Nautilus macromphalus specimens were collected on the outer shelf of New Caledonia (22°33′S; 166°25′E). Small pieces of gill tissue from frozen (−20 °C) specimens were fixed in 70% ethyl alcohol. DNA was extracted as described in Bonnaud et al. [6].

The complete 18S rDNA gene was amplified in polymerase chain reaction using the following primers: ATGCTTGTCTCAAAGATTAAGCC (between region 4 and 5 – Fig. 1) and AATTACCGCGGCTGCTGGCA (helix 20), TACTTGGATAACTGTGGTAA (helix 9) and AAATCGCTCCACCAACTAAGAA (helix 40), GGGCGGACACCGGTAGGATTGAC (helix 37) and AATGATCCTTCCGCAGGTTCAC (helix 50). PCR products were purified and cloned. Several clones of each 18S rDNA portion were sequenced in both strands (EMBL accession number AJ301606). The complete 18S rDNA gene sequence of N. macromphalus (submitted to EMBL under the accession number AJ301606), N. pompilius and N. scrobiculatus (accession numbers AF207641 and AF120504, respectively) were compared taking into account the secondary structure features.

For karyotype determination, three mature males were used. One was dissected on the day of capture (A), another was injected with colchicine 0.5% (0.3 ml per 100 g of fresh animal) and kept in aquarium for 24 h prior to dissection (B), the third was brought back alive to Paris and dissected upon arrival (C). Gills, mantle and testis were sampled and directly dilacerated in the hypotonic media (i.e. 0.55% KCl and distilled water) where they were kept for 30 to 70 min prior to fixation with three baths (10, 20, 30 min) of a freshly prepared solution of iced methanol-acetic acid (3:1). The pellet from the last centrifugation was resuspended in 2 ml fixative, then deposited on slides and stained with Giemsa 4%.

3.1 Nuclear 18S rDNA gene analysis

Hairpin 11, hairpin 17 and helix E10-1 are identical in the three nautilus species, but helix E10-1 is longer in nautiloids (169 nucleotides) than in other molluscs (46 to 69 nucleotides) (Figs. 1 and 2). The 3′ end of the molecule is highly variable (essentially in helices 23 to E23-6 in coleoids). Among cephalopods, only Nautilus has a huge insertion of around 200 nucleotides in this region (almost identical in the three species). From helix 25 to the molecule end, N. pompilius and N. macromphalus sequences are very similar, whereas large insertions are present in N. scrobiculatus, especially in loops 43 and 49 (Fig. 2).

3.2 Karyotype of Nautilus

Coupled with gene-sequence analyses, insertions and/or the composition of hairpins and loops are beginning to be considered of interest in phylogenetic reconstruction, systematic grouping and for explanation of molecular mechanisms in an evolutionary perspective [11,20,21]. Partial sequences of mitochondrial and nuclear genes for Nautilus species are now available [4,22]; however, total sequences are rare.

The present results show that 18S rDNA is longer in Nautilus than in any known mollusc, including other cephalopods, where the size of 18S rDNA is around 1850 nucleotides [9].

Being a highly conserved molecule, it was expected that 18S rDNA sequences would be very similar amongst Nautilus species. This is the case for Nautilus macromphalus and N. pompilius, but the N. scrobiculatus 18S rDNA sequence shows significant insertions mainly in the last region of the molecule (Fig. 2). This region could be an interesting tool to estimate genetic variation between Nautilus species, like helix 17 within Rhabditidae (Nematoda) [20] or helix 23 within Rotifera-Acanthocephala [11].

Although five to six morphotypes are generally recognised based on shell morphology (size, colour pattern, general shape and form of the umbilical region), only three species are universally accepted: (1) Nautilus pompilius Linnaeus, 1758, the type species of the genus, distributed over almost the entire range of the genus and possessing a closed umbilical region; (2) N. macromphalus Sowerby, 1849, endemic to New Caledonia, with an intermediate umbilicus opening; (3) N. scrobiculatus [Lighfoot 1786], endemic to Papua-New Guinea, sympatric with N. pompilius, with a large and open umbilical area.

In nautiloids, phenotypic variations do however not necessarily relate to genetic variability. N. pompilius and N. macromphalus are two forms easily recognised by their shell characteristics but neither mitochondrial genes [4] nor 18S (present study) nucleotide differences correspond to the morphological species variation, suggesting a close phylogenetic affinity between these morphologically distinct species. Moreover, the chromosome number is the same, indicating that there has not been a drastic transformation of the genome. Therefore, N. macromphalus is probably a geographic variant of N. pompilius, a hypothesis supported by 16S mitochondrial gene analysis [4].

The N. scrobiculatus 18S rDNA sequence shows specific large insertions in some loops leading to a unique form among nautiloids. Moreover, there is intraspecific variation in helix E10-1 (unlike N. pompilius, data not shown): the database complete sequence is identical to that of the other Nautiloids (169 nt long), whereas the partial 18S rDNA sequence previously obtained from another specimen [22] is only 116 nucleotides long (gaps of 48 and 5 nucleotides – Fig. 2). An intraspecific variation was also observed between 16S rDNA sequences obtained from different specimens of N. scrobiculatus [4]. Polymorphism could be at the origin of speciation events. Ward and Saunders [5] proposed to elevate N. scrobiculatus to a generic rank, coining the genus Allonautilus. Although this raised some controversy [23], the present differences in 18S rDNA sequences observed between N. scrobiculatus and the other representatives of the genus, its apparent intraspecific variability, its morphological characters and the phylogenetic analysis of the different species based on 16S [4], all warrant that N. scrobiculatus is to be placed in the distinct genus, Allonautilus. The karyotype of N. scrobiculatus is still to be established.

A. scrobiculatus is often considered as representing the oldest lineage in Nautilus evolution. Therefore, its larger 18S rDNA gene with many insertions could be the ancestral form of 18S, suggesting that the tendency in nautiloid evolution might be a reduction of the molecule size. Similarly, the ancestral form is the large rDNA gene of nautiloids (compared to other cephalopods), with many insertions. The loss of partial sequences in loops during evolution has lead to smaller rDNA molecules in the coleoid lineage.

Gene sequences, chromosome number and morphology can all be used as phylogenetic indicators [23], conservatism of chromosome number being reported for many mollusc groups. The chromosome numbers of cephalopods (52–112) are the highest among Mollusca (28–32) [16] and their progression appears non-random, following the successive appearance of the lineages (Fig. 3). As Nautilus displays the lowest number of chromosomes among cephalopods, it can be speculated that a small number of chromosomes, shared by Nautilus and octopods, is the ancestral state for cephalopods. However, within octopods, when the chromosome number is almost the same, the forms are clearly different. Eledone cirrhosa and Octopus vulgaris have a higher proportion of ‘small’ chromosomes than O. ocellatus. The number of chromosomes and the proportion of ‘large’ chromosomes are higher in Decabrachia than in Octopoda. Thus, both number and morphology appear to change radically during evolution. If Nautilus chromosome characteristics, i.e. small number and small size, are considered to be close to the ancestral state, then the decabrachian karyotype probably results from chromosome duplication events having potentially occurred at the base of sepiids and teuthoids.

Polyploidy leads to gene duplications. In many ‘recent’ cephalopods (mainly Decabrachia), the existence of different forms of 18S rDNA genes, identified as ‘pseudogenes’ has been demonstrated [9]. Preliminary in situ hybridisation results indicate that Nautilus is apparently the only cephalopod with only one form of 18S rRNA (Bonnaud, unpublished data). Accordingly, ‘pseudogenes’ of 18S rDNA in Decabrachia might be a consequence of a genomic duplication, their chromosome number being higher than in Nautilus.

The presence of insertions in 18S rDNA and the small number of chromosomes could be an ancestral condition in Nautilus and suggest that the derived state is an increased chromosome number and a shorter 18S rDNA gene. Parallel to additional molecular studies, karyological data on a wider range of cephalopod species could be of considerable help to further understand the relationships of mollusc groups in an evolutionary perspective.