A synthetic description of Drosophila melanogaster spermatogenesis comes from light microscopy studies 123, as well as electron microscopy studies 456789. At the tip of Drosophila testis a group of 8–9 staminal cells mitotically divide forming another staminal cell and a primary spermatogonium. Primary spermatogonia divide four times producing, after the last division, sixteen primary spermatocytes. Primary spermatocytes undergo a relatively long (~90 hours) maturation phase, during which they slowly increase their nuclear volume so that by the end of the growth phase they are 25–30 times larger than spermatogonia. Morphologically, at the beginning of their development primary spermatocytes are very similar to spermatogonia; however, the completely developed nuclei are characterized, in addition to their size, by the presence of three filamentous structures called lampbrush-like loops. Primary spermatocytes at late meiotic prophase I show an extensive fragmentation of these loops 1011. Meiosis produces 64 haploid spermatids which are easily recognizable by the association of a phase lucent nucleus with a phase dense mitochondrial derivative (nebenkern) of the same size (approximately 7 μm). In the last phase of spermatogenesis, spermatid nuclei reduce in volume by 200 times as a byproduct of DNA condensation and lose cytoplasmic organelles, while the nebenkern elongates and divides into two symmetric halves between which the sperm axoneme is formed. After sperm tail formation, spermatozoa transfer into seminal vesicles and they are ready to be inseminated into the female. Mature spermatozoa in drosophilids are characterized by very long sperm tails: their size is approximately 1.8 mm in D. melanogaster, 23 mm in D. hydei and almost 60 mm in D. bifurca 5121314.

The Y chromosome of D. melanogaster is a submetacentric, completely heterochromatic element representing 12% of the male genome 15. The main genetic function of the Y chromosome is male fertility: X/0 flies are phenotypically normal males, but they are completely sterile 16. It has been demonstrated that Y-associated fertility factors play a role only in the male germ line 17 and more specifically inside primary spermatocytes 5. Three fertility factors (namely kl-5, kl-3 and ks-1) have huge physical dimensions 18, showing a DNA content of ~4,000 Kb each, this being 100 times longer than an average eukaryotic gene. These uncommon sizes might be partly explained by the fact that they form, inside primary spermatocyte nuclei, three giant lampbrush-like loops 10, and indeed one of the most striking features of Y-loops is their DNA content. Most of the Y chromosome DNA is represented by transposable elements 19 and simple sequence satellite DNA 202122. Furthermore, some of these satellites are abundantly transcribed in the ks-1 and kl-5 loops, but the corresponding transcripts do not migrate into the cytoplasm, and disintegrate together with the loops during meiotic prophase I 23. The unusual behavior of these transcripts has lead to a debate regarding the functional role of Y-loops in drosophilids. It has been proposed that Y-loops are merely the cytological manifestation of unusually long genes (see 24 for review). Goldstein and coworkers 25 found that mutants lacking kl-3 and kl-5 loci in D. melanogaster do not express some high molecular weight polypeptides and lack the external dynein arms in the sperm axonemes, suggesting that these regions indeed harbor the structural genes for these proteins. In support of this hypothesis Gepner and Hays 26 found an open reading frame (ORF) coding for a dynein heavy chain inside the kl-5 locus (Dhc-Yh3); this gene is conserved also in Drosophila hydei in both location (Y chromosome) and cytological phenotype (a loop called Thread) 27. Further support for this hypothesis comes from Carvalho and coworkers 28 who found another dynein heavy chain gene inside the kl-3 fertility factor and an occludin-related gene (ORY) inside the ks-1 locus 29. However, although the Y loops appear to represent single giant transcription units, the presence of such extremely large introns full of repetitive DNA suggests the non-protein coding regions of the transcription units might serve other purposes. Loops can be selectively stained by Giemsa at pH10, as well as by dansyl-chloride and Coomassie blue 10; furthermore, several antibodies have been raised which specifically recognize these structures 103031323334353637, showing that they accumulate high quantities of proteins. An interesting example is provided by T53-1, a monospecific, polyclonal antibody that recognizes Loopin-1, a testis-specific autosomal protein of Drosophila melanogaster. Loopin-1 is present onto both the Y-chromosome kl-3 loop and the sperm tails 30; notably, T53-1 also stains the Pseudonucleolus and, partially, the Cones Y-loops of D. hydei. The accumulation on the loops of proteins encoded by autosomal genes suggests that the Y-loops, besides any coding function, might also play a structural role, possibly as a nuclear framework useful for docking and/or modifying proteins to be used later in spermatogenesis 103038.

In the present paper I studied the male germ lines of 13 Drosophila species and used the T53-F1 antibody (a second antiserum prepared against Loopin-1 which has the same specificity of the original T53-1 antiserum), to assay the evolutionary conservation of lampbrush-like loops. In all the 13 species of drosophilids analyzed, intranuclear structures were detected in primary spermatocytes using phase-contrast microscopy. Such structures resemble the Y-chromosome loops of D. melanogaster 5 and D. hydei 40, and in each species one loop is intensely stained by the antibody. The antibody shows no specific staining in spermatids but strongly decorates sperm tails. These data strongly suggest that lampbrush-like loops are a conserved, and consequently a necessary feature of drosophilid meiosis. Moreover, the conservation of spatial and temporal immunostaining in all species suggests a conserved, although unknown, function for Loopin-1. Comparison of these data with other spermatogenesis parameters, such as the size of primary spermatocyte nuclei or the sperm length, did not reveal any evident relationship, suggesting that these aspects of male germ line are not directly correlated.

In the present study an extensive analysis of the male germ line of several Drosophila species was performed, in order to find conserved features. The selected species are scattered along their genealogical tree, so that it was possible to analyze even drosophilids which diverged approximately 58 million years ago (Figure 1a). Cytological analysis was performed using both phase-contrast microscopy and immunofluorescence, and the same fixing procedure was used, so it is possible to argue that the structures stained are probably evolutionary conserved. The results reported here indicate that in all species tested it is possible to recognize cells which are probably mature primary spermatocytes, as both their size and the presence of intranuclear structures are similar to both D. melanogaster and D. hydei (Figure 2). Moreover, with regard to their size and shape, the intranuclear structures are species-specific, and at least one of them is always specifically immunostained by T53-F1, a monospecific polyclonal antibody raised against the Loopin-1 protein of D. melanogaster 39. In a pattern similar to that seen in D. melanogaster, the staining by T53-F1 is not present in the spermatids of any species, but the antibody intensely stains sperm tails from when they start developing (data not shown). This indicates that a Loopin-1-like protein is loaded onto lampbrush-like loops before meiosis and then accumulates along sperm tails in all the drosophilids analyzed.

Several cytological parameters were considered in order to find possible correlations among them, to elucidate a possible role of Loopin-1, since no mutation is known to date in the loopin-1 gene. However, it was not possible to find any relationship. I found that sperm length, amount of sperm tail inside the egg upon fertilization, loop staining size and loop shape are independent parameters (Table 1 and Figure 2).

Whereas Loopin-1 immunostaining was found to be conserved with regard to its presence in male germ lines, the shape and staining size of the corresponding lampbrush-like loops were found to be species-specific. In fact, although all species tested have loops stained by T53-F1 and similarly sized primary spermatocytes nuclei, all possible combinations can be found. The loop appears very small in D. americana, D. bifurca and D. funebris (very distant species) and it appears very prominent in D. acanthoptera, D. littoralis, D. melanogaster and D. mercatorum (similarly distant species) (Figure 1). However, it seems that the shape of primary spermatocyte nuclei may be conserved. The only three species having a non spherical nucleus are D. bifurca, D. hydei and D. mercatorum, which are evolutionary close to each other (Figure 1a). The first two show a "pear-like" nucleus, while the last species has an "dumbbell-shaped" nucleus (Figure 2 and Table 1). Notably, all three species have loops of a dense nature, suggesting a similar molecular organization. In this respect it is interesting to note that a filamentous loop recognized by T53-F1 is present only in some species of the melanogaster subgroup (Figure 1B), and in the closely related species D. novamexicana, D. virilis and D. littoralis (Table 1). The conservation of nuclei shape, staining size and appearance is even more interesting if we compare it with previous reports 4243. The sperm tail length and sperm head size are very variable not only among species, but also among different populations of the same species and even inside the same individual 44. This suggests that mechanisms controlling premeiotic stages are more stringent than those controlling sperm development.

Thus far the biological role of Loopin-1 in Drosophila melanogaster male germ line is unknown. The loopin-1 gene maps to polytene region 53C 39 and its product was originally described as a tektin-like protein 30. Subsequent studies, however, demonstrated that the loopin-1 gene corresponds to the CG4750 locus of D. melanogaster and is homologous to a leucine aminopeptidase 3945; therefore, it is unrelated to the tektin family of proteins. The protein encoded by this gene is conserved at least in D. simulans (55% identities, 72% positives) which diverged from D. melanogaster 3My ago, and in D. pseudoobscura (78% identities, 88% positives) which diverged from D. melanogaster 40My ago (Figure 1). This is not surprising, since it has been recently demonstrated that leucine aminopeptidases represent an important fraction of the Drosophila sperm proteome 49, although their role in sperm tail is still unclear. It is also known that Loopin-1 localization in D. melanogaster sperm tails is not dependent upon loop binding 30. Indeed, its pattern of localization is quite peculiar: the protein recognized by the T53-F1 antibody appears inside primary spermatocyte nuclei during their maturation, binding to the kl-3 loop throughout its development 30. At the beginning of meiosis, as soon as Y-chromosome lampbrush-like loops disgregate, staining fades and no other sign of the protein is present during meiosis or spermatid maturation. The staining once again becomes intense in sperm tails at the time they begin to elongate. Notably, this peculiar behavior was found to be similar in all the 13 species examined, despite their considerable phylogenetic distance (Figure 1). The fact that T53-F1 reacts specifically in a conserved pattern strongly suggests that D. melanogaster Loopin-1 is a conserved protein and that the formation of loop-like structures during primary spermatocyte development is a necessary step during spermatogenesis in all species examined.

The data reported here show that a Loopin-1-like protein is detectable in both primary spermatocyte nuclei and sperm tails in all species tested, and possibly in all drosophilids. These data strongly suggest that this protein might have an important functional role during spermatogenesis through its binding to intranuclear structures during the development of primary spermatocyte nuclei. Notably, the structures decorated by the T53-F1 antibody during primary spermatocyte nuclei development clearly resemble the well known Y-chromosome lampbrush-like loops of D. melanogaster and D. hydei. It is likely that lampbrush-like loops might be a sign of Y chromosome activity during spermatogenesis in all drosophilids tested. Therefore it would not be surprising to find that the Y chromosome plays a similar role in the development of lampbrush-like loops in all drosophilids.

Evolutionary data demonstrate that drosophilids diverged at least 60 million years ago. Data shown here support the hypothesis that a Loopin-1-like protein exists in all species tested, indicating that it is evolutionary conserved. Moreover, data presented here also demonstrate that primary spermatocyte nuclei of all species tested show the presence of discrete intranuclear structures, indicating that this cytological feature is also conserved. The lack of a loopin-1 mutation however has impeded the assessment of its role during spermatogenesis, and the function of Loopin-1 loading onto lampbrush-like loops in the male germ line. Interestingly, mutations in other genes that, like loopin-1, are located on the autosomes have recently been shown to affect the formation of the kl-3 loop and to alter the external dynein arms of the sperm axonemes 47 which are believed to be the site of action of the dynein encoded by the kl-3 fertility factor. The spermatozoa of these mutants either degenerate before maturation or have major structural defects leading to their complete immobility. These data suggest that these autosomal proteins might be involved in building the structure of the lampbrush-like loops – either to set up structures necessary for transcription of the kl-3 gene or to perform other functions associated with the loop.

RP designed the study, carried out all experiments and prepared the manuscript.