Germline-Restricted Chromosomes and Autosomal Variants Revealed by Pachytene Karyotyping of 17 Avian Species

Birds provide an interesting model to study chromosome evolution. They have undergone rapid speciation and evolved various adaptations to a wide variety of habitats. Yet, their karyotypes (chromosome sets) are very conservative. They are bimodal, i.e., composed of macro- and microchromosomes, the feature inherited from their reptilian ancestors [Romanov et al., 2014; Damas et al., 2018]. The diploid chromosome number (2n) of most karyotypes of bird species is about 78–82 [Degrandi et al., 2021]. The number of chromosome arms, so-called fundamental number (FN), also varies in narrow limits: 90–110. Unfortunately, less than 10% of bird species have been karyotyped. In passerines, the most speciose and diverse order of birds, the portion of karyotyped species is even smaller (7%) [Degrandi et al., 2020]. Paradoxically, the karyotype is unknown even in the species whose genomes have been sequenced, annotated, and studied in detail, such as the European pied flycatcher (Ficedula hypoleuca) [Kawakami et al., 2017].

Most avian karyotypes have been described in the 1960–1970s with the use of conventional methods of chromosome preparation and staining available at that time [Degrandi et al., 2021]. Development of the methods of chromosome analysis at the pachytene stage of meiosis provided a rather efficient karyotyping tool [Moses et al., 1977]. These methods are based on visualization of the synaptonemal complex (SC), the structure that mediates synapsis and recombination of homologous chromosomes. The SC is composed of 2 lateral elements, to which the chromatin loops are attached, and the central element that holds homologous chromosomes together. At the pachytene stage of meiotic prophase, the chromosomes are less compacted than at metaphase [Pigozzi, 2016]. Therefore, analysis of the SC provides higher resolution than the conventional analyses of metaphase chromosomes. This is especially important in the cytogenetic studies of the bimodal karyotypes because a morphology of the microchromosomes and even their number are rather difficult to assess in conventional chromosome spreads [Lisachov and Borodin, 2016; Kichigin et al., 2019]. The SC analysis is particularly efficient in the detection of heterozygosity for all types of chromosome rearrangements: inversions, translocations, deletions, and duplications. The heterozygotes for structural variants and heterogametic organisms produce characteristic heteromorphic SCs [Poorman et al., 1981; del Priore et al., 2015; Lisachov et al., 2017].

Analysis of pachytene chromosomes led to the discovery of the germline-restricted chromosome (GRC) [Pigozzi and Solari, 1998]. The GRC was present in germline cells and absent in somatic cells in all 22 species of passerine birds examined so far [Torgasheva et al., 2019; Poignet et al., 2021; Sotelo-Muñoz et al., 2022]. It has not been found in any non-passerine bird [Torgasheva et al., 2019]. In female germ cells, the GRC is usually present in 2 copies, which synapse and recombine with each other. At the meiotic prophase, the GRC bivalents are practically indistinguishable from normal autosomal bivalents. They can only be revealed by a comparative subtractive analysis of pachytene and somatic karyotypes. In male germ cells, the GRC is usually present in 1 copy, which forms a univalent at pachytene, easily distinguishable from the bivalents by its thin, coiled and often fragmented SC. The GRC univalents in male pachytene cells are always surrounded by dense chromatin clouds heavily labeled by anticentromere antibodies. Among the Passeriformes, the GRC varies in size and morphology [Pigozzi and Solari, 1998; Torgasheva et al., 2019, 2021a; Malinovskaya et al., 2020; Sotelo-Muñoz et al., 2022].

In this study, using immunolocalization of SYCP3, the main protein of the lateral elements of the SC, centromere proteins and histone H3, di- and trimethylated at lysine 9 (H3K9me2/3), marking heterochromatic regions, we examined male pachytene karyotypes of 16 passerine species with special attention to the GRC and 1 outgroup species, the common cuckoo (Cuculus canorus). We compared the pachytene karyotypes with the karyotypes described earlier and mostly obtained by conventional methods of chromosome preparation and staining.

Adult males were sampled at the beginning of the breeding season (April–May). The sources of the material and the number of specimens are shown in online supplementary Table 1 (for all online suppl. material, see www.karger.com/doi/10.1159/000524681).

Chromosome spreads were prepared by the drying-down method [Peters et al., 1997]. Immunostaining was performed according to Anderson et al. [1999] using rabbit polyclonal anti-SYCP3 (1:500; Abcam, cat. No. ab15093), human anticentromere (1:100; Antibodies Inc., cat. No. 15–234), and mouse monoclonal anti-H3K9me2/3 (1:100, Cell Signaling, cat. No. 5327) primary antibodies. The secondary antibodies used were Cy3-conjugated goat anti-rabbit (1:500; Jackson ImmunoResearch, cat. No. 111-165-144), AMCA-conjugated donkey anti-human (1:100; Jackson ImmunoResearch, cat. No. 709-155-149), and FITC-conjugated goat anti-mouse (1:100; Jackson ImmunoResearch, cat. No. 115-095-003). The slides were incubated overnight with primary antibodies and for 1 h with secondary antibodies at 37°C in a humid chamber. Slides were mounted in Vectashield antifade mounting medium (Vector Laboratories, USA, cat. No. H-1000-10).

The mitotic metaphase chromosomes were obtained from the cell culture of the gonads. The primary cell cultures were established from the minced gonads that were successively treated with 100 ng/mL collagenase I (Sigma, cat. No. SCR103) for 20 min and after that with 0.25% Trypsin-EDTA (Sigma, cat. No. T4174) for 20 min, both treatments at 37°C. The cell cultures were maintained in Dulbecco’s Modified Eagle’s medium (Gibco, cat. No. 41965039) supplemented with 10% fetal bovine serum (Gibco, cat. No. 10270106), 2% chicken serum (Gibco, cat. No. 16110082), GlutaMAX supplement (ThermoFisher, cat. No. 35050038), and penicillin/streptomycin (Sigma-Aldrich, cat. No. P0781). Preparation of metaphase chromosomes from the bird’s cell culture was performed at the 2nd to 3rd passage according to a modified protocol of chicken spermatogonial and follicle cell culture [Volkova et al., 2016; Zhao et al., 2018]. Slides were mounted in Vectashield antifade mounting medium with DAPI (Vector Laboratories, cat. No. H-1200).

The preparations were visualized with an Axioplan 2 imaging microscope (Carl Zeiss) equipped with a CCD camera (CV M300, JAI), CHROMA filter sets, and the ISIS4 image-processing package (MetaSystems GmbH). In each specimen at least 20 well-spread pachytene nuclei containing complete chromosome sets were photographed. The brightness and contrast of all images were enhanced using Corel PaintShop Photo Pro X6 (Corel Corp).

The centromeres were identified by anticentromere antibodies. The length of the SC of each chromosome arm was measured in micrometers, and the positions of centromeres were recorded using MicroMeasure 3.3 [Reeves, 2001]. All raw measurements are shown in the online supplementary Excel spreadsheet [https://meiosislab.com/projects/chromosomes/Malinovskaya_17_karyo_supplementary_data.xls]. We plotted idiograms after measuring at least 20 cells for each species. In each cell, we sorted SCs by their length in micrometers, identified those that can be distinguished unambiguously by the length and centromere index (most macrochromosomes in most species), and measured the average of these parameters across all cells. We grouped SCs that could not be unambiguously distinguished by their lengths and/or centromere indices (most microchromosomes in most species and ZZ bivalent in all species except the European pied flycatcher), ranged them within the group and measured the average parameters for each rank.

Statistica 6.0 software package (StatSoft) was used for descriptive statistics.

Figure 1 shows microphotographs of the SC spreads after immunolocalization of SYCP3 (red) and centromere proteins (pseudo-colored green) of all species examined. Schematic representations of the 7 largest SCs are shown in online supplementary Figure S1. Idiograms of pachytene karyotypes of the studied species are shown in online supplementary Figure S2. The haploid chromosome number (n) equal to the number of SCs in the pachytene cell, the haploid number of the chromosome arms, the total SC length, and a brief description of GRC are shown in Table 1.

The descriptions of the chromosome morphologies are based on the estimates of their centromeric indices. Those with a centromeric index between 0 and 0.1 are scored as acrocentrics, between 0.1 and 0.4 as submetacentrics, and between 0.4 and 0.5 as metacentrics. In all species, we identified the 7 largest SCs as macro-SCs, the others as micro-SCs.

We estimated the size of the GRC based on the size of GRC chromatin labeled by anticentromere antibodies and the length of the lateral element of the GRC SC if it was completely formed. We classified GRCs of sizes comparable to macrochromosomes of the basic set as macro-GRCs, the others as micro-GRCs. In most species, we found a variation in the SC appearance of macro- and micro-GRCs between the cells. It could form a complete, fragmented, or dot-like lateral element of SC or do not form it at all.

The somatic karyotype of the common cuckoo has been described by Bian and Li [1989]. They indicated 2n = 78. The pachytene karyotype of this species comprises 40 chromosome pairs (2n = 80). The total length of its SC set is one of the largest among the bird species examined so far. The excess of the SC length is mainly due to the very large metacentric SCs 1 and 2 and submetacentric SC3. Macro-SC4 is a metacentric. Macro-SCs 5 and 6 and micro-SCs 8 and 9 are submetacentrics, macro-SC7 and all other 30 micro-SCs are acrocentrics. The common cuckoo, as well as all other non-passerine birds studied so far, does not have a GRC in its pachytene karyotype.

The somatic karyotype of the rook has not been studied yet. The diploid number of the closely related species common raven (Corvus corax) is 80 [Roslik and Kryukov, 2001]. Our results indicate that the basic diploid chromosome set of the rook is 80. The pachytene karyotype of rook comprises 40 bivalents and 1 univalent of a GRC surrounded by a small cloud of anticentromere antibodies. All of the rook’s macro-SCs are submetacentrics. Among the micro-SCs, there are 10 submetacentrics and 23 acrocentrics. In most pachytene cells, the GRC forms a dot-like lateral SC element, while in some cells a GRC SC is not formed.

The somatic karyotype of the Blyth’s reed warbler remains unknown. The diploid number of the closely related species Paddyfield warbler (Acrocephalus agricola) is 78 [Bulatova, 1981]. The pachytene karyotype of the Blyth’s reed warbler comprises 40 chromosome pairs (2n = 80). Its macro-SCs 1, 2, 3, 6, and 7 are submetacentrics, SC4 is a metacentric, SC5 is an acrocentric. Most micro-SCs are acrocentrics. There are 1 metacentric and 6 submetacentric micro-SCs.

The GRC in the Blyth’s reed warbler occurs as a large acrocentric univalent with fragmented lateral SC element surrounded by a chromatin cloud labeled by anticentromere antibodies. The level of fragmentation varies between the cells from evenly labeled SC to a dispersed series of short fragments.

The first description of the Eurasian skylark somatic karyotype was provided by Udagawa [1952] on the basis of histological sections of embryonic testes and ovaries. The spermatogonial 2n was estimated as 78, oogonial as 77. Li and Bian [1988] estimated the diploid number of the Oriental skylark (Alauda gulgula) as 2n = 76±. We found that the pachytene karyotype of the Eurasian skylark contains 40 chromosome pairs (2n = 80). In terms of SC sizes, it is extremely asymmetric. Metacentric SCs 1 and 2 comprise 36% of the total SC length. This is in agreement with the description of the Alaudidae karyotype by Li and Bian [1988]. They described the Z and W chromosomes as the largest elements, metacentric and submetacentric respectively. The 1st macrochromosome was the second largest metacentric.

One of the 2 largest SCs in the pachytene karyotype of the Eurasian skylark is probably the neo-Z chromosome. Genomic analysis indicates that the Z chromosome of several skylark species resulted from fusions between parts of the chromosomes Z, 3, 4A, and 5. It is the largest sex chromosome found in birds (about 200 Mb) [Sigeman et al., 2019, 2020]. Another exceptionally large chromosome has probably also evolved via several chromosome fusions. Surprisingly, the chromosome number of the Eurasian skylark is the same as in most songbirds. This means that the fusions leading to the formation of the 2 largest macrochromosomes were preceded or followed by a series of chromosome fissions. Besides the 2 largest macrochromosomes, the karyotype of the Eurasian skylark contains 1 submetacentric macrochromosome. All other macrochromosomes and microchromosomes are acrocentrics.

The GRC of the Eurasian skylark occurs as a large chromatin cloud heavily labeled with anticentromere antibodies. It forms a thin, pale, and fragmented lateral SC element.

The genome of the European pied flycatcher has been sequenced, annotated, and studied in detail [Kawakami et al., 2017], yet its karyotype remains unknown. Its close relative, the red-breasted flycatcher (Ficedula parva), contains 40 chromosome pairs (2n = 80) [Bian et al., 1991]. The pachytene karyotype of the European pied flycatcher contains 40 bivalents (2n = 80). Almost all its macro-SCs are submetacentrics, SC4 is an acrocentric, SC5 is a metacentric. We are sure that SC4 is ZZ because analysis of pachytene oocytes of this species revealed a heteromorphic ZW bivalent with acrocentric Z axis [Torgasheva et al., 2021b]. Among micro-SCs, there are 5 metacentrics and 4 submetacentrics.

We observed extended centromeric regions in all meta- and submetacentric macro-SCs (but not in the acrocentric ZZ) and in several of the largest micro-SCs. They appeared as beads of several dots, which cover from 9% of SC1 to 41% of SC11/13 (online suppl. Table 2). The antibodies against histone H3, di- and trimethylated at lysine 9 (H3K9me2/3), marking transcriptionally inactive chromatin, produced strong signals at the standard centromeres of the pachytene chromosomes, but not at the extended ones except for SC11 (Fig. 2a, b). In the corresponding somatic metaphase chromosomes of this species, we observed extended primary constrictions (Fig. 2c).

The GRC in the European pied flycatcher usually appears as a chromatin cloud heavily labeled with anticentromere and H3K9me2/3 antibodies containing several fragments of lateral SC element. All cells of one specimen contained 1 GRC. Another specimen was mosaic: 93% of its pachytene cells (126 out of 135) contained 2 univalents of GRC, the remaining cells contained 1 GRC. Earlier polymorphism and mosaicism for GRC number has been detected in the pale martin and great tit [Malinovskaya et al., 2020; Torgasheva et al., 2021a].

The pachytene karyotype of the Bengalese finch comprises 40 chromosome pairs (2n = 80). This coincides with the description given by del Priore and Pigozzi [2014] and differs from the earlier description of its somatic karyotype as 2n = 78 [Takagi, 1972; Ray-Chaudhuri, 1976; Christidis, 1986a]. Its SCs 1 and 4 are metacentrics, SCs 2, 3, 5, and 6 are submetacentrics. All other SCs are acrocentrics.

The Bengalese finch GRC has been first described by del Priore and Pigozzi [2014] and then by Torgasheva et al. [2019]. It appears as the largest acrocentric univalent in the pachytene nucleus. We found a variation in the appearance of GRC SC between the cells: from complete lateral element to dot-like SC. Despite variation in the degree of polymerization of the lateral SC element, the cloud of anticentromere antibodies labeling the GRC was similarly large in all cells, which allowed us to classify the GRC of Bengalese finch as macro-GRC in accordance with the previous descriptions [del Priore and Pigozzi, 2014; Torgasheva et al., 2019].

Christidis [1986a] described 2n of the gouldian finch as 78. We found 40 chromosome pairs (2n = 80) in pachytene cells of this species. SCs 1 and 2 are metacentrics, SCs 3 to 6 are submetacentrics. Macro-SC7 and all micro-SCs are acrocentrics. SCs 1, 2, and 3 have extended centromeres, similar to those described in the European pied flycatcher. They occupy around 13–20% of SC length (online suppl. Table 2). In half of the cells analyzed, these regions are asynapsed whereas all other SCs are completely synapsed (Fig. 1). The delayed synapsis could be caused by an unusually extended centromeric heterochromatin.

We found heteromorphic SCs 4, 5, and 25. SC4 has 2 centromeres in metacentric and submetacentric positions. SC5 has 2 centromeres in different submetacentric positions. SC25 has centromeres in acrocentric and metacentric positions. The GRC appears as a moderately sized chromatin cloud comparable to the largest microchromosomes labeled with anticentromere antibodies surrounding 1–3 fragments of the lateral SC element.

In the pine bunting pachytene spermatocytes, we observed 40 bivalents (2n = 80), while Radzhabli et al. [1970] and Bulatova [1973] described the diploid number of the somatic karyotype of this species as 2n = 78. The pachytene karyotype of this species is rather similar to that of the Bengalese finch described above. There are 3 differences: SC4 is a submetacentric, SC7 is a submetacentric, and one of the micro-SCs is a metacentric. The GRC of the pine bunting is labeled by a small cloud of anticentromere antibodies. In most cells, the GRC does not form the lateral element of SC; in some cells we detected a dot-like signal of anti-SYCP3 antibodies.

The common chaffinch is an important model for evolutionary genetic studies. Its high-quality genome assembly has recently been published [Recuerda et al., 2021]. Our estimate of the common chaffinch karyotype coincides with the earlier published one (2n = 80) [Piccinni and Stella, 1070]. Its pachytene karyotype comprises 40 chromosome pairs. Total SC length and the lengths of the macro-SCs in the common chaffinch are much longer than in the closely related pine bunting (Table 1). However, the morphology of their chromosomes is rather similar except SC1, which is submetacentric in the common chaffinch.

The GRC of the common chaffinch forms a dot-like lateral element of SC labeled by a small cloud of anticentromere and H3K9me2/3 antibodies. To our surprise, in half of the pachytene spermatocytes (26 out of 49 examined), we observed a univalent of a small microchromosome. Unlike the GRC, it was not labeled either with anticentromere or with H3K9me2/3 antibodies (Fig. 3a). We suggest that this univalent originated by a premeiotic nondisjunction of one of the microchromosomes.

The somatic karyotype of the brambling has been described as 2n = 78 [Li and Bian, 1987]. We observed 40 pairs of chromosomes (2n = 80) which were morphologically similar to the pairs detected in the common chaffinch. There are a few differences: SC6 is a metacentric, and 2 of the micro-SCs are metacentrics and one is a submetacentric.

The specimen examined here has heteromorphic SCs 7 and 9. Each of them contains 2 centromeres in all cells examined. One homolog of chromosome 7 is a submetacentric as in most fringillids examined here. The other homolog of chromosome 7 is a metacentric, which had probably resulted from pericentric inversion or centromere shift, as well as the metacentric homolog of the acrocentric chromosome 9 typical for the fringillids.

Similar to the GRC of the pine bunting, the GRC of the brambling is labeled by a small cloud of anticentromere and H3K9me2/3 antibodies and does not form the lateral element of SC in most analyzed cells (Fig. 3b). In some cells, we detected dot-like signals of anti-SYCP3 antibodies.

Li and Bian [1987] described the somatic karyotype of the Eurasian bullfinch as 2n = 78. The pachytene cells of this species contain 41 chromosome pairs (2n = 82). All macro-SCs are submetacentrics. Among micro-SCs we observed 4 metacentrics and 5 submetacentrics.

The GRC of this species has been described by Torgasheva et al. [2019]. It forms a long acrocentric univalent evenly labeled by SYCP3 antibodies.

The diploid chromosome number of the European greenfinch is 76 according to Hammar and Herlin [1975] and 80 according to Christidis [1986b]. We detected 41 bivalents (2n = 82) in pachytene cells. The morphology of the 7 largest macro-SCs is very similar to that described for the Eurasian bullfinch except for the morphology of macro-SCs 3 and 4. They are metacentrics in the European greenfinch. The 3 smallest macro-SCs and 8 largest micro-SCs are submetacentrics. There are also 2 metacentric micro-SCs while all the others are acrocentrics.

The GRC of the European greenfinch appears as a partially formed lateral element of SC comparable in size to the largest microchromosomes and labeled by anticentromere antibodies.

According to Li and Bian [1987] the diploid chromosome number of the common redpoll is 78. According to our estimate, it is 82. All macro-SCs are submetacentrics. Among micro-SCs, 22 are submetacentrics and 12 are acrocentrics.

The pachytene karyotype of the common redpoll also contains a large acrocentric univalent of GRC evenly labeled by antibodies to SYCP3. In some cells, the lateral element of the GRC SC was fragmented.

The karyotype of the domestic canary has been first described by Ohno et al. [1964] as 2n = 80± and recently by da Silva dos Santos et al. [2017] and Kiazim et al. [2021] as 80. The domestic canary genome has been assembled and annotated but not yet to a chromosome level [Frankl-Vilches et al., 2015].

Pachytene cells of the domestic canary contain 41 chromosome pairs (2n = 82). Its macro-SCs are larger than those in the fringillids described above, although their morphology is similar. Macro-SC4 is a metacentric, all other macro-SCs are submetacentrics. Eight micro-SCs are submetacentrics. All other micro-SCs of the domestic canary are acrocentrics.

The domestic canary shows unusually extended centromeres of macro-SCs and the largest micro-SCs, similar to those detected in the European pied flycatcher and the gouldian finch. They are about 2–3 μm long covering about 10% of the SCs 1, 2/3, 4, 6, and about 15% of the SC5 (Fig. 2d; online suppl. Table 2). The antibodies against H3K9me2/3 label both the standard and extended centromeres of the pachytene chromosomes (Fig. 2e). We also detected substantially extended primary constrictions at the corresponding metaphase chromosomes (Fig. 2f).

The GRC of this species has been described by Torgasheva et al. [2019]. It appears as a small cloud of anticentromere antibodies with the dot-like lateral element of SC.

Christidis [1986b] described the somatic karyotype of the European goldfinch as 2n = 82. Our data confirm this diploid chromosome number. We observed 41 bivalents in the pachytene cells of the European goldfinch. Its 7 largest macro-SCs are smaller than those of the domestic canary but similar in morphology. We observed 1 metacentric and 6 submetacentric micro-SCs, all other micro-SCs are acrocentrics.

The GRC of the European goldfinch has been described by Torgasheva et al. [2019] as micro-GRC. In most cells, the GRC was present as 1 or 2 dot-like lateral SC fragments labeled by a small cloud of anticentromere antibodies. In some cells, the GRC formed a complete lateral SC element.

Bulatova [1973, 1981] found 82 chromosomes in bone marrow cells of the common linnet. This is in good agreement with our observation. We found 41 bivalents. Pachytene karyotypes of the common linnet and the European goldfinch are almost the same. The difference concerns the morphology of the micro-SСs. The common linnet has only 2 homomorphic metacentric and 2 heteromorphic metacentric/acrocentric micro-SCs. Micro-SC9 is a submetacentric. All other micro-SCs are acrocentrics.

The heteromorphic SCs 12 and 15 contain 2 centromeres in all cells examined: one in acrocentric and one in metacentric position. The metacentric homologs have probably resulted from pericentric inversions or centromere shifts.

The GRC of the common linnet occurs as an acrocentric univalent with fragmented lateral element of SC labeled by a large cloud of anticentromere and H3K9me2/3 antibodies (Fig. 3c).

The diploid chromosome number in the Eurasian siskin was estimated as 78 [Takagi, 1972; Bulatova, 1973; Li and Bian, 1987; Christidis, 1990]. We found 41 chromosome pairs in the pachytene cells of this species (2n = 82). The morphology of the macro-SCs is rather similar to that in 2 fringillid species described above. However, the karyotype of the Eurasian siskin contains only 2 submetacentric micro-SCs, all other micro-SCs are acrocentrics.

Torgasheva et al. [2019] classified its GRC as macro-GRC. In our spreads, it occurs as a large cloud of anticentromere antibodies with fragmented lateral SC element.

The main results of this study are: (1) the first description of the karyotypes of 3 species; (2) the correction of the published data on the karyotypes of 10 songbird species; (3) the first detection of the extended centromeres in 3 model species; (4) the first detection of chromosomal heteromorphism in 3 species; and (5) the detailed characterization of the GRCs of all songbird species examined.

Why are we sure that our descriptions of the karyotypes are more precise and reliable than the published data based on conventionally prepared and stained somatic metaphase spreads?

There are at least 4 reasons for this. First, the pachytene chromosomes of birds are about 2–3 times longer than somatic metaphase chromosomes (see Fig. 2). Second, each object, which we identified by microphotography as the bivalent of the basic chromosome set, was simultaneously labeled with 2 different antibodies: against SYCP3, the main protein of the lateral elements of the SC, and the centromere proteins. Univalents were distinguished from the bivalents by less intense SYCP3 labeling. Therefore, a misidentification of a microchromosome as cell debris and cell debris as a microchromosome was unlikely. Third, these 3 criteria were applied to at least 20 well-spread pachytene cells containing complete chromosome sets. This made our estimates statistically sound. Forth, the bivalents were not just counted, they were measured. This made possible an objective estimate of the morphology of the macrochromosomes and, what is more important and almost impossible at the somatic metaphases, the morphology of the microchromosomes.

A comparison between our and previously published chromosome numbers of the examined species shows that the earlier researchers often undercounted 1 or 2 chromosome pairs. These errors were due to the very small sizes of the smallest microchromosomes and the relatively low specificity of the chromosome dyes. For these reasons, it was almost impossible to estimate the morphology of the small microchromosomes at the mitotic metaphase spreads. It was believed that all of them were acrocentrics [Christidis, 1990]. Our data show that this is true for the common cuckoo, Eurasian skylark, and Bengalese finch. Other songbirds have at least 1 meta- or submetacentric microchromosome, and most of the species have many of them.

Using anticentromere antibodies, we revealed unusual extended centromeres in almost all macrochromosomes and the largest microchromosomes of 3 model species (the European pied flycatcher, gouldian finch, and domestic canary). They are visible at both the pachytene and mitotic metaphase spreads of these species. They are also visible in the published images of the domestic canary metaphase chromosomes and shown in the idiograms [da Silva dos Santos et al., 2017; Kiazim et al., 2021], but did not attract much attention.

Such long centromeres are rare. They have been detected in a few species of legumes [Neumann et al., 2015, 2016], fire ants [Huang et al., 2016], Indian muntjac [Brinkley et al., 1984], and marsupial hybrids [Metcalfe et al., 2007]. However, they have not yet been described in birds. Robertsonian translocations and centromere drive have been suggested among the possible causes of the centromere extension [Brinkley et al., 1984; Huang et al., 2016]. The results of H3K9me2/3 immunolocalization indicate a variation in the epigenetic status of the extended centromeres. In the domestic canary, both extended and standard centromeres are H3K9me2/3-positive indicating their heterochromatic state. In the European pied flycatcher, the standard centromeres are H3K9me2/3-positive, while the extended centromeres are H3K9me2/3-negative except SC11, which was H3K9me2/3-positive. The genetic composition of the extended centromeres deserves special attention.

One more advantage of the SC analysis is its high efficiency in the detection of structural heterozygosity. In this study, we detected heteromorphic SCs in the gouldian finch, brambling, and common linnet. In all cases, the bivalents displayed 2 misaligned centromeres. This may indicate heterozygosity either for pericentric inversions or for centromere repositions (centromere shift) [Schubert and Lysak, 2011]. Both types of chromosome rearrangements are implicated in karyotypic macroevolution of birds [Zhou et al., 2014; Kiazim et al., 2021]. The inversions play an especially important role in the restriction of the gene flow between sympatric and parapatric species [Hooper and Price, 2017]. However, the intraspecific polymorphism for inversion is poorly studied in birds. Our findings indicate the targets for future studies.

Our study revealed a wide variation in GRC size and appearance. Torgasheva et al. [2019] suggested to classify them as macro- and micro-GRCs to fit the criteria for macro- and microchromosomes of the basic set. Indeed, all analyzed chromosomes fell into one of these categories. However, some micro-GRCs were much smaller than the smallest microchromosomes of the basic set that can be inferred from the size of the chromatin cloud labeled by anticentromere and H3K9me2/3 antibodies. This variation and the lack of phylogenetic clustering by size confirm the highly dynamic nature of GRCs [Kinsella et al., 2019; Sotelo-Muñoz et al., 2022]. It was shown that GRCs of different species contain different multiply repeated sequences, which probably can be accumulated and/or be lost rather quickly [Torgasheva et al., 2019].

The SC of GRCs in pachytene spermatocytes of most species examined here appeared fragmented, however, the degree of fragmentation varied between cells. It is unclear if this feature is related to the different properties of GRCs in different species (its genetic content, the degree of heterochromatinization) or the intercellular and interindividual variation in the effectiveness of cohesin loading and SC polymerization since for most birds studied to date, only one sample was analyzed. The interindividual variation in GRC appearance was observed in pachytene spermatocytes of the great tit [Torgasheva et al., 2021a] and several species of the genus Lonchura [Sotelo-Muñoz et al., 2022]. The intense chromatin labeling with antibodies to centromere proteins and H3K9-modified histone allowed us to distinguish between GRC and accidental autosomal univalents.

The results of our study indicate several lines of future research. To minimize sacrificing birds, we described the karyotypes of most species by single specimens examined. The sample size might be increased at least for the most interesting species by targeted examination of the somatic karyotypes, which can be obtained from short-term fibroblast cultures derived from blood or feather pulp.

It seems important to estimate the frequency, geographic distribution, and probable adaptive significance of the inversion polymorphism detected in the gouldian finch, brambling, and common linnet, the species with wide breeding and residence areas.

Another interesting species is the Eurasian skylark with 2 giant chromosomes. The origin of its Z/W chromosomes has been resolved [Sigeman et al., 2019, 2020]. The genetic content of another giant chromosome of the Eurasian skylark remains unknown. FISH with universal BAC probes [Damas et al., 2017] might shed light on its origin. The microchromosomes of this species also deserve close attention because despite the fusions of several macrochromosomes in the neo-Z and in another giant chromosome its chromosome number remains the same as in most songbirds.

The nature, evolution, and adaptive significance of the extended centromeres of gouldian finch, European pied flycatcher, and domestic canary are of special interest. Recent advances in sequencing and bioinformatic analysis of the repetitive DNA of birds [Weissensteiner and Suh, 2019] make it possible to address these questions

The birds were handled and euthanized in accordance with the approved national guidelines for the care and use of laboratory animals. The euthanization was performed by isoflurane overdose. The study is reported in accordance with ARRIVE guidelines (https://arriveguidelines.org). This study protocol was reviewed and approved by the Animal Care and Use Committee of the Institute of Cytology and Genetics SB RAS (protocol # 114).

The authors have no conflicts of interest to declare.

Data collection and analysis was funded by Russian Science Foundation, grant number 20-64-46021. Microscopy was carried out at the Core Facility for Microscopy of Biologic Objects of Siberian Branch of Russian Academy of Sciences, Novosibirsk, Russia (regulation no. 3054) funded by the Ministry of Science and Higher Education of the Russian Federation, grant numbers 0259-2021-0011 and #2019-0546 (FSUS-2020-0040). The funding bodies played no role in the design of the study and collection, analysis, and interpretation of data and in writing the manuscript.

Conceptualization: P.M. Borodin; investigation: L.P. Malinovskaya, A.Y. Slobodchikova, E.O. Grishko, I.E. Pristyazhnyuk; writing and original draft preparation: P.M. Borodin; writing, review and editing: L.P. Malinovskaya, P.M. Borodin, A.A. Torgasheva; visualization: L.P. Malinovskaya, P.M. Borodin, A.A. Torgasheva; supervision: A.A. Torgasheva; project administration: A.A. Torgasheva.

All data generated or analyzed during this study are included in this article and its online supplementary material. A preprint version of this article is available on bioRxiv [Slobodchikova et al., 2022].

Lyubov P. Malinovskaya and Anastasia Y. Slobodchikova share first authorship.