was observed, however, we cannot say that these females cannot be differentiated by wing geometric morphometry, as no other identification method such as by molecular biology was carried out to ensure that there are no different species within these population. Another hypothesis is that the males of this complex present greater variation in wing shape than females, enabling their differentiation. This hypothesis is relevant because similar patterns have already been observed in the filariasis vector mosquito species of the Mansonia subgenus⁵³.

Although this is the first study to use wing geometric morphometry to differentiate species within the An. nuneztovari complex, this technique has already been successfully applied to other cryptic mosquito species complexes^52,54 and other medically important insects such as sandflies^55,56 and triatomines^57,58.

Using the morphometric analysis, it was also possible to verify the sexual dimorphism in wing shape and size, that is, males and females of An. nuneztovari formed distinct groups without intersection in the morphospace of canonical variables, and the dendrogram derived from Mahalanobis distances showed a large distance between males and females with different dissimilarity values.

Centroid size analysis revealed a significant difference (p<0.05) between sexes, wings of females are larger than those of males. This difference in wing size was observed by other authors in mosquitoes of the genus Culex⁵⁹ and the genus Aedes^29,60-62.

Although distinction between male and female anophelines is commonly based on morphological characteristics of the antennae, palps, and genitalia⁴⁵ and this distinction is important for public health, our findings show that the analysis of the wings, and conditions of An. nuneztovari, also reveals sexually dimorphic characteristics. Wing shape is a sex-specific phenotypic trait among these anophelines. This result is in line with Virginio et al⁶³ who explained sexual dimorphism in several mosquito species of the genera Aedes, Anopheles, and Culex, as well as those by Manimegalai et al⁶⁴ in Cx. quinquefasciatus from India, Devicari et al²⁹ in Aedes scapularis, Vidal et al³⁰ in Aedes albopictus, Lorenz et al⁶⁵Anopheles cruzii, and Christe et al⁶⁶ in Aedes fluviatilis.

Therefore, wing morphometry can provide useful information for assessing sexual dimorphism in medically important culicids, which is relevant because males and females perform distinct ecological roles and use their wings differently⁴⁵.

CONCLUSIONS

Wing geometric morphometry proved to be an effective approach for taxonomic differentiation and the evaluation of sexual dimorphism, and showed considerable potential for distinguishing male mosquitoes within the nuneztovari complex. The results indicate that wing shape may represent a valuable morphological marker for taxa that are difficult to identify based on conventional external morphology, especially cryptic species.