Crossing experiments detect partial reproductive isolation among populations of Triatoma longipennis (Hemiptera: Reduviidae: Triatominae)

Abstract

Chagas disease is an important vector-borne disease endemic in Mexico. Of the 33 triatomine species found in Mexico, Triatoma longipennis (Usinger) is considered among the most important because of its infection indices, capacity for transmitting Trypanosoma cruzi (Chagas), and its distribution areas. Here, we describe the results of a reproductive isolation analysis among 5 populations of T. longipennis collected from representative areas of Mexico. Fertility and segregation of morphological characteristics were examined in two generations of hybrids. The percentage of pairs with (fertile) offspring varied from 30% to 100% in the parental crosses, while these values varied from 0 to 100% in the intersite crosses. Our results indicate partial reproductive isolation among these populations. These findings shed light on the potential presence of a cryptic species complex of T. longipennis in Mexico.

Graphical Abstract

Open in new tabDownload slide

Chagas disease, also known as American trypanosomiasis, is a potentially life-threatening illness caused by the protozoan parasite Trypanosoma cruzi (Chagas) (Trypanosomatida: Trypanosomatidae). Between 6 and 7 million people are estimated to be infected worldwide with T. cruzi. The disease is found mainly in endemic areas of 21 continental Latin American countries, where it is transmitted to humans and other mammals primarily by contact with feces or urine of some hemipterans of the family Reduviidae, included in the subfamily Triatominae (WHO 2022). Up to 30% of people chronically infected with T. cruzi develop cardiac alterations (e.g., cardiomyopathies), and up to 10% develop digestive, neurological, or mixed alterations, which may require specific treatment (WHO 2022). In Mexico, the national estimated seroprevalence of T. cruzi infection was 3.38%, suggesting that there are 4.06 million cases (Arnal et al. 2019). Studies focused on pregnant women, who may transmit the parasite to their newborn during pregnancy, reported an estimated seroprevalence of 2.21%, suggesting that there are 50,675 births from T. cruzi-infected women per year and 3,193 cases of congenitally infected newborns per year (Arnal et al. 2019).

In Mexico, more than 30 species of triatomines have been recorded, 11 of which are considered epidemiologically important vectors of T. cruzi based on their distribution, domestic habits, and infection indices (Salazar-Schettino et al. 2019). One of the most important species is Triatoma longipennis (Usinger), which has been reported in 10 states of western, north, northwestern, and northcentral Mexico (Martínez-Ibarra et al. 2021). This species is considered the main vector in the states of Nayarit, Jalisco, and Zacatecas, where the prevalence of T. cruzi human infections has reached 12.2%, infestation and colonization indices have reached 67.3% and 50%, respectively, in some areas, and the percentages of infected triatomines have been reported from 32.4% to 44.8% (Martínez-Ibarra et al. 2021).

Triatoma longipennis has been considered a single species throughout its range (Salazar-Schettino et al. 2010); however, as more studies have been conducted, noticeable differences have been observed when comparing diverse populations of T. longipennis (Martínez-Ibarra et al. 2013a,b, Martínez-Hernández et al. 2021). A molecular study using Cyt b to characterize the population genetics of T. longipennis showed that the population from the state of Guanajuato had a high haplotype differentiation, separate from the other 13 studied populations. In the same study, 3 clades were identified; one clade included the specimens from the states of Colima, Michoacan, Zacatecas, Nayarit, and all populations from Jalisco (except one specimen from Autlan de Navarro); a second clade included the two populations from Chihuahua and one specimen from Jalisco (Autlan de Navarro) and one from La Palma, Michoacan (Fig. 1). Finally, the third (and more distinguishable) clade included the population from Guanajuato (Martínez-Hernández et al. 2021).

Life cycle varied from 193.9 to 197.8 days in T. longipennis from the states of Guanajuato, Michoacan (La Palma), and Jalisco (Sayula) in contrast to 211.8–224.7 days in specimens from the states of Chihuahua (Saucito) and Durango (Huazamota) (Fig. 1). Also, mortality rates varied from 18.9% to 20.8% in T. longipennis from Chihuahua, Guanajuato, and Jalisco to 50.9% in specimens from Michoacan and 80.7% in T. longipennis from Durango. Fecundity varied from 1.2 to 1.4 eggs/♀/day in most populations to 1.9 eggs/♀/day in T. longipennis from Jalisco. Likewise, the egg eclosion rates were different when comparing populations from Durango (46.5%), Chihuahua (54.9%), Michoacan (59%), Guanajuato (64.6%), and Jalisco (79.2%) (Fig. 1) (Martínez-Ibarra et al. 2013a,b).

We detected differences in the feeding and defecation behaviors of populations from areas with high (states of Nayarit, Jalisco, and Zacatecas) and low (states of Chihuahua, Aguascalientes, and Guanajuato) prevalence of human infection by T. cruzi (Fig. 1). Times for starting a blood meal in third-, fourth-, and fifth-instar nymphs and adults of T. longipennis varied according to populations: Chihuahua and Aguascalientes (3.9 to 8.0 min); Guanajuato (3–7 min); and Zacatecas, Nayarit, and Jalisco (1–3) (Martínez-Ibarra et al. 2013a, b). It has been reported that the percentages of specimens defecating when eating, immediately after feeding, and in less than 2 min postfeeding [when young (first- to third-instar nymphs) and old (fourth- and fifth-instar nymphs)] varied when comparing populations of T. longipennis. The percentages of defecation in young nymphs from Chihuahua, Aguascalientes, and Guanajuato were noticeably lower (4.0–6.7%) than of T. longipennis from Zacatecas, Jalisco, and Nayarit (20.1–26.5%). A similar phenomenon was observed when comparing percentages of old nymphs of T. longipennis from Chihuahua and Aguascalientes (17.1–19.2%), Guanajuato (25.8%), Nayarit (56.1%), and Zacatecas and Jalisco (67.2–71.7%) (Martínez-Ibarra et al. 2021).

Courtship of triatomines and copulation behavior is brief. Chemical attractants are common for triatomines, and these chemical factors may reduce the need to rely on a patterned approach (Chiang and Chiang 2021). To initiate copulation after encountering a receptive female, the male mounts the dorsal-lateral side of the female facing the same direction. The male holds onto the female by placing his legs on her body while the female stays on the substrate. With his genital capsule next to the female genitalia, the male lifts his parameres from their grooves at the posterior dorsal rim of his genital capsule and extends them toward the female’s external genital plates. These parameres then hook onto the dorsal edge of the female dorsal genital plate and may serve to clasp this segment and direct the aedeagus (the intromittent organ of the male insect, also named phallus) into the female (Chiang and Chiang 2021). During the first minute of copulation, the phallus is inserted into the vagina. After 5–6 min, the endosome everts, and at about 10 min of copulation, it is completely inflated, exposing its sclerotized substructures. At this point, the spermatophore (a male reproductive structure that packages sperm cells to aid in their transmission to females during mating) is completely formed into the endosome, and 2 min later, the endosome collapses, leaving the spermatophore in the vagina. Copulation is successful when viable sperm reach the spermathecae (Téllez-García et al. 2019). Polyandry (a female with different sexual couples) has been recorded in different species of triatomines (Chiang and Chiang 2021).

Experimental crosses have an important role in systematic issues in Triatominae (Alevi et al. 2021). Even when observed experimentally, reproductive isolation is one of the best criteria to evaluate the taxonomic status of morphologically or genetically close populations (Alevi et al. 2018, Villacís et al. 2020, Delgado et al. 2021, Madeira et al. 2021, Vicente et al. 2022). Degrees of reproductive incompatibility can be estimated by determining the pre and postzygotic reproductive barriers (i.e., isolation mechanisms) acting when two taxa (e.g., species, subspecies, and even populations) are experimentally crossed (Campos-Soto et al. 2016, Neves et al. 2020, Cesaretto et al. 2021, Ravazi et al. 2021, dos Reis et al. 2022, Martínez-Ibarra et al. 2023).

Different biological parameters can be used to research the effect of some isolation mechanisms over the crosses of 2 taxa, such as the number of pairs having offspring, the fecundity of females, egg fertility, mortality rates, survivorship of offspring until adulthood, and fertility of offspring (Campos-Soto et al. 2016, dos Reis et al. 2022, Martínez-Ibarra et al. 2023).

Therefore, we carried out experimental crosses, and we evaluated the possible pre and postzygotic reproductive barriers, aiming to understand the potential presence of a cryptic species complex of T. longipennis in Mexico.

Material and Methods

In order to evaluate the genetic and reproductive compatibility and thereby glimpse the potential presence of a complex of cryptic T. longipennis species (Ravazi et al. 2021, Martínez-Hernández et al. 2021) in Mexico, reciprocal crossing experiments were carried out between 5 different populations of T. longipennis from diverse locations along the of the area of distribution of this species.

Biological Material

Specimens used in these crossing experiments were obtained from colonies established for at least 5 generations with Triatominae from allopatric populations. The 5 colonies used in the experiments comprised 5 populations of T. longipennis for which previous studies (Martínez-Ibarra et al. 2013a, b, 2017, 2021, Martínez-Hernández et al. 2021) have reported genetic and behavioral differences, suggesting an ongoing divergence process, leading to the formation of new species. The colonies were started by the collection of at least 30 specimens of T. longipennis from each locality: Jala (21° 03ʹ N, 104° 26ʹ W), in the state of Nayarit; Guadalupe y Calvo (26° 05ʹ N, 106° 57ʹ W), Chihuahua; Santa Catarina (21° 08ʹ N, 100° 04ʹ W), Guanajuato; Sayula (19° 52ʹ N, 103° 34ʹ W), Jalisco, and La Palma (20° 12ʹ N, 101 °44ʹ W), Michoacan (Fig. 2). The localities where the founders of the studied colonies were initially collected are located in different ecoregions (Level III) of Mexico: Jala and La Palma are located in ecoregion 21 (Hills and mountains with coniferous, oak and mixed forests in central Mexico), Guadalupe y Calvo is in ecoregion 17 (Hills of Sonora and Sinaloa and canyons of the Sierra Madre Occidental with xeric scrub and deciduous forest), Santa Catarina is in ecoregion 20 (Hills and plains of the interior with xerophytic scrub and low mesquite forest), and Sayula is in the ecoregion 6 (Coniferous, oak and mixed forests of the Sierra Madre of southern Jalisco and Michoacan). Specimens of T. longipennis from western and central Mexico were collected in peridomestic and domestic areas of human dwellings, while those from Guadalupe y Calvo (northern Mexico) were collected in sylvatic areas (caves).

Specimens were identified based on the dichotomous keys published by Lent and Wygodzinsky (1979) and exhibited the typical morphological characteristics of T. longipennis, with slight variations in size and color (from dark yellow to orange) of the connexivum (Fig. 3).

Crossing Experiments

The fifth generation of each population was used to generate the experimental crosses for this study. For those crosses, 10 pairs of each set were placed in plastic jars (5 cm diameter × 10 cm height) (Table 1): Sayula × Guadalupe y Calvo, Sayula × Jala, Sayula × La Palma, Jala × Guadalupe y Calvo, Jala × La Palma, Guadalupe y Calvo × La Palma, Sayula × Santa Catarina, Jala × Santa Catarina, La Palma × Santa Catarina, and Guadalupe y Calvo × Santa Catarina. Specimens were maintained within incubators at 27 ± 1°C and 70 ± 5% relative humidity, with a photoperiod of 12:12 h of light:dark, and were fed fortnightly on New Zealand rabbits (Oryctolagus cuniculus L.). In a previous study, these were favorable conditions for raising T. longipennis (Martínez-Ibarra et al. 2021). The rabbits were maintained under laboratory conditions and were handled and anesthetized following the Norma Oficial Mexicana NOM-062-ZOO-1999 by intramuscular inoculation with 0.25 ml/kg of ketamine (SAGARPA 1999). The ethical committee approved our study with the report CEI/017/2020. Crosses of the 5 parental lineages involved in the study were used as controls.

To record fecundity, all crosses were checked daily to record the copulation event; for direct observation of the copula or by observation of spermatophore elimination once the spermatophore received by the female is expelled after copulation and after the separation of the couple (Chiang and Chiang 2021). To check egg fertility, eggs of each cross were collected for 30 days, assuming an egg production of 1.2–2.4 egg/♀/day, and incubated under previously described conditions (Martínez-Ibarra et al. 2013a, b, 2021).

To obtain 200 F1 experimental individuals, 20 first-instar nymphs from each of the 10 crosses of each set were obtained; they were placed in plastic jars (10 in each jar). In those crossed sets where not all performed crosses were successful, more F1 nymphs were obtained from successful crosses until reaching 200 specimens. All studied specimens were fed on rabbit blood and reared. To prevent uncontrolled mating, the fifth instar was sexed, and females were separated from males in different plastic jars, such as those previously described, until they reached the adult stage (Martínez-Ibarra et al. 2021). The first 100 specimens of each set reaching adulthood were included in the results.

In those cases where crosses were unsuccessful (no offspring), if eggs were laid, those inviable eggs were observed under a stereoscopic microscope to verify embryo formation. Complementarily, F1 males were placed with a virgin female of their parental lineage, while the studied F1 females were placed with a parental male. When the females laid numerous fertile eggs, which produced F1 nymphs, fertility was confirmed (Martínez-Ibarra et al. 2011).

Crosses Between F1 Individuals

We followed our standardized methodology for crossing triatomines, described in a previous study (Martínez-Ibarra et al. 2011). Briefly, to determine whether the F1 offspring of crosses between the 5 populations in the study were fertile, 10 females and 10 males of each type were crossed. After F2 was obtained, viability and fertility were determined. The experimental order of crosses was 10 F1 females × 10 siblings F1 males. Once again, 20 F1 nymphs (hoping to obtain at least 10 adults, given mortality rates) from each of the crosses of each set were placed in plastic jars (10 to a jar). Nymphs were fed on rabbit blood and reared to adults. Likewise, to ascertain the fertility of F2, 5 F2 females and 5 sibling F2 males were crossed. As soon as first-instar nymphs were obtained, F2 fertility was proven.

However, male or female specimens involved in F1 × F1 crosses that produced no F2 progeny were backcrossed with specimens of a parental lineage to ascertain if they were fertile (Southwood and Henderson 2000). If no offspring were obtained, they were backcrossed with specimens of the other parental lineage. In both cases, fertility was confirmed when the females laid numerous fertile eggs, which produced first-instar nymphs. A similar procedure was followed when male or female specimens involved in F2 × F2 crosses produced no F3 progeny (Martínez-Ibarra et al. 2011).

Statistical Analyses

The Holm–Sidak method was used to compare the mean number of eggs laid per female. A chi-square test was used to compare frequencies. Sigma Stat 3.1 software (Version 3.1 for Windows, Systat Software, Inc., San Jose, CA, USA) was used for statistical analysis. The results were considered to be statistically significant when P < 0.05.

Results

Most morphological characteristics (except for the overall size and color of the connexivum) were similar among the 5 studied populations (Fig. 2). Those specimens from Guadalupe y Calvo had the longest overall size (female ≤ 3.5 cm, male ≤ 3.2 cm). The overall sizes (length) of the females and males of the other 5 studied populations were ≤ 3.2 cm and ≤ 3.0 cm, respectively. Most populations from western Mexico (Jala, Sayula, La Palma) had an orange connexivum in contrast to dark yellow in specimens from Guadalupe y Calvo and Santa Catarina (Fig. 2). The offspring of all crosses were similar in overall size to most parental populations (≤ 3.2 cm in females and ≤ 3.0 cm in males). Similarly, all the offspring of crosses involving a population from western Mexico (Jala, Sayula, and La Palma) had a light orange connexivum, in contrast to the offspring of the crosses Guadalupe y Calvo × Santa Catarina (in both directions), whose descendants exhibited a pale yellow connexivum (Fig. 3).

A 100% success rate (100% of couples with offspring) was observed in only 2 parental crosses (Sayula × Jala in both directions and Guadalupe y Calvo × La Palma in both directions) (χ² = 73.80–82.47, df = 1, P < 0.01). Control crosses from the 5 populations studied also had a 100% success rate. Success rates between 60% and 80% were recorded for 2 other sets of crosses where the population from La Palma was involved (× Sayula and × Jala, in both directions), as well as in that of Jala × Guadalupe y Calvo (in both directions) (χ² = 70.7–79.8 df = 1, P < 0.01) (Table 1). Evidence of copulation events was observed in all those sets of crosses.

The lowest success percentages (mostly 30%) were observed in the 4 sets of crosses where the population of Santa Catarina was involved and the set of crosses of Sayula × Guadalupe y Calvo (in both directions) (χ² = 66.48–74.11, df = 1, P < 0.01; Table 1). In all those crosses, evidence of copulation events was observed.

In most sets of crosses (including most controls), the mean egg production rate of parental crosses varied from 1.9 to 2.4 eggs/♀/day (Table 1). In contrast, 2 sets involving Jala (× Guadalupe y Calvo and × La Palma) and Sayula × Guadalupe y Calvo (female from Sayula) had a significantly lower mean number of produced eggs (1.2–1.5 eggs/♀/day) (t = 4.33–6.22, df = 9, P < 0.01; Table 1).

Percentages of embryonated eggs from unsuccessful parental crosses varied from 88.1% to 92.5%, with no significant differences among sets (P > 0.05; Table 1). Those specimens involved in unsuccessful crosses were shown to be fertile when crossed with virgin specimens of the same parental cohort.

A similar phenomenon to that in their parental crosses was observed in the percentage (100%) of success in F1 × F1 crosses (in both directions) of Sayula × Jala and Guadalupe y Calvo × La Palma and in all control crosses (Table 1). In contrast, even when copulation events were observed, a lower percentage of successful crosses concerning parental crosses was observed in those 4 crosses involving Santa Catarina (in both directions) and of Sayula × Guadalupe y Calvo (in both directions), which produced offspring in only 10–20% of crosses (Table 1).

Similar to parental crosses, the lowest success percentages (0–20%, mostly <10%) were observed in those 4 sets of crosses (in both directions) where the population of Santa Catarina was involved and the set of crosses (in both directions) of Sayula × Guadalupe y Calvo (Table 1). Meanwhile, crosses of La Palma × Sayula and × Jala and Jala × Guadalupe y Calvo were successful in only half of the attempts (40–60%) (Table 1), with lower success rates compared to parental crosses. In all cases, copulation events were observed. Those specimens involved in unsuccessful crosses were fertile when crossed with virgin specimens of the same parental cohort.

The mean egg production of F1 × F1 crosses varied from 1.1 to 1.5 eggs/♀/day in most sets of crosses (Table 1). In contrast, 2 sets involving Sayula (× Jala and × La Palma and Guadalupe y Calvo and × La Palma, in both directions), as well as most control crosses (except Santa Catarina) produced the highest mean number of eggs (1.9–2.4 eggs/♀/day) (t = 3.33–7.18, df = 9, P < 0.01; Table 1).

Like parental crosses, percentages of embryonated eggs from unsuccessful F1 × F1 crosses varied from 88.0% to 93.4%, with no significant difference among sets (P > 0.05; Table 1). Those specimens involved in unsuccessful crosses were fertile when crossed with virgin specimens of the same parental cohort.

The percentage of success in F2 × F2 crosses showed a similar pattern to that in F1 × F1 crosses, where crosses (in both directions) of Sayula × Jala and Guadalupe y Calvo × La Palma as well as all control crosses produced offspring in all crosses (100%). The percentages of success of crosses of Sayula × La Palma (in both directions) and of Jala × Guadalupe y Calvo remained similar (40–50%) to those in the F1 × F1 crosses (Table 1). Those crosses involving Santa Catarina (in both directions) had offspring in only one cross in each direction when crossed with Guadalupe y Calvo and had no offspring when crossed with the other 3 studied populations (Table 1). Similarly, no offspring were obtained when specimens from Sayula × Guadalupe y Calvo (in both directions) were crossed (Table 1). Those specimens involved in unsuccessful crosses were fertile when crossed with virgin specimens of the same parental cohort.

Discussion

Studies on hybridization are useful tools for analyzing phenomena that may lead a population to divergence and speciation. Although the results of experimental crosses under laboratory conditions can be biased since organisms of different groups (species, subspecies, or isolated populations) are forced (e.g., breaking ecological and geographical barriers) into contact, these kinds of studies may help in understanding the systematics of a group (Martínez-Ibarra et al. 2011, Villacís et al. 2020).

Integrative taxonomy drives the integration of different analyses to characterize a taxon by cumulation or congruence, which contributes to establishing the proper status (e.g., species, subspecies, population) of that taxon (Dayrat 2005, Alevi et al. 2020, 2021, Vicente et al. 2022). Even when T. longipennis is considered a single species (Salazar-Schettino et al. 2010), some important differences have been noticed when comparing populations, including those recorded in the current study (Martínez-Ibarra et al. 2013a, b, Martínez-Hernández et al. 2021). In contrast, crosses were performed between populations of T. sordida (Stål) from Brazil and Bolivia (possibly representing cryptic species), and the viability of the offspring was analyzed. Based on the absence of reproductive barriers and the high genetic distance, these allopatric populations were confirmed to represent a single taxon (Madeira et al. 2021).

The results of the current study indicated different degrees of reproductive isolation among the studied populations. A low degree of isolation was observed between the populations from Jala and Sayula and between Guadalupe y Calvo and La Palma. The geographic proximity of the populations from Jala and Sayula is consistent with that closeness recorded from a genetic study, where those populations were genetically close and weakly isolated from each other (Martínez-Hernández et al. 2021). No isolation mechanisms were apparent between those two populations. Similar results were obtained after comparing some biological parameters (life cycle, fecundity, fertility, percentages of females at the end of the cycle) of those populations (Martínez-Ibarra et al. 2013a, b). This congruence in the results of studies of multiple and complementary parameters led us to consider populations of Jala and Sayula as belonging to the same species, T. longipennis.

In contrast, even when our current data support the proposal of low isolation between the populations of Guadalupe y Calvo and La Palma, despite being geographically distant (≈ 900 km), complementary studies on biological parameters, as well as genetic distances, do not support this statement (Martínez-Ibarra et al. 2013a, b, Martínez-Hernández et al. 2021). This could be accounted for if both populations had originated from a possible recent diversification from the common ancestor (Pinotti et al. 2021) and were founded by some specimens likely originating in Jalisco, as previously proposed (Martínez-Hernández et al. 2021). That lack of congruence between different studies precludes concluding the proper status of those two populations of T. longipennis.

The population from La Palma has a moderate postzygotic degree of isolation from those from Sayula and Jala. A similar degree of isolation was observed between this last population and that from Guadalupe y Calvo. The first relationship fits with the moderate degree of isolation among La Palma concerning Jala and Sayula, molecularly detected in a previous study (Martínez-Hernández et al. 2021). In agreement, the results of most different studied biological parameters reported in previous studies were similar among the 3 studied populations (Martínez-Ibarra et al. 2013a, b, 2017). Since evidence of copulation events was observed, and embryo formation was verified among all studied pairs of T. longipennis, it is suggested that a postzygotic mechanism could be acting by elimination of hybrids when they are in the zygote or embryo stage (hybrid inviability), precluding the possibility that some pairs had had offspring. This evidence suggests that those populations from La Palma, Sayula, and Jala but are very likely under an incipient process of differentiation (Martínez-Hernández et al. 2021). A similar phenomenon of hybrid inviability was observed when specimens of Mepraia spinolai (Porter) were crossed with M. gajardoi Frias, Henry & Gonzalez (Campos-Soto et al. 2016), and specimens of T. sordida were crossed with T. rosai (Alevi et al. 2020). When the results of our current study are grouped with those from previous studies on biological parameters (Martínez.-Ibarra et al. 2013a, b), feeding and defecation behaviors (Martínez-Ibarra et al. 2021) as well as genetic distances of the populations of Jala and Guadalupe y Calvo (Martínez-Hernández et al. 2021), they show that these two populations had a notable degree of differentiation. That congruence between different studies led to the conclusion that those two populations of T. longipennis were founded in 2 separate events by some specimens likely originating in Jalisco, as previously proposed, and probably initiated a diversification process (possibly representing incipient species) (Martínez-Hernández et al. 2021).

Our data show that the populations from Sayula and Guadalupe y Calvo are highly isolated from each other since only 10% of the F1 × F1 experimental crosses were successful, and no F3 offspring were obtained. These data fit with the isolation degree between these two populations, which was previously detected in a study on the population genetics of T. longipennis (Martínez-Hernández et al. 2021). Both studies are congruent with the important differences recorded when the biological parameters of both populations were compared (Martínez-Ibarra et al. 2013a). Apparently, some different postzygotic reproductive barriers are present (i.e., hybrid inviability, hybrid breakdown) since very few offspring were obtained. The presence of similar postzygotic isolation mechanisms was observed in crosses of M. spinolai and M. gajardoi, two sympatric but highly reproductive isolated triatomine species from the coast of Chile (Campos-Soto et al. 2016). This accumulation of diverse evidence in our current study suggests that the populations of T. longipennis from Sayula and Guadalupe y Calvo are highly reproductively isolated (even under isolation by distance) and under an advanced differentiation process. This leads to their separation into cryptic species (Martínez-Hernández et al. 2021), as recently reported for T. pallidipennis (Stål), a species phylogenetically close to T. longipennis (Cruz and Arellano 2022).

Finally, the results of the current study show that the population from Santa Catarina is markedly reproductively isolated from the other 4 studied populations. None of the 8 sets of parental crosses surpassed 30% success, decreasing from 0 to 20% in F1 × F1 crosses to 0% (except in those crosses × Guadalupe y Calvo, 10%) in F2 × F2 crosses. It is likely that some postzygotic (hybrid inviability and hybrid breakdown) mechanisms are contributing to these findings, as well as observed for hybrids resulting from crosses between T. lenti Sherlock and Serafim and T. sherlocki Papa et al. (Mendonça et al. 2014).

These results fit with a molecular study, and after phylogenetic reconstruction and genetic structure of population analysis, the population from Santa Catarina was found to be genetically separated from the populations from Jala, Sayula, La Palma, and Guadalupe y Calvo (Martínez-Hernández et al. 2021). Likewise, some previously studied biological parameters related to the life cycle, fecundity, fertility, feeding, and defecation behaviors of the population from Santa Catarina distinguish it from other studied populations (Martínez-Ibarra et al. 2013a, b, 2021). Considering all available evidence, among the 5 studied populations, the Santa Catarina population is the most differentiated and apparently represents a cryptic species related to T. longipennis, as previously suggested (Martínez-Hernández et al. 2021).

The analysis of the overall size as well as of the color of the spots in the connexivum showed an apparent influence of local environmental conditions since those specimens from western Mexico (Jala, Sayula, La Palma) were “short” and had orange spots, whereas those from the North (Guadalupe y Calvo) were “long” and had yellow spots. Finally, those specimens from Santa Catarina (a semidry area) were “short” and had yellow spots. That morphological variability influenced local particular environmental conditions has been previously recorded among populations of T. mexicana (Herrich-Schaeffer) from different localities of the states of Hidalgo, Queretaro, and Guanajuato in Central Mexico (Rivas et al. 2021). The morphological characteristics of short overall length and the orange color in the connexivum were dominant in the obtained offspring, concerning those longer specimens, some with yellow spots in the connexivum. The dominance of some morphological characteristics over others (e.g., color in the pronotum, in the scutellum, and the connexivum) has been previously recorded when T. brasiliensis Neiva species complex (Pinotti et al. 2021b) or some triatomine species of the former genus Meccus (synonymized to Triatoma [Cesaretto et al. 2021]) were intercrossed (Martínez-Ibarra et al. 2011).

Using integrative taxonomy data (supported by data available in the literature), we can draw some conclusions. The studied populations of T. longipennis could initially be separated into 3 groups: (i) Sayula (Jalisco), Jala (Nayarit), and La Palma (Michoacan); (ii) Guadalupe y Calvo (Chihuahua); and (iii) Santa Catarina (Guanajuato). The populations from Sayula, Jala, and La Palma (all in western Mexico) are lineages of a metapopulation, defined as “an inclusive population made up of connected subpopulations extended through time” (Pavan et al. 2021). A similar conclusion was reached with some studied populations of T. mexicana from the states of Hidalgo and Guanajuato in central Mexico (Martínez-Ibarra et al. 2023). The populations from Guadalupe y Calvo, the extreme northern distribution of this species (Salazar-Schettino et al. 2019), are undergoing a process of divergence caused by the low genetic flow with geographically far (900–1,000 km.) populations of T. longipennis (isolation by distance) and by the very difficult access to its distribution area in Chihuahua. The proposal of a divergence process is also supported by the fact that the populations of T. longipennis from Chihuahua have been collected only from sylvatic areas (mainly caves) close to human dwellings (Licón-Trillo et al. 2010). In contrast, specimens of T. longipennis from western and central areas of Mexico were collected in peridomestic and domestic areas of human dwellings (Salazar-Schettino et al. 2019).

The studied population from Santa Catarina is the extreme southern distribution of T. longipennis in Mexico (Salazar-Schettino et al. 2019). The results of this current study, as well as previous population genetics and biological parameters, lead to the conclusion that this population is apparently under an allopatric process of speciation since this population is geographically isolated from other populations of T. longipennis because Santa Catarina is located inside the Sierra Gorda mountain range (GEG 2022). This population of T. longipennis is also isolated by the distance of some previously collected populations of this species in Guanajuato (≈100 km north) (López-Cárdenas et al. 2005). If the hypothesis is true that gene flow between Santa Catarina and the remaining populations is absent or, if present, largely irrelevant, that population would accumulate mutations independently, develop some degree of genetic divergence, and might become genetically isolated. Complete allopatric speciation can occur if populations of incipient species develop pre- or postzygotic barriers to reproduction (Pavan et al. 2021). A similar phenomenon seems to be present in some populations of T. mexicana, a phylogenetically close species to T. longipennis (Aguilera-Uribe et al. 2020). Recent studies (morphological, morphometrical, on population genetics and crosses of populations) have been carried out to better understand the variability of T. mexicana populations. Those studies were congruent, recording important differences between some populations from the Mexican states of Hidalgo and Queretaro concerning those from Guanajuato (Rivas et al. 2021, 2022, Martínez-Ibarra et al. 2023). Interestingly, the locality (Tierra Blanca) from which the population of T. mexicana from Guanajuato was collected is no farther than 20 km from Santa Catarina, the locality where a remarkably different population of T. longipennis was collected. This locality is also inside the Sierra Gorda mountain range, which isolated it, such as Santa Catarina (GEG 2022).

Our data are limited by the small number of studied populations of T. longipennis, as well as because of the discontinuous collections in the distribution areas of this species, skipping the state of Sinaloa, placed between Nayarit and Chihuahua. Collection in this area will resume as soon as it is safe to travel to Sinaloa. Even when limited, our results contribute to the elucidation of the potential presence of a cryptic species complex of T. longipennis in Mexico.

Further studies with a wider collection, additional analyses (morphometric, molecular, biological), and more populations of T. longipennis are now necessary to fully define the variability among their populations.

Acknowledgments

We thank Alejandro Martínez-Pérez for his technical advice. We thank inhabitants of studied localities for the donation of specimens.

References

Author notes