Unraveling the Morphological Variation of Triatoma infestans in the Peridomestic Habitats of Chuquisaca Bolivia: A Geometric Morphometric Approach

Simple Summary Triatoma infestans is the main vector of the Chagas disease transmission and has been for years one of the main sanitary problems in Bolivia, particularly for the movement between isolated population to the urban areas. In the following research, we analyze the pattern of biological adaptation of this vector species from two different areas in Bolivia (areas of the Bolivian Chaco with the inter-Andean valleys). Using advanced geometric morphometric tools, it was possible to unravel T. infestans morphological variation and understand the biological adaptation of this important insect species. Abstract Morphometrics has been used on Triatomines, a well-known phenotypically variable insect, to understand the process of morphological plasticity and infer the changes of this phenomenon. The following research was carried out in two regions of the inter-Andean valleys and two Chaco regions of Chuquisaca-Bolivia. Triatoma infestans adults were collected from the peridomestic (pens and chicken coops) along a geographic gradient in order to evaluate the morphological differentiation between groups and their pattern of sexual shape dimorphism. Geometric morphometric methods were applied on the wings and heads of T. infestans. The main findings include that we proved sexual dimorphism in heads and wings, determined the impact of environmental factors on size and shape and validated the impact of nutrition on head shape variation. These results show that geometric morphometric procedures can be used to provide key insight into the biological adaptation of T. infestans on different biotic (nutrition) and abiotic (environment) conditions, which could serve in understanding and evaluating infestation processes and further vector control programs.


Introduction
In Bolivia, Triatoma infestans (Klug) Hemiptera Reduviidae is the main vector of Trypanosoma cruzi, a parasite that causes Chagas disease. T. infestans is a synanthropic insect it is found in seven countries in Latin America [1]. In Bolivia, T. infestans is not only limited to the intradomiciliary and peridomestic habitat, but in some localities of the inter-Andean valleys and the Chaco, there are sylvatic foci [1][2][3]; however, in Chuquisca no sylvatic foci outbreaks have been identified yet [4]. Environmental characteristics, such as temperature and relative humidity, as well as the structure of the pens and chicken coops, are important factors to understand the morphological characteristics of the insect. The wet Chaco is also found in the Tucumano-Boliviano region; it is a region of lower mountains, the height of the mountains does not exceed 2600 m above sea level, at the base of the mountains, the altitude is 900 m above sea level. It has a warm and humid climate; the temperature is around 28 °C per year and humidity is close to 50% [25].
The dry Chaco region is located at the east of the Eastern Mountain Range, a region of flat arid lands, related to Paraguay and Northern Argentina, which has a warm climate with annual temperatures above 30 °C, and low humidity, around 20%, it is denominated the Bolivian Boreal Chaco [25].

Insect Sampling and Preparation
A total of 110 adults of Triatoma infestans were examined, 57 females and 53 males, distributed as follows: Tarabuco/Sarufaya (TS) 11; 11, Sucre/Surima (SS) 22; 18, Huacaya/Imbochi (HI) 13; 12, Monteagudo/Cañon Largo (MC) 11; 12 females and males, respectively. All were collected in peridomestic locations (pens and chicken coops). We used a gripper to collect the insects, and each insect was placed in a plastic container. The collection was between July and September 2018. The collection of insects was carried out for three weeks per location. The wet Chaco is also found in the Tucumano-Boliviano region; it is a region of lower mountains, the height of the mountains does not exceed 2600 m above sea level, at the base of the mountains, the altitude is 900 m above sea level. It has a warm and humid climate; the temperature is around 28 • C per year and humidity is close to 50% [25].
The dry Chaco region is located at the east of the Eastern Mountain Range, a region of flat arid lands, related to Paraguay and Northern Argentina, which has a warm climate with annual temperatures above 30 • C, and low humidity, around 20%, it is denominated the Bolivian Boreal Chaco [25].

Insect Sampling and Preparation
A total of 110 adults of Triatoma infestans were examined, 57 females and 53 males, distributed as follows: Tarabuco/Sarufaya (TS) 11; 11, Sucre/Surima (SS) 22; 18, Huacaya/Imbochi (HI) 13; 12, Monteagudo/Cañon Largo (MC) 11; 12 females and males, respectively. All were collected in peridomestic locations (pens and chicken coops). We used a gripper to collect the insects, and each insect was placed in a plastic container. The collection was between July and September 2018. The collection of insects was carried out for three weeks per location.
Adult insects were introduced in a plastic container. At least 11 insects of the same sex were collected per location and were preserved in alcohol (96%) for further analyses. In the laboratory, wings were mounted in slides with Euparal ® for the analysis using the right and left wings in all cases. Each head was excised at the collar and mounted on a pin attached to a metal support. All wings and heads were photographed and measured with a Celestron handheld digital microscope pro 5MP.

Morphometric Analysis
Eight landmarks were selected for dorsal views of the head, and nine landmarks of the wings ( Figure 2) and digitized using the software TpsDig2 V.231 [26]. For all digitized individuals, the shape information was extracted using a Procrustes superimposition analysis, which is a procedure that removes the information of size, position, and orientation to standardize each specimen according to centroid size [27]. Adult insects were introduced in a plastic container. At least 11 insects of the same sex were collected per location and were preserved in alcohol (96%) for further analyses. In the laboratory, wings were mounted in slides with Euparal ® for the analysis using the right and left wings in all cases. Each head was excised at the collar and mounted on a pin attached to a metal support. All wings and heads were photographed and measured with a Celestron handheld digital microscope pro 5MP.

Morphometric Analysis
Eight landmarks were selected for dorsal views of the head, and nine landmarks of the wings ( Figure 2) and digitized using the software TpsDig2 V.231 [26]. For all digitized individuals, the shape information was extracted using a Procrustes superimposition analysis, which is a procedure that removes the information of size, position, and orientation to standardize each specimen according to centroid size [27]. The measurement error (ME) was calculated using a Procrustes ANOVA in order to detect digitizing errors in morphometric data. For this procedure, the original dataset was compared with a control of repeated measures, and the values of the mean squares (MS) of the individual values were compared with the error (dataset of the repeated measurement) [28,29].
To characterize the head and wing shape variation, a principal component analysis (PCA) was carried out based on the covariance matrix of shape. Canonical variate analysis (CVA) methods were used to amplify the shape variation and visualize the sexual shape dimorphism between T. infestans populations [30][31][32]. Mahalanobis and Procrustes morphological distances were calculated and reported with their respective p-values after a permutation test (10,000 runs). Multivariate regression of shape (dependent variable) on centroid size (independent variable) was performed to analyze if the size has an influence on the shape distribution (allometric effect) of T. infestans populations of Inter-Andean valleys and Chaco. All the analyses were performed using the software MorphoJ V.1.06 [29] and the R package Momocs [33]. The measurement error (ME) was calculated using a Procrustes ANOVA in order to detect digitizing errors in morphometric data. For this procedure, the original dataset was compared with a control of repeated measures, and the values of the mean squares (MS) of the individual values were compared with the error (dataset of the repeated measurement) [28,29].
To characterize the head and wing shape variation, a principal component analysis (PCA) was carried out based on the covariance matrix of shape. Canonical variate analysis (CVA) methods were used to amplify the shape variation and visualize the sexual shape dimorphism between T. infestans populations [30][31][32]. Mahalanobis and Procrustes morphological distances were calculated and reported with their respective p-values after a permutation test (10,000 runs). Multivariate regression of shape (dependent variable) on centroid size (independent variable) was performed to analyze if the size has an influence on the shape distribution (allometric effect) of T. infestans populations of Inter-Andean valleys and Chaco. All the analyses were performed using the software MorphoJ V.1.06 [29] and the R package Momocs [33].

Results
The Procrustes ANOVA for assessing the measurement error of head shape showed that the mean square for individual variation exceeded the measurement error: MS error: 0.0000354342 < MS individual: 0.0001282749. The measurement error in the wings showed Principal component analysis showed that the first three PCs accounted for 52.882% of the head shape variation (PC1:22.440%, PC2:16.373%, PC3:14.069%). The PCA of the wings view showed that the first three PCs accounted for 61.055% (PC1:33.499%, PC2:16.613%, PC3:10.743%). In order to localize the shape variation, the average shape was extracted for the two localities of Inter-Andean valleys (Sucre/Surima and Tarabuco/Sarufaya) and two localities of Chaco (Huacaya/Imbochi and Monteagudo/Cañon Largo). The dorsal head view showed that individuals of T. infestans from HI were clearly different from the TS, but with few superpositions of individuals with MC and SS and the wing view, the superposition of individuals was more evident for all populations. (Figures 3-5).

Results
The Procrustes ANOVA for assessing the measurement error of head shape showed that the mean square for individual variation exceeded the measurement error: MS error: 0.0000354342 < MS individual: 0.0001282749. The measurement error in the wings showed that the mean square for individual variation exceeded the measurement error (MS error: 0.000099151< MS individual: 0.0010029078).
Principal component analysis showed that the first three PCs accounted for 52.882% of the head shape variation (PC1:22.440%, PC2:16.373%, PC3:14.069%). The PCA of the wings view showed that the first three PCs accounted for 61.055% (PC1:33.499%, PC2:16.613%, PC3:10.743%). In order to localize the shape variation, the average shape was extracted for the two localities of Inter-Andean valleys (Sucre/Surima and Tarabuco/Sarufaya) and two localities of Chaco (Huacaya/Imbochi and Monteagudo/Cañon Largo). The dorsal head view showed that individuals of T. infestans from HI were clearly different from the TS, but with few superpositions of individuals with MC and SS and the wing view, the superposition of individuals was more evident for all populations. (Figures 3-5).     The scatterplot of CVA shows differentiation between females and males (sexual shape dimorphism) in heads and wings of T. infestans populations of Chaco and Inter-Andean valley ( Figure 6).
After extracting Mahalanobis and Procrustes distances (permutations 10,000 runs), T. infestans did not show sexual shape dimorphism using Procrustes distances for both structures and also for Mahalanobis was not evident in the head of the inter-Andean valleys. According to the relationship between the Mahalanobis Distance (p < 0.0001), T. infestans populations of Chaco showed sexual shape dimorphism in the head, but not for wings, and, in addition, T. infestans populations of inter-Andean valleys presented dimorphism in wings (Table 1).  The scatterplot of CVA shows differentiation between females and males (sexual shape dimorphism) in heads and wings of T. infestans populations of Chaco and Inter-Andean valley ( Figure 6).  After extracting Mahalanobis and Procrustes distances (permutations 10,000 runs), T. infestans did not show sexual shape dimorphism using Procrustes distances for both structures and also for Mahalanobis was not evident in the head of the inter-Andean valleys. According to the relationship between the Mahalanobis Distance (p < 0.0001), T. infestans populations of Chaco showed sexual shape dimorphism in the head, but not for wings, and, in addition, T. infestans populations of inter-Andean valleys presented dimorphism in wings (Table 1).   The multivariate regression showed that, although the allometric percentage was lower, the influence of size was noticeable in the different traits evaluated, where shape variation showed influence by allometry in head and wings; dorsal head view 3.73242% p-value 0.0001 and wings view 3.1994% p-value < 0.0001. It is possible to identify that T. infestans from the Inter-Andean valley are bigger than the specimens from Chaco (see set of gray points at the left of Figure 7A and set of blue points at the right of Figure 7B.  Figure 7). When the analysis is separated between the four analyzed populations, a clear sexual size dimorphism was also observed were males from inter-Andean valleys, and Chaco was smaller in size for both of the traits compared to the females (Figure 8). variation showed influence by allometry in head and wings; dorsal he p-value 0.0001 and wings view 3.1994% p-value < 0.0001. It is possib T. infestans from the Inter-Andean valley are bigger than the specimens set of gray points at the left of Figure 7A and set of blue points at the r (Figure 7). When the analysis is separated between the four analyzed po sexual size dimorphism was also observed were males from inter-And Chaco was smaller in size for both of the traits compared to the females

Discussion
This study analyzed the morphological plasticity and sexual shape dimorphism of T. infestans in two geographical environments, the inter-Andean valleys and Chaco and found the following results: (A) sexual dimorphism in heads and wings; (B) significant effect of environmental factors on size and shape; (C) impact of nutrition on head shape variation.
(A) In Bolivia, sylvatic-Andean T. infestans inhabits rock piles and feeds on animals living in burrows. Morphologically it has a yellow connexivum, similar to the intradomiciliary variety. At the same time, the specimens from boreal Chaco that live in trees and

Discussion
This study analyzed the morphological plasticity and sexual shape dimorphism of T. infestans in two geographical environments, the inter-Andean valleys and Chaco and found the following results: (A) sexual dimorphism in heads and wings; (B) significant effect of environmental factors on size and shape; (C) impact of nutrition on head shape variation.
(A) In Bolivia, sylvatic-Andean T. infestans inhabits rock piles and feeds on animals living in burrows. Morphologically it has a yellow connexivum, similar to the intradomiciliary variety. At the same time, the specimens from boreal Chaco that live in trees and feed on birds have a dark connexivum [34][35][36]. Males often made their movements by flying and prefer peridomestic and intradomicile environments; conversely, females have limited dispersal capabilities, they remain in a single habitat, and they do not discriminate their food source [21]. This research confirmed a significant morphological variation in the head and wings of T. infestans. In sylvatic environments, investigated populations of T. infestans showed high levels of sex-based dimorphism. Results showed that females are larger than males (discussed in Djuradin et al. [9]). In species like insects, a sexual dimorphism observed in smaller size and shape of males is often revealed in many species of Diptera, Lepidoptera, Hymenoptera and Coleoptera [37][38][39][40][41][42][43]. After further analyses, this survey found that sexual dimorphism depends on geographic region. T. infestans populations from Chaco had sexual shape dimorphism in the head but were not observed for wings, while T. infestans populations of inter-Andean valleys had sexual dimorphism in wings. According to Fairbairn [44] and Cox et al. [45], sexual dimorphism may be the result of ecological and reproductive pressure. Sexual dimorphism in triatomines can be related to feeding habits and population density [24,46,47]. In the Bolivian Chaco (Huacaya/Imbochi and Monteagudo/Cañon Largo), T. infestans were found in the intra-and peridomestic habitat, with increased population density compared to populations from the valleys. High population density may determine an intraspecific competition for food, consequently having males with smaller heads than females [48]. As discussed in Mikac et al. [49] for coleopteran species and in Lemic et al. [42] for dipteran species, it is thought that bigger wings are probably more aerodynamic and may also be useful for mated females that are known to engage in migratory flights. Considering presented results and based on literature review [43,[49][50][51][52], this study provides opposite morphological evidence than shown in Hernández et al. [21] that the migration in T. infestans can be attributed to the females of this species.
(B) The insect size and shape were influenced by environmental factors. The geography of inter-Andean valleys featuring mountains, temperatures between 17 and 24 • C and humidity above 40%, and the structure of peridomestic environments, such as chicken coops and pens, built with earth blocks that make them dark and wet, are unfavorable factors for the development of T. infestans microcolonies. In these unfavorable environmental conditions, females have the priority in feeding, which has a direct influence on their bigger wing size and shape [53][54][55]. T. infestans fly in the warmer months when temperatures get close to 30 • C. At temperatures below 20 • C, T. infestans do not fly; therefore, re-infestation of peridomicile to intradomicile becomes difficult because the insects must walk for feeding [56]. Centroid size was used in this research as a measure of overall head and wing size differences among T. infestans populations. Because of these high temperatures in boreal Chaco (Huacaya/Imbochi) (temperature higher than 30 • C and humidity lower than 20%), the specimens were smaller compared to those from the valley. Vilaseca et al. [12] similarly observed that the centroid size of T. infestans was larger in populations from the inter-Andean valleys compared to specimens from Chaco. In Chaco, variations were observed in T. infestans heads; males had a smaller head and a different shape compared to female heads. According to Hernández et al. [48], changes in the males' head morphology have a dispersal genetic component. The variation in size among populations suggests strong differential selection and sensitivity to changes in environmental conditions [57][58][59][60][61]. Reproductive studies on insects have shown that the biological cycles associated with high temperatures would be shorter; therefore, the specimens tend to be smaller [8,9,53,[62][63][64][65]. In contrast to size, analyses of organismal shape, which was proofed to be influenced by allometry, provide more reliable information about the phenotypic variation of populations representing high and stable heritability [66,67]. Although T. infestans head and wing shape comparisons revealed some differences between populations, clear site-specific population differentiation was not found. However, an environmental pattern in head and wing shape variation was detected when populations were pooled by sex (as already described in part A).
(C) Nutrition profile has been observed in variability in this study. Unfavorable environmental conditions lead to circumstances in which females have the priority to feed, therefore resulting in smaller males (in shape and size). Consequently, the females become bigger than the males as an adaptation mechanism to the environmental pressures in order to enhance fertility [53][54][55]. Males carry out the re-infestation process, so they tend to stay in the peridomicile (Chicken coops and pens) [21]. T. infestans mostly feed on bird blood. According to Natero et al. [68] and Lunardi et al. [69], when triatomines feed on bird blood, no variability in the head shape will occur. In this research, no dimorphism was observed in the shape of the head because both females and males fed on bird blood. When analyzing centroid size, females' heads occur longer than male heads. Except for the intensity of feeding, the type of blood host also could have an influence on morphological variation in this species. According to Natero et al. [68], when T. infestans feeds on mammalian blood, a morphological widening and shortening effect of the head is observed, in contrast to individuals feeding on bird blood. Lunardi et al. [69] found phenotypic plasticity in Triatoma williami based on feeding on a mammal or bird blood as an adaptation process to the host. In this research, T. infestans from Chaco was found in the intra and peridomicile, and the males tend to fly from the peridomestic to intradomicile and feed on human/mammal blood, which evidently had a direct influence on the variability in size and shape of the male heads comparing with female heads.
At the population level, the variation of the first three principal components showed that there was no clear-cut separation between the populations, which is in agreement with a high proportion of misassignments and the findings of low population structure and no isolation by distance (using mitochondrial cytochrome b gene and microsatellite loci), as demonstrated by Giordano et al. [4] and Marcet et al. [70]. According to these authors' findings, T. infestans populations are genetically similar. The slight wing shape differences detected (especially between wings) may be the result of emerging phenotypic plasticity. Phenotypic plasticity is often defined as the change in the phenotypic expression of a genotype in response to environmental factors [71] and has been shown to have significant evolutionary consequences [71,72].

Conclusions
The morphological variation of T. infestans from two different environments has been determined. The inter-Andean and Chaco populations showed the sexual size and shape dimorphism in relation to environmental factors and nutrition. The results of the present survey will serve as a starting point in further understanding the re-infestation processes to redesign the science-based vector control programs.