Abstract
Hybrid male sterility contributes to speciation by restricting gene flow between related taxa at the beginning stages of postzygotic isolation. However, we have limited knowledge about cellular processes and molecular mechanisms that come into play when fertility is first affected. Hybrids between closely related species of the Anopheles gambiae complex offer opportunities to identify spermatogenic errors that arise early during speciation. To investigate possible cellular causes of hybrid male sterility, we performed crosses between sibling species of the An. gambiae complex. Our results demonstrate that testes are severely underdeveloped in hybrids between male An. merus and female An. gambiae or An. coluzzii. No meiotic chromosomes are identified in these hybrid males. However, testes have nearly normal morphologies and sizes but produce mostly nonmotile spermatozoa in hybrids from the reciprocal crosses. Using chromosome X– and Y-specific fluorescent probes, we followed the process of meiosis in each species and their F1 hybrids between female An. merus and male An. gambiae or An. coluzzii. Unlike for pure species, sex chromosomes in meiotic prophase I of F1 hybrids are largely unpaired and all chromosomes show various degrees of insufficient condensation. Instead of entering the reductional division in meiosis I, primary spermatocytes undergo an equational mitotic division producing abnormal diploid sperm. Meiotic chromosomes of some F1 hybrid individuals are involved in de novo genome rearrangements. Yet, the germline-specific genes β2-tubulin, Ams, mts, and Dzip1l express normally in these hybrid males. Thus, our study identified cytogenetic errors in hybrids that arise during the early stages of postzygotic isolation. This knowledge will inform the development of innovative mosquito control strategies based on population suppression by manipulating reproduction via genetic technologies.
Author Summary The genetic basis and molecular mechanisms of hybrid male sterility are of considerable interest as they inform our understanding of both speciation and normal fertility function. Studies of sterility in male hybrids between recently evolved species offer opportunities to identify developmental errors that arise early in speciation. We performed crosses between sibling species of the Anopheles gambiae complex to gain insights into a cellular basis of postzygotic isolation. We demonstrate that hybrid male sterility in the malaria mosquitoes is caused by two processes in reciprocal crosses: premeiotic arrest in germline stem cells and the failure of the reductional meiotic division in primary spermatocytes. The meiotic abnormalities also include unpairing of sex chromosomes, chromatin decompaction, and, in some cases, de novo genome rearrangements. The failure of the reductional division in meiosis I results in the production of diploid nonmotile sperm. Despite these meiotic errors, tested germline-specific genes express normally in these hybrid males. Thus, our study identified cellular errors in hybrids that arise during the early stages of postzygotic isolation. Studying molecular mechanisms of the developmental abnormalities in testes of hybrids between closely related species of mosquitoes will improve our knowledge of speciation and empower the sterile insect technique.
Introduction
Hybrid male fertility is among the first phenotypes affected as the postzygotic isolation between species is being established [1]. Therefore, genetic factors, cellular basis, and molecular mechanisms of hybrid male sterility are of considerable interest, as they inform our understanding of both speciation and normal fertility function. Genetic loci responsible for hybrid male sterility in plants and animals are often being identified using quantitative trait loci (QTL) mapping [1–8]. In a variety of organisms, hybrid sterility can be caused by the genetic factors on the X chromosome or by epistasis between X-linked and autosomal alleles [4, 8–15]. Studies in animals commonly assess testis shape, size or weight, as well as sperm morphology, density or motility to define male sterility phenotypes. However, a cellular basis of hybrid male sterility is rarely investigated. Cytological studies of spermatogenesis in sterile hybrids indicate that multiple mechanisms of functional sterility are possible. Sterile hybrids between Mus musculus domesticus and M. m. musculus have spermatogenic arrest in early meiosis I with disrupted both homoeologous chromosome pairing and meiotic sex chromosome inactivation [16, 17], as well as reductions in spermatocyte and spermatid number, increased apoptosis of primary spermatocytes, and more multinucleated syncytia [18]. Studies of failed spermatogenesis in different Drosophila hybrids observed arrests at the premeiotic stage [19], reduced chromosome pairing and unequal chromosome segregation in meiosis [20, 21], characteristic spermiogenic arrests [19], spermatid abnormalities [22], and problems in sperm bundling and motility [23, 24]. Detailed analyses of cellular and developmental abnormalities in testes of various hybrid organisms could help better define the causes of infertility. Cytological studies of spermatogenesis in hybrids may link different sterility phenotypes to the early or advance stages of postzygotic isolation. Moreover, identification of first-evolved cytogenetic errors in hybrids between closely related species could improve our understanding of speciation.
The Anopheles gambiae complex consists of at least eight morphologically nearly indistinguishable sibling species of African malaria mosquitoes [25, 26]. Genome-based estimations of the age of the An. gambiae complex vary from 1.85 [27] to as young as 0.526 million years [28]. Genomic introgression has been found prevalent in autosomal regions of several species indicating naturally occuring interspecies hybridization [27, 28]. Experimental crosses of species from the An. gambiae complex often produce sterile F1 hybrid males confirming Haldane’s rule of sterility or inviability of the heterogametic sex [13, 29–31]. Early crossing experiments between members of the complex have found that hybrid male sterility is associated with various degrees of atrophies of testes and underdevelopment of sperms [30]. Examination of hybrid sterility in backcrosses of An. gambiae and An. arabiensis using an X-linked white-eye marker demonstrated a large effect of the X chromosome on male hybrid sterility [32]. The large X effect on hybrid sterility has been supported with a QTL mapping approach [13]. However, an introgression of the Y chromosome from An. gambiae into a background of An. arabiensis has shown no apparent influence on male fertility, fitness, or gene expression [29]. A recent study of hybridization in nature has shown that postmating isolation is positively associated with ecological divergence of An. coluzzii, An. gambiae, and An. arabiensis [33]. However, the cellular basis and molecular mechanisms of hybrid male sterility in mosquitoes are unknown. Because of the recent evolution and ease of hybridization, sibling species of the An. gambiae complex offer great opportunities to provide insights into mechanisms of speciation.
A recent World Health Organization report has shown that after a successful period of global malaria control, progress has stalled since 2015. There are still 219 million cases and 435,000 related deaths from malaria without a tendency to decrease [34]. In the 60s-70s, the sterile insect technique (SIT) was one of the first methods used to control the malaria-transmitting Anopheles mosquitoes as it aimed to interfere with their reproduction by introducing sterile males into natural populations. However, because of the low competitiveness of X-ray-irradiated males and the lack of institutional commitments, the traditional SIT largely failed in controlling malaria mosquitoes [35]. Novel genetic approaches based on CRISPR-Cas9 gene drives [36–40] show potential toward the generation of self-sustainable and species-specific mosquito control strategies. A better understanding of the normal male fertility function and mechanisms of naturally occurring hybrid male sterility will inform the development of novel SIT tools and implementation of pre-existing technologies for the control of malaria-transmitting mosquitoes.
Here, we investigate possible cellular causes of male sterility in hybrids between sibling species of the An. gambiae complex. We asked three specific questions: (i) Which cellular processes are involved in causing infertility in hybrid mosquito males? (ii) Are the defects leading to hybrid male sterility premeiotic, meiotic, or postmeiotic? and (iii) What cytogenetic errors trigger the spermatogenic breakdown? We demonstrate that hybrid male sterility in malaria mosquitoes is caused by two cellular processes in reciprocal crosses: premeiotic arrest in germline stem cells and the failure of the reductional meiotic division in primary spermatocytes. Our data suggest that the meiotic abnormalities in hybrid males stem from the unpairing of the sex chromosomes and chromatin decondensation. Thus, our study identified first cytogenetic errors in hybrids that arise during the early stages of postzygotic isolation.
Results
Hybrid male sterility phenotypes at the cellular level
To obtain sterile males, we performed reciprocal crosses between An. merus MAF and An. gambiae ZANU or An. coluzzii MOPTI and MALI. Backcrossing of F1 males to parental females resulted in no progeny demonstrating sterility of F1 hybrid males. This result confirms Haldane’s rule for the majority of interspecies crosses in the An. gambiae complex, except for crosses between An. gambiae and An. coluzzii: F1 females are fertile while F1 males are sterile [13, 29–31, 41, 42]. To investigate the developmental phenotypes associated with hybrid sterility, we dissected testes from adult males obtained from interspecies crosses and from pure species. Normal testes of pure species have a spindle-like shape (Fig 1A). We found obvious asymmetry in morphology and sizes of testes when reciprocal crosses were compared. Hybrid males from crosses between female An. merus and male An. gambiae or An. coluzzii display normal-like reproductive organs (Fig 1B). In contrast, F1 males from crosses between female An. gambiae or An. coluzzii and male An. merus show severely underdeveloped testes (Fig 1C). We then tested if normal-like and underdeveloped testes of interspecies hybrids produce any sperm. In squashed testes of 2-day-old adults of pure species, large amounts of mature spermatozoa with long tails can be seen. After we crushed the testes, spermatozoa with vibrant motility escaped from the ruptures (Fig 1A, S1 Movie). However, mature sperm or sperm motility hardly could be seen in squashed or crushed normal-like testes of 2-day-old adult hybrids from crosses when An. merus was the mother. Instead, we see delayed spermatid differentiation, fewer spermatids, and mostly nonmotile spermatozoa with large heads and often two short tails growing from opposite ends of the head (Fig 1B, S2 Movie). Only undifferentiated round cells could be seen in underdeveloped testes of hybrids from reciprocal crosses when An. merus was the father (Fig 1C). Given the small size of the degenerate testes, these round cells may represent germline stem cells. Thus, neither normal-like nor underdeveloped testes of the interspecies hybrids produce mature motile spermatozoa.
The normal progress of meiosis in males of pure species
Before we investigate meiosis of interspecies hybrids, we provide the first description of chromosome behavior in meiosis of pure Anopheles species. The chromosome complement of Anopheles males consists of three chromosome pairs: two autosomes (2 and 3) and X and Y sex chromosomes. With the help of sex-chromosome-specific fluorescent probes, we followed the progress of meiosis in An. gambiae, An. coluzzii, and An. merus. The following DNA sequences were used as the probes for fluorescence in situ hybridization (FISH) (Table 1). Retroelement zanzibar is specific to the Y chromosome of An. gambiae and An. coluzzii but is absent in An. merus; 18S ribosomal DNA (rDNA) labels only the X chromosome in An. gambiae or An. coluzzii but labels both X and Y chromosomes in An. merus; and satellite AgY53B [43] labels both X and Y chromosomes in all three species [44]. Fig 2 shows normal activities of meiotic chromosomes in testes of An. gambiae ZANU. In primary spermatocytes, all homologous autosomes and sex chromosomes pair and display chiasmata in diplotene/diakinesis of prophase I, they align with each other at the cell equator in metaphase I, and then move from each other in anaphase I. In secondary spermatocytes, sister chromatids of each chromosome align with each other at the cell equator in metaphase II and go to opposite poles of the cell during anaphase II. Meiotic divisions produce spermatids that contain a haploid set of autosomes and either a Y or X chromosome. An. coluzzii males show similar morphology and behavior of chromosomes during meiosis (S1 Fig). Because retroelement zanzibar is absent in An. merus, we used satellite AgY53B and 18S rDNA to label sex chromosomes in this species. Each of these probes hybridized with both X and Y chromosomes in An. merus, making discrimination between X and Y more difficult in this species. However, we could differentiate metaphase sex chromosomes by the euchromatic arm of the X chromosome and by the slightly larger distal heterochromatic block of the Y chromosome (S2 Fig). Heterochromatic parts of the X and Y chromosomes are relatively large and structurally similar in An. merus in comparison with An. gambiae or An. coluzzii. In the latter two species, the heterochromatic parts of the X and Y chromosomes are relatively small and substantially different from each other both in size and genetic content [44]. Despite these differences between An. merus and either An. gambiae or An. coluzzii, the activities of meiotic sex chromosomes in testes are similar in all three species.
Meiotic failures in F1 males of interspecies hybrids
To determine possible cytogenetic mechanisms of hybrid male sterility, we analyzed chromosome behavior in testes of hybrids from interspecies crosses. Meiotic chromosomes were present in normal-like testes of hybrids from the ♀An. merus × ♂An. coluzzii/An. gambiae rosses and we identified important abnormalities of meiosis in these males (Fig 3). In primary spermatocytes, homoeologous autosomes pair and form chiasmata in prophase I as in pure species, but X and Y chromosomes do not display chiasmata in diplotene/diakinesis of prophase I. We found this pattern consistent in all analyzed hybrid males. In hybrids, metaphase chromosomes are visibly longer than at the same stage in pure species, indicating insufficient chromatin condensation. Besides, homoeologous chromosomes in hybrids do not segregate during anaphase. Instead, sister chromatids move to opposite poles of the dividing cell. Because reductional division does not occur in hybrid males, both X and Y chromatids move to the same pole during anaphase. As a result, haploid secondary spermatocytes do not form in these males. Our FISH analysis demonstrated that each spermatid in testes of pure species normally contains either an X or Y chromosome. In contrast, we found that both X and Y chromosomes present in each spermatid of the hybrids (S3 Fig). Moreover, the abnormal spermatids are larger in size due to insufficient chromatin condensation and the double chromosome content. Thus, we discovered that chromosomes in normal-like testes of hybrid males start with a meiotic behavior in prophase and then switch to a mitotic behavior in anaphase. The equational division of primary spermatocytes results in dysfunctional diploid sperm in hybrids if An. merus is the mother. In contrast, degenerate testes of F1s from the reciprocal ♀An. coluzzii/gambiae × ♂An. merus crosses have only undifferentiated round germline stem cells. To visualize sex chromosomes, we performed whole-mount FISH with labeled 18S rDNA and satellite AgY53B to mark the X chromosomes of An. coluzzii and the Y chromosomes of An. merus, respectively (S4 Fig). Only interphase sex chromosomes in nuclei of germline stem cells are detected indicating that meiosis does not start in the underdeveloped testes of F1 hybrids if An. merus is the father. Thus, the premeiotic arrest in the degenerate testes is the reason for the lack of spermatids in these hybrids.
Chromosomal and molecular abnormalities in interspecies hybrids
Here, we performed quantitative analyses of chromatin condensation and the X-Y chromosome pairing in pure species and their hybrids. We also tested if germline-specific genes express in sterile male hybrids from the reciprocal crosses. To determine the extent of chromatin condensation in normal-like testes of interspecies hybrids in comparison with pure species, we measured the lengths of metaphase chromosomes (S1 Table). The results of a statistical analysis with a two-sample pooled t-test show that chromosomes in F1 hybrids are typically longer than chromosomes in An. coluzzii, An. gambiae, or An. merus (Fig 4A, S5 Fig). For example, chromosomes of the An. merus origin in a hybrid background always show significant (P<0.001) elongation by at least 1.3 times in comparison with pure An. merus. The X chromosome of An. merus suffered the most serious undercondensation in hybrids exceeding the length of the X chromosome in the pure species background by 1.6-1.9-fold. Probably because the An. merus X is the longest chromosome in the hybrid karyotype, it can be subject to segregation delays during anaphase (S6 Fig), possibly causing slowing down of the cell division.
In our cytogenetic study of interspecies hybrids, the X and Y chromosomes do not show pairing or chiasmata in diplotene/diakinesis of prophase I (Fig 3). We hypothesized that sex chromosome pairing is affected in the early prophase I when individual chromosomes cannot be distinguished by direct visualization. To analyze the X-Y chromosome pairing at the pachytene stage of prophase I, we performed a whole-mount FISH and examined spatial positions of the X– and Y-specific fluorescent signals in confocal optical sections of nuclei in testes of pure species and their hybrids (S7 Fig). We recorded the number of nuclei with X and Y fluorescent signals colocalized versus X and Y fluorescent signals located separately (Fig 4B). The results of a statistical analysis with a two-sample pooled t-test demonstrate that X and Y chromosomes pair in more than 90% of primary spermatocytes in the An. coluzzii, while they pair in less than 30% of primary spermatocytes in F1 hybrids of ♀An. merus × ♂An. coluzzii (Fig 4C).
To test if premeiotic or meiotic failures are associated with the misexpression of germline-specific genes, we analyzed the presence of the postmitotic germline transcripts Ams, mts, Dzipll [45], and β2-tubulin [47, 48] in F1 hybrid males (Table 1). The reverse transcription polymerase chain reaction (RT-PCR) results show that these genes express at similar levels in the reproductive tissues of An. coluzzii and F1 hybrids from the ♀An. merus × ♂An. coluzzii MOPTI cross (Fig 4D). However, Ams, mts, Dzip1l, and β2-tubulin are strongly down regulated in gonads of the ♀An. coluzzii MOPTI × ♂An. merus F1 hybrids, supporting our observation that meiosis does not occur in these hybrids. In contrast, a pre-meiotic gene, vasa [49] (Table 1), expresses at similar levels in reproductive tissues of all hybrids and pure species, indicating that germline stem cells present even in degenerate testes of interspecies hybrids.
Besides the abnormalities that affect all progeny of the ♀An. merus × ♂An. coluzzii/An. gambiae crosses, chromosomes of some hybrid individuals were involved in de novo genome rearrangements (Fig 5). For example, a large fragment of the An. merus X chromosome, including the rDNA locus, was translocated to the 2L arm of An. coluzzii in the male hybrid from the ♀An. merus × ♂An. coluzzii MALI cross (Fig 5A). In addition, we detected a duplication of a chromosomal segment involving the rDNA locus within the X chromosome of An. merus in the ♀An. merus × ♂An. gambiae ZANU cross (Fig 5B). These rearrangements were not observed in pure species, and their occurrence in F1 hybrid males suggests an increased genome instability as a result of interspecies hybridization.
Discussion
Cellular and molecular phenotypes associated with hybrid male sterility
In this study, we performed the first detailed cytological analysis of spermatogenesis in pure species and hybrids of mosquitoes. Hybrid male sterility phenotypes in the An. gambiae complex are clearly asymmetric between the reciprocal crosses (Fig 6), as has been demonstrated earlier for mosquitoes and other animals [8, 20, 30, 50]. The observed cellular and molecular abnormalities suggest the following developmental scenarios in testes of the interspecies hybrids. In hybrids from crosses with An. merus as the father, mitotic divisions of germline cells are impaired, no meiosis occurred, and expression of postmitotic germline-specific genes is repressed (Fig 4D). As a result, these hybrids have degenerate testes with arrested germline stem cells. Degenerate testes have also been observed in F1 hybrids between sibling species of the An. albitarsis [51] and the An. barbirostris [52] complexes. However, in hybrids from the reciprocal crosses with An. merus as the mother, mitotic divisions of germ cells occur as normally as in pure species. Moreover, germ cells in these hybrids go through meiotic prophase I as evidenced from our cytogenetic analyses (Fig 3) and expression of the postmitotic germline genes (Fig 4D). However, the downstream developmental processes in these hybrids are impaired. The abnormalities begin in meiosis I where homoeologous chromosomes fail to segregate. Testes of these hybrids show delayed spermatid differentiation, a smaller number of these cells, and eventual formation of immotile abnormal (including two-tailed) sperm with insufficient chromatin condensation (Fig 6). Thus, we demonstrate that premeiotic and meiotic failures associate with hybrid male sterility in malaria mosquitoes. This observation is at odds with the commonly accepted view that more often hybrid males suffer postmeiotic sterility problems [1, 2]. Although many postmeiotic defects seen in Drosophila hybrids are indeed related to problems in sperm bundling and motility [2, 19, 23], some sperm abnormalities may stem from meiotic failures. For example, our results demonstrate that such sperm abnormalities as nonmotility, two-tailed heads, and chromatin decompaction in sperm heads result from the impaired meiosis I. However, very few studies have been devoted to the detailed cytological investigation of meiosis in interspecies hybrids, especially in dipteran insects. The closest studies to our own were performed by Theodosius Dobzhansky in 1933 and 1934, in which he described pronounced defects in meiosis I in sterile male hybrids from the ♀D. pseudoobscura × ♂D. persimilis cross [20, 21]. Also, a histological investigation of spermatogenesis in hybrid mice identified defects in meiosis I as a primary barrier to reproduction [18]. Additional studies of interspecies hybrids of various organisms should determine if the first meiotic division commonly fails when fertility is affected.
Cytogenetic mechanisms of hybrid male sterility
Our cytogenetic investigation of meiosis in pure Anopheles species revealed that the X and Y chromosomes normally pair tightly with each other throughout prophase I (Figs 2, 4BC, and S1, S2 Figs). Sex chromosomes in malaria mosquito males genetically recombine [44] highlighting an important difference from male meiosis in fruit flies [53–55]. As a result of the recombination, some mosquito species, such as An. merus, have homomorphic heterochromatic parts in their X and Y chromosomes (S2 Fig). Future sequencing and assembly of the heterochromatin in malaria mosquitoes using long reads may yield important insights into structural and functional organization of the sex chromosomes, as it has been demonstrated for Drosophila [56].
Unlike pure species, the meiotic prophase I in F1 mosquito hybrids shows cytogenetic anomalies—low percentage of the X-Y chromosome pairing and insufficient chromatin condensation (Fig 7). Disruption of chromosome pairing and synapsis is frequently observed in interspecies hybrids of various organisms. Similar to the mosquitoes, pairing of X and Y chromosomes is more adversely affected than that of autosomes during the prophase I in male hybrids between Campbell’s dwarf hamster and the Djungarian hamster [57]. A recent work has clarified that the autosomes of male hybrids between these hamster species undergo paring and recombination as normally as their parental forms do, but the heterochromatic arms of the X and Y chromosomes show a high frequency of asynapsis and recombination failure [58]. Another study has demonstrated a high rate of synaptic aberrations in multiple chromosomes of male hybrids between two chromosome races of the common shrew [59]. It has been proposed that asynapsis of heterospecific chromosomes in prophase I may provide a recurrently evolving trigger for the meiotic arrest of interspecific F1 hybrids of mice [8, 16]. Indeed, the presence of chiasmata and tension exerted across homologs ensures that yeast cells undergo reductional segregation [60]. Whenever it occurs, an asynapsis of chromosomes almost invariably triggers pachytene checkpoint and meiotic breakdown [61]. The genetic determinants of chromosome pairing and synapsis in both plants and animals are of great interest. In bread wheat, the Ph1 locus has the largest effect on preventing homeologous pairing in meiosis [62]. In mice, a null mutation in the PRDM9 gene causes chromosome asynapsis, arrest of spermatogenesis at pachynema, impairment of double-strand break repair, and disrupted sex-body formation [63].
The PRDM9-null phenotype resembles observed univalents and frequent X-Y dissociation in interspecies hybrids of mice [17]. A recent study has demonstrated a genetic link between meiotic recombination and hybrid male sterility [12]. It has been suggested that within species there may be selection to maintain sequence homozygosity for meiosis genes because their divergence can lead to reproductive isolation through failures of double-strand break formation and synapsis in hybrids [62].
It is possible that the absence of chiasmata between the X and Y chromosomes causes failure of reductional segregation in the mosquito hybrids. Unlike hybrids of mice, primary spermatocytes in the mosquito hybrids do not undergo arrest in meiosis I. Instead, they continue dividing by mitosis, in which sister chromatids move to the opposite poles of the dividing cell (Fig 7). This meiosis-mitosis switch is likely caused by the bi-orientation instead of the normal monoorientation of sister kinetochores, in which tension across the centromere regions and proteins shugoshins plays a key role [53, 64, 65]. It is known that kinetochore-microtubules attachment errors are more commonly found in meiosis I than in mitosis [66, 67]. A mathematical model explains why kinetochore-microtubules attachment errors occur more frequently in the first meiotic division than in mitosis. The model suggests that the gradual increase of microtubules may help turn off the spindle assembly checkpoint in meiosis I leading to chromosome mis-segregation errors [68]. In mosquito hybrids, change in the orientation of sister kinetochores may be caused by reduced tension across the centromeres of homologous chromosomes due to chromatin decompaction. Changes in chromatin condensation have been documented in interspecies hybrids of diverse groups of organisms [69, 70]. In a study of F1 hybrids between Arabidopsis thaliana and A. lyrata, the A. thaliana chromatin became more compact, whereas the A. lyrata chromatin became moderately less compact [70]. An intriguing spatial arrangement model suggests that that Y chromosome-linked variation may alter spatial position and packaging of other chromosomes in the nucleus [71]. Chromosome condensation and spatial position could also be affected in mosquito hybrids by improper function of condensins. A study of Drosophila male meiosis demonstrated that condensin II subunits, Cap-H2 and Cap-D3, are required to promote chromosome territory formation in primary spermatocyte nuclei. Moreover, anaphase I is abnormal in Cap-H2 mutants as chromatin bridges are formed between segregating heterologous and homologous chromosomes [72]. Thus, the interplay between the two phenotypes in mosquito hybrids—sex chromosome unpairing and chromatin decompaction— may result in failing a reductional meiotic division and proceeding to an equational mitotic division.
Evolution of spermatogenesis errors in interspecies hybrids
Reciprocal crosses in malaria mosquitoes and other organisms usually produce hybrids with sterility phenotypes of different degrees of severity [20, 30, 50]. This observation suggests that multiple stages of postzygotic isolation exist based on malfunction of the spermatogenesis. At the early stages of evolution, germline-specific genes still express normally and meiosis proceeds until metaphase I and then switches to an equational anaphase producing diploid sperm cells in hybrids. In this case, meiotic errors are characterized by the dramatically decreased pairing between sex chromosomes and insufficient chromatin condensation, which are seen in sterile hybrids from the ♀An. merus × ♂An. coluzzii/An. gambiae crosses (Figs 3, 7). The observed incidental chromosome rearrangements in mosquito hybrids (Fig 5) support the notion that genome instability is a common characteristic trait of hybrid incompatibility that may be associated with increased transposable element activity, ectopic recombination, and doublestrand DNA breaks [73]. A recent work demonstrated that biogenesis of Piwi-interacting RNA (piRNA) is enhanced in testes of hybrids between Drosophila buzzatii and D. koepferae. The study argues that interspecies hybridization causes a genomic stress that can activate the piRNA response pathway to counteract transposable element deregulation [74]. Because piRNAs predominantly target long terminal repeats (LTR) retrotransposons in both Anopheles and Drosophila [75], future studies of the piRNA expression in mosquito hybrids may identify common and specific responses to the genomic stress in dipteran insect species. As species continue to diverge, meiotic errors become more prominent, and new hybrid phenotypes appear as has been seen, for example, in sterile male hybrids from the ♀D. pseudoobscura × ♂D. persimilis cross [20, 21]. At this stage of postzygotic isolation, an abnormal anaphase is characterized by unequal segregation of chromosomes resulting in spermatids with unbalanced chromosome content. The more advanced interspecies divergence of meiosis is manifested by the unpairing of the most chromosomes and by the malfunction of the abnormally elongated spindle in sterile male hybrids [20, 21]. The next stage of evolution is spermatogenic arrest in early meiosis I with disrupted homoeologous chromosome pairing, meiotic sex chromosome inactivation, and increased apoptosis of spermatocytes as seen in mice hybrids [16–18]. Interestingly, chromosome unpairing and asynapsis in male hybrids stand out as common spermatogenic phenotypes associated with sterility. It has been hypothesized that nongenic repetitive sequences, as the fastest diverging component of the genome, may facilitate asynapsis in hybrids, thus, representing suitable candidates for a Dobzhansky-Muller incompatibility [16]. Findings in Drosophila hybrids demonstrate that rapid evolution of heterochromatin may indeed result in hybrid incompatibilities [69, 73, 76, 77]. Also, the centromere drive model has been proposed to explain how paired chromosomes at meiosis I can be subject to nondisjunction leading to infertility in Drosophila males [78, 79]. Accordingly, incompatibilities between rapidly evolving centromeric components of emerging species, such as co-evolving the CENP-A histone variant and its chaperone CAL1 [80], may account for species incompatibility between centromeric histones and for postzygotic reproductive isolation. Finally, spermatogenic abnormalities in hybrids can happen even before meiosis starts. A spermatogenic arrest at the premeiotic stage is characterized by the repression of germline-specific genes and by the lack of spermatocytes or spermatids in degenerate testes of hybrid males. These phenotypes have been observed in sterile hybrids from the ♀D. mauritiana × ♂D. sechellia cross [19] and in sterile hybrids from the ♀An. coluzzii/An. gambiae × ♂An. merus crosses (our study). Thus, cytological analyses of spermatogenesis in interspecies hybrids can associate sterility phenotypes with early or advance stages of speciation. Such studies of interspecies hybrids from diverse groups of organisms may highlight general patterns and mechanisms in the origin and evolution of postzygotic isolation.
Conclusions
Charles Darwin in the chapter “Hybridism” of his Origin of Species rightly argued that hybrid sterility ‘‘is not a specially endowed quality, but is incidental on other acquired differences [81].’’ Identification of the cellular and molecular differences acquired during the early stages of postzygotic isolation between species is crucial to explaining both speciation and normal fertility function. The cross between a female An. merus and a male An. gambiae or An. coluzzii produces sterile hybrid males with spermatogenic abnormalities in the first meiotic division. Our study uncovered differences in chromosome behaviors between pure Anopheles species and their hybrids. The obtained data suggest that the meiotic abnormalities in hybrid males stem from the unpairing of the sex chromosomes and chromatin decondensation. Therefore, these malaria mosquitoes represent a great new system for studying a genetic basis and molecular mechanisms of species incompatibilities at early stages of postzygotic reproductive isolation.
Knowledge of the mechanisms of reproductive isolation in Anopheles has important implications not only for evolutionary biology but also for malaria control. The success of malaria transmission highly depends on the rate of mosquito reproduction. The development of novel approaches to control the reproductive output of mosquitoes must include the understanding of how fertility is regulated. Determining the mechanisms of sterility in male hybrids between closely related species of mosquitoes can also empower the sterile insect technique. In this respect, our results shed new light on the cellular processes and possible mechanisms in spermatogenesis that first appear when fertility is affected.
Materials and Methods
Mosquito maintenance and crossing experiments
The laboratory colonies of An. gambiae ZANU (MRA-594), An. coluzzii MOPTI (MRA-763), An. coluzzii MALI (MRA-860), and An. merus MAF (MRA-1156) were obtained from the Biodefense and Emerging Infections Research Resources Repository (BEI). Authentication of the species was performed by a cytogenetic analysis and by PCR diagnostics [82, 83]. Mosquitoes were reared at 27±1°C, with 12-hour photoperiod and 70±5% relative humidity. Larvae were fed fish food, and adult mosquitoes were fed 1% sugar water. To induce oviposition, females were fed defibrinated sheep blood (Colorado Serum Co., Denver, Colorado, USA) using artificial blood feeders. To perform interspecies crosses, male and female pupae were separated to guarantee virginity of adult mosquitoes. We differentiated males and females at the pupal stage using sex-specific differences in the shape of their terminalia [84]. After the emergence of adults, crossing experiments were performed by combining 30 females and 15 males in one cage. Five days after random mating, the females were fed sheep blood. Two days later, an egg dish, covered with moist filter paper to keep the eggs from drying out, was put into the cage. Backcrosses of F1 males and parental females were done using a similar method. At least two blood meals were fed to females with three repeats of each crosses.
Male gonad and sperm observation
Male gonads were dissected and photographed using an Olympus SZ 61 stereoscopy microscope (Olympus, Tokyo, Japan) with an Olympus Q-Color 5 digital camera (Olympus, Tokyo, Japan). We observed the testes of 30 males and took pictures of the testes of five males from each cross. For sperm observation, testes were mounted in 20 μl sperm assay buffer containing 4 mM KCl, 1.3 mM CaCl2, 145 mM NaCl, 5 mM D-glucose, 1 mM MgCl2, and 10 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (Hepes) [85]. After gently covering testes with a coverslip, they were crushed, and sperm motility was observed under a phase contrast microscope BX 41 (Olympus, Tokyo, Japan). A Movie of sperm motility for at least five males of hybrids from each cross and five males of pure species was recorded using a digital camera UC90 (Olympus, Tokyo, Japan).
Chromosome preparation
Testes with male accessory glands were dissected from male pupae and 0-12 hours-old adults in 0.075% potassium chloride (KCl) hypotonic solution on a frosted glass slide (Thermo Fisher Scientific, Waltham, MA, USA) under an Olympus SZ 61 dissecting microscope (Olympus, Tokyo, Japan). To observe meiotic chromosomes, male accessory glands and other tissues were removed, and only testes were left on the slide. Immediately after dissection, a drop of 50% propionic acid was added to the testes, and they were covered with a 22×22-mm coverslip (Thermo Fisher Scientific, Waltham, MA, USA). Preparations were tapped using the flat rubber end of a pencil and observed under an Olympus CX41 phase-contrast microscope (Olympus, Tokyo, Japan). The slides were frozen in liquid nitrogen, and a sharp razor was used to take off the coverslips. The preparations were placed in 50% ethanol chilled at -20°C for a minimum of 2 hours. Later, serial dehydrations were performed in 70%, 90%, and 100% ethanol for 5 min each at room temperature (RT). Subsequently, the best preparations from at least 10 slides/individuals with desired meiotic stages were chosen for further studies.
DNA probe labeling and FISH
Three DNA probes were used in this study: retroelement zanzibar, which is specific to the Y chromosome of An. gambiae and An. coluzzii, 18S rDNA, which is specific to the X chromosome of An. gambiae and An. coluzzii but labels both X and Y chromosomes of An. merus, and satellite AgY53B, which labels X and Y chromosomes of all three species [44]. The AgY53B satellite was amplified using primers that target the AgY53B/AgY477 satellite array (Table 1). The probes were labeled by Cy3 or Cy5 fluorochromes using PCR with genomic DNA as a template. Genomic DNA was extracted using DNeasy Blood & Tissue Kits (Qiagen, Hilden, Germany) from virgin males of An. gambiae ZANU or An. coluzzii MOPTI for labeling zanzibar, virgin males of An. merus MAF for labeling satellite AgY53B, and virgin females of An. gambiae or An. coluzzii for labeling 18S rDNA. Each 25 μl of a PCR mix consisted of 1-μl genomic DNA, 12.5-μl ImmoMix™ 2x reaction mix (Bioline USA Inc., Taunton, MA, USA), 1 μl of 10-μM forward and reverse primers, and water. PCR labeling was performed in the Mastercycler® pro PCR thermocycler (Eppendorf, Hamburg, Germany) starting with a 95°C incubation for 10 min followed by 35 cycles of 95°C for 30 sec, 52°C for 30 sec, 72°C for 45 sec; 72°C for 5 min, and a final hold at 4°C. FISH was performed as previously described [86, 87]. Briefly, slides with good preparations were treated with 0.1-mg/ml RNase at 37°C for 30 min. After washing with 2×SSC for 5 min twice, slides were digested with 0.01% pepsin and 0. 037% HCl solution for 5 min at 37°C. After washing slides in 1×PBS for 5 min at RT two times, preparations were fixed in 3.7% formaldehyde for 10 min at RT. Slides were then washed in 1×PBS and dehydrated in a series of 70%, 80%, and 100% ethanol for 5 min at RT. Then 10 μl of probes were mixed, added to the preparations, and incubated at 37°C overnight. After washing slides in 1×SSC at 60°C for 5 min, 4×SSC/NP40 solution at 37°C for 10 min, and 1×PBS for 5 min at RT, preparations were counterstained with a DAPI-antifade solution (Life Technologies, Carlsbad, CA, USA) and kept in the dark for at least 2 hours before visualization with a fluorescent microscope.
Cytogenetic analyses
Cytogenetic analyses of meiotic and mitotic stages and of behaviors of FISH-labeled chromosomes were performed on at least 10 male individuals of each pure species and of each hybrid. To visualize and photograph meiotic chromosomes after FISH, we used an Olympus BX61 microscope (Olympus, Tokyo, Japan) with a connected camera Olympus U-CMAD3 (Olympus, Tokyo, Japan). Lengths of well-spread 10 metaphase I chromosomes of each species and of each hybrid were measured using the ruler tool in Adobe Photoshop CS6 (Adobe Inc., San Jose, CA, USA). Since the variances for both groups were equal, a statistical two-sample pooled t-test was done with the JMP 13 software (SAS Institute Inc., Cary, NC, USA).
Whole-mount FISH and analysis of chromosome pairing
Testes from one-day-old adults of pure species and hybrids were dissected in 1×PBS solution and fixed in 3.7% paraformaldehyde in 1×PBS with 0.1% tween-20 (PBST) for 10 min at room temperature. After incubation with 0.1-mg/ml RNase for 30 min at 37 ºC, testes were penetrated with 1% triton/0.1M HCl in PBST for 10 min at room temperature. After adding labeled DNA probes, testes were incubated at 75 ºC for 5 min (denaturation) and 37 ºC overnight (hybridization). Later, testes were washed with 2×SSC and mounted with a DAPI-antifade solution (Life Technologies, Carlsbad, CA, USA). Testes from 6 individuals of pure species and of hybrids were scanned and analyzed. Visualization and z-tack 3D scanning were performed on the whole testes with an interval of 1.25 μm between two optical sections under a 63 × oil lens of Zeiss LSM 880 confocal laser scanning microscope (Carl Zeiss AG, Oberkochen, Germany). Based on the size of the testis cells and on the limitations of microscope imaging, we chose from the first to sixteenth optical layers with strong and clearly detected fluorescent signals to analyze sex-chromosome-pairing events. Specifically, cell nuclei at the early stages of meiotic prophase I from 4th±1, 8th±1, and 12th±1 optical layers were used to count paring events. A total of 418 and 489 nuclei at the early stages of meiotic prophase I were analyzed for paring and unpairing of the sex chromosomes in pure species and hybrids, respectively (S2 Table). The percentage of the numbers of cells with pairing and no paring of sex chromosomes was used to compare parental species and hybrids. Since the variances for both groups were equal, a statistical two-sample pooled t-test was done using the JMP 13 software (SAS Institute Inc., Cary, NC, USA).
RNA extraction and RT-PCR
RNA was extracted from 40 abdomens (only distal segments that include testes) and from 20 carcasses of 1– to 2-day-old virgin adult males of An. coluzzii MOPTI, as well as of interspecies hybrids from crosses ♀An. coluzzii MOPTI × ♂An. merus and ♀An. merus × ♂An. coluzzii MOPTI using a Direct-ZolTM RNA MiniPrep Kit (Zymo Research, Irvine, California, US). cDNA for selected genes (Table 1) was generated in a 40-μl reaction that contained 8 μl of 5× first-strand buffer, 4 μl of 0.1-M dithiothreitol (DTT), 2 μl of 50-μM random hexamer primer (Thermo Fisher Scientific, Waltham, MA, USA), 1 μl of 25-mM dNTP mix solution (Thermo Fisher Scientific, Waltham, MA, USA), 1 μl of 40-U/μl RNAsin (Promega, Madison, WI, USA), 2 μl of 200-U/μl M-MLV Reverse Transcriptase (Thermo Fisher Scientific, Waltham, MA, USA), 1ng to 2μg of RNA and water. After incubation at 42°C for 1h, the reaction was stopped by incubation at 95°C for 5 min. Two-step RT-PCR was performed to analyze the gene expression. Each 25-μl PCR mix consisted of 1-μl cDNA, 12.5-μl ImmoMix™ 2× reaction-mix (Bioline USA Inc., Taunton, MA, USA), 1 μl of 10-μM forward and reverse primers, and water. PCR was performed in the Mastercycler® pro PCR thermocycler (Eppendorf, Hamburg, Germany) starting with a 95°C incubation for 10 min followed by 35 cycles of 95°C for 30 sec, 57°C for 30 sec, 72°C for 30 sec; 72°C for 10 min; and a final hold at 4°C. Amplification products were visualized in a 1-1.5% agarose gel.
Acknowledgments
We thank the Malaria Research and Reference Reagent Resource (MR4) at the BEI for providing us the laboratory colonies of malaria mosquitoes and Melissa Wade for proofreading the text.
References
- 1.↵
- 2.↵
- 3.
- 4.↵
- 5.
- 6.
- 7.
- 8.↵
- 9.
- 10.
- 11.
- 12.↵
- 13.↵
- 14.
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.
- 38.
- 39.
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.
- 55.↵
- 56.↵
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.↵
- 78.↵
- 79.↵
- 80.↵
- 81.↵
- 82.↵
- 83.↵
- 84.↵
- 85.↵
- 86.↵
- 87.↵