- Split View
-
Views
-
Cite
Cite
Jay T. Dalet, Cynthia P. Saloma, Baldomero M. Olivera, Francisco M. Heralde, Karyological analysis and FISH physical mapping of 18S rDNA genes, (GATA)n centromeric and (TTAGGG)n telomeric sequences in Conus magus Linnaeus, 1758, Journal of Molluscan Studies, Volume 81, Issue 2, May 2015, Pages 274–289, https://doi.org/10.1093/mollus/eyu090
- Share Icon Share
Abstract
Karyological analysis of gill tissue from the marine gastropod Conus magus showed a diploid chromosome number of 32. Three major groups of chromosomes were identified: 22 median region, 4 submedian region and 6 subterminal region chromosomes. The haploid count was verified using chromosomal spreads from ovarian cells. Fluorescence in situ hybridization (FISH) physical mapping of an 18S rDNA sequence showed a wide distribution of major, medium and minor hybridization sites. These hybridization sites were detected in two to four different regions—paracentromeric, centromeric, interstitial or telomeric—per chromosome. Identical 18S rDNA FISH signals were found in the putative pairs of homologous chromosomes. FISH profiles of tandem simple-sequence repeats (SSRs) were physically mapped in locations near centromeres (GATA)n or telomeres (TTAGGG)n, as well as in noncentromeric (GATA)n regions and nontelomeric (TTAGGG)n interstitial regions of whole chromosome arms. Similar SSR chromosome organization FISH patterns were observed in two chromosomal spreads from the same individual: (1) telomeric (TTAGGG)n sequences in both p and q terminals of 15 chromosomes (10 median region, 2 submedian region and 3 subterminal region) and (2) centromeric (GATA)n sequences in 23 chromosomes (13 median region, 4 submedian region and 6 subterminal region). The C. magus DAPI karyotype and genomic landmarks based on the FISH profiles of 18S rDNA, (TTAGGG)n and (GATA)n sequences may contribute to the elucidation of the evolution of the karyotypes of Conus species and in the detection and localization of Conus chromosomal genes using FISH.
INTRODUCTION
Little is known about the karyological profile of Conus species. Chromosomal classification and characterization in Conus are prerequisites for chromosomal depiction of genomic structure and allelic mapping of toxin genes in Conus species. Conus species are venomous and predatory marine snails that are of recent interest for their potential as rich sources of diverse pharmacologically-active peptides. Omega conotoxin-MVIIA (otherwise known as ziconotide or Prialt) is an example of one of these conopeptides that has been made available commercially for the treatment of chronic pain. It was isolated from C. magus and discovered to have unique pharmacological activity against N-type calcium channels (Cav2.2) (Olivera, 2006). Some Conus pepdtide families were found to have a high frequency of post-translational modifications of prepropeptide gene products. This biochemical mechanism entails a genomic structure different from what is basically encoded by cDNA clones and plays a significant role in the physiological properties of some conotoxins (Buczek, Bulaj & Olivera, 2005; Olivera, 2006). The Conus venom is a suite of different conotoxins. The organization of the conopeptide genes in the genome, their location, arrangement and interaction are as yet unknown. However, knowledge of these may assist in elucidating the mechanisms governing the molecular expression of these conopeptides. So far, conopeptide gene locations along the chromosomes have not been determined. Developing a method for the identification and classification of chromosomes in Conus by characterizing molecular landmarks may contribute to localization of these genes at the chromosome level. This could provide a lead to target the genes directly or indirectly associated with conotoxins and may provide information on the organization of conopeptide genes in situ. Such a system may also help in the verification and identification of conserved gene sequences, as well as their organization in the genome under certain biological conditions.
Fluorescence in situ hybridization (FISH) provides a versatile approach for genomic characterization and is an excellent tool for chromosome identification and for studies of genome organization, chromosome evolution, cytotaxonomy and introgression (Liu et al., 2003). Repetitiveness of nucleotide sequences is an important feature of all genomes, although the extent to which it occurs within them varies considerably (Grover & Sharma, 2011). 18S rDNA (which is part of the 45S multigene family) as well as telomere- and centromere-associated microsatellite repeats have been widely used as hybridization probes for FISH studies of chromosomes (Porter et al., 1991; Heng, Squire & Tsui, 1992; Gortner et al., 1998; Liu et al., 2003; Vitturi et al., 2005; Bouilly et al., 2008; Gross et al., 2010). In Conus species, the physical mapping of 18S rDNA genes, (GATA)n centromeric and (TTAGGG)n telomeric sequences has not previously been undertaken.
Most of the pioneering in vivo gastropod FISH karyotyping analyses have involved freshwater molluscs (Vitturi et al., 2005; Bouilly et al., 2008). In vitro FISH mapping has been reported for the Biomphalaria glabrata (Bge) cell line (Odoemelam et al., 2009). In bivalves, a short-term culture method was developed and found effective for the preparation of chromosomes from somatic tissues of adult Mytilus edulis (Cornet, 1993). A similar procedure was applied to obtain proliferating cells for chromosome preparation from gill tissue cultures of Crassostrea gigas (Cornet, 2000). The in vitro karyotyping approach and gene mapping done on Bge embryonic cell lines have provided additional useful information on host-pathogen interaction in Biomphalaria (Odoemelam et al., 2009). Hence, tissue dissociated cells such as those maintained in primary cultures and cell lines (Freshney, 2000) are of better utility for karyotyping in conjunction with in situ hybridization procedures.
In vitro karyotyping of Conus species has not yet been accomplished. Based on the review by Thiriot-Quiévreux (2003), only in vivo based karyotypes have been reported, mainly on three conoidean species that include C. mediterraneus and C. coronatus (Conidae; Vitturi & Catalano, 1984) and Terebra nebulosa (Terebridae; Ebied et al., 2000). The two species of Conus exhibited haploid chromosome counts of n = 35 and n = 36, and the karyotypes included a majority of subtelocentric chromosomes, while that of Terebra nebulosa was n = 17 with a karyotype that included a majority of telocentric chromosomes (Vitturi & Catalano, 1984; Ebied et al., 2000; Thiriot-Quiévreux, 2003).
DAPI (4′,6-diamidino-2-phenylindole) has gained wide application in chromosomal studies for its specific binding to the AT-rich regions of the DNA minor groove (Trotta et al., 1995). Such affinity is characterized by a marked increase in DAPI fluorescence quantum yield (Barcellona, Cardiel & Gratton, 1990) with a consequent benefit of revealing both structural and DNA sequence information. DAPI staining intensity of GC-rich euchromatin is weak, while the staining intensity of heterochromatic regions is high (Zink, Sadoni & Stelzer, 2003). This differential staining response, combined with advanced image-processing techniques, could provide a better means of analysing DAPI-counterstained chromosomes, particularly if the size of chromosomes being studied is of the order of a micrometre. Detecting the centromere position is fundamental to karyotyping and is now possible through the application of mathematical algorithms (Arachchige et al., 2010; Abd-El-Haliem, 2012; Mohammadi, 2012; Landini, 2014b). Among others, ImageJ software is a versatile open-source computer program (maintained at US National Institutes of Health, Bethesda, MD) that is capable of compiling and running such computer applications (Rasband, 2010; Schneider, Rasband & Eliceiri, 2012). A reliable nomenclature of chromosome classification based on the p and q arm ratio (r) has been developed by Levan, Fredga & Sandberg (1964). The range of r values provided in this nomenclature has been adopted as a universal tool in chromosome classification (Levan, Fredga & Sandberg, 1964; Guerra, 1986). Thus, a combined approach where DAPI counter-stained chromosomes are analysed with advanced image-processing software, applied in the context of the chromosome classification scheme of Levan et al. (1964), should address issues of apparent ambiguity in chromosomal counting and karyotype analysis.
In this study we report an in vivo and in vitro chromosome count, an in vitro DAPI-karyotype and physical mapping of 18S rDNA genes, GATAn centromeric and TTAGGGn telomeric sequences by FISH in Conus magus Linnaeus, 1758.
MATERIAL AND METHODS
Specimen source and identification
Live individuals of Conus magus were collected from Libas Buenavista, Marinduque Province, Philippines (13°14′ 40.9986″N, 121°57′29.001″E). The snails were maintained at the aquarium facility of the Conus-Turrid Project of the Marine Science Institute, University of the Philippines Diliman, Quezon City. A voucher specimen was submitted and the identity of C. magus was verified by Ms Vivian Ang of the Zoology Division of the Philippine National Museum (Fig. 1A).
Specimen preparation
All materials were sterilized prior to use. Two individuals of C. magus were dissected. The gill and ovary were isolated from one snail and placed in separate petri dishes containing Hepes calcium-magnesium-free (HCMF) dissection solution in preparation for primary cell culture. The ovary was isolated from the second snail and incubated for 4 to 6 h in colcemid (10 µg/ml) solution. The tissue was subsequently fixed in prechilled methanol-acetic acid mixture (3:1).
Primary cell culture
Gill and ovary were minced separately in HCMF dissection solution. Minced samples were dispensed in two sets of four 15-ml Falcon tubes. Tissue cells were dissociated by treatment with 300 µl 0.5% trypsin-EDTA and incubated at 37°C for 5 min. Trypsinization was stopped with the addition of 400 µl Dulbecco's modified Eagle's medium (DMEM). Trituration was done after repeated pipetting (Saloma, Torres-Villanueva & Mirano, 2003). The tubes were centrifuged for 3 min using a Millipore tabletop centrifuge and cells were resuspended in 400 µl DMEM (modified from Cornet, 1993). Cells were seeded into a 35-mm culture dish and maintained and allowed to grow in a CO2 incubator at 5% CO2 and 37 °C. Images of gill and ovarian cells were taken using an Olympus photomicroscope.
Metaphase spread preparation and chromosome counting
In vivo. A volume of 0.5 ml 0.025% trypsin was added to a portion of fixed ovarian sample and incubated for 10 min at 37 °C. Cells were recovered by centrifugation at 2,000 rpm for 2 min and resuspended in 0.5 ml hypotonic KCl for 15–20 min at room temperature. Methanol-acetic acid (3:1) mixture was added, followed by 2 min centrifugation. Cells were subsequently resuspended in methanol-acetic acid mixture (3:1). A drop of sample was released on a slide. The slide was tilted to finely spread the drop and incubated until dry at 95 °C inside a drying oven (modified from Freshney, 2000).
In vitro. A solution of 1% (v/v) colcemid (10 µg/ml) was added to 10-day old gill cells in the culture dish. Cells were incubated for 4 to 6 h at 37 °C followed by incubation with 5% (v/v) 0.025% trypsin for 10 min at 37 °C. Cells were recovered by centrifugation at 2,000 rpm for 2 min and resuspended in 0.5 ml hypotonic KCl for 15–20 min at room temperature. Methanol-acetic acid (3:1) was added, followed by centrifugation for 2 min. Cells were resuspended in methanol-acetic acid mixture (3:1). A drop of sample was released on a slide and tilted to spread the drop finely, then incubated until dry at 95 °C in a drying oven (modified from Freshney, 2000).
Metaphase spreads were viewed and initial counting of chromosomes was done using a Wild Leitz phase-contrast microscope. Images were taken using a Mintron CCD camera. Chromosome counting was done using JMicrovision and Digital Scientific software.
Synthesis of probes and fluorescence labelling
The probe for 18S rDNA was designed through online Primer3Plus software using the 18S rDNA sequence of Paracentrotus lividus (Cantone et al., 1993). The base sequence of probes targeting the centromere (GATA)n and telomere (TTAGGG)n were based on human chromosomes (Ijdo et al., 1991; Vitturi et al., 2005). Probe sequences were submitted to Invitrogen for synthesis and labelling. The 18S rDNA, centromere (GATA)7 and telomere (TTAGGG)5 probes were labelled with Alexa 488, 532 and 568 fluor dyes respectively.
For clarification, the terms ‘(GATA)n’ and ‘(TTAGGG)n’ represent the target SSR sequences of varying number of repeats for FISH physical mapping. These repetitive sequences are characterized by corresponding signal intensities. The terms ‘(GATA)7’ and ‘(TTAGGG)5’ represent the probes of known repeat number used to target (GATA)n and (TTAGGG)n repetitive sequences.
FISH probe and hybridization buffer preparation
All probes were reconstituted in 1X TE buffer (Invitrogen) according to the manufacturer's instructions. Hybridization buffer was prepared using 2X saline sodium citrate (SSC) buffer (Invitrogen), 50% formamide (Ultra Pure Invitrogen), 10% dermatan sulphate from 50% stock (Santa Cruz Biotechnology) and distilled water. The hybridization buffer was left at room temperature and was pre-warmed at 37 °C prior to addition of FISH probes.
Fluorescence in situ hybridization
A FISH procedure (modified after Abbott Laboratories, 2014a, b) was performed. A solution of 2X SSC in a coplin staining jar was pre-warmed at 37 °C. The gill sample slide was incubated in 2X SSC for 30 min at 37 °C. The slide was sequentially dehydrated in 70, 85 and 100% ethanol for 2 min each. One ml of each probe was added to pre-warmed (37 °C) hybridization buffer, centrifuged, vortexed, recentrifuged and heated for 5 min at 73 °C in a waterbath (DigiSystems). The slide of gill cells was denatured for 5 min in 70% formamide/2X SSC at 73 °C in a waterbath and subsequently dehydrated in 70%, 85% and 100% ethanol for 2 min each. After being air dried, 10 µl of the denatured Alexa-labelled probes was added to the slide and covered with a coverslip. The slide was placed in a prehumidified chamber and incubated in the dark overnight at 42 °C. After incubation, the coverslip was removed and the slide initially washed in 2X SSC/0.1% Tween 20 solution for 2–5 min at 73 °C. For final washing, the slide was transferred to 2X SSC/0.1% Tween 20 for 1–2 min at 37 °C. After the slide was air dried inside a black box, 20 µl of DAPI was applied to the sample area and covered with a coverslip. The slide was kept in the dark inside a slide holder with a temperature range of 1.7 to 3 °C until viewed using a fluorescence microscope.
Fluorescence microscopy
The gill slide preparations were viewed using a BX51T Olympus microscope with BX2-FLA2 Olympus reflected-light fluorescence attachment. Three filter cubes (mirror units) for routine fluorescence labelling U-MWU2 (BP330–385, BA420, DM400), U-MNIB3 (BP470–495, BA510IF, DM505) and U-MNG2 (BP530–550, BA590, DM570) were used. Compatibility of flourochromes (DAPI, Alexa 488, Alexa 532 and Alexa 568) used in the study with the corresponding mirror units was verified with the Olympus universal infinity system (UIS) fluorescence mirror units technical specifications (Olympus, 2003, 2008). Simultaneous excitation/emission profiles of Alexa flour dyes were also checked using a fluorescence spectra viewer (Tsurui et al., 2000; Chroma Technology, 2014; Life Technologies, 2014). The images were acquired using the applied spectral imaging (ASI) FishView v. 4.5.
DAPI karyotyping
The chromosomes were analysed using ImageJ v. 1.46o. Two images of DAPI-stained chromosome spreads (sl99cl49 and sl48cl06) were converted to 8-bit greyscale format. An image-processing algorithm in ImageJ (‘Process and Analyze’ menu) that involves morphological operations (enhancement with normalization, background subtraction, image filtering and thresholding) and particle analysis was applied (Vincent & Soille, 1991; Mei, 2003; Grathwohl et al., 2009; Rasband, 2010; Sage, 2011). The blurred version of the image was subtracted from the 8-bit greyscale original format with the unsharp mask (Gaussian blur radius sigma = 5.0 and mask weight = 0.9). A subsequent background subtraction was applied with a rolling ball radius size = 10.0. Binary operations were applied to segment and reconstruct the chromosomes. Segmentation of the chromosomes was achieved after applying greyscale erosion at a minimum pixel value = 5.0, followed by a greyscale dilation at maximum pixel value = 3.0 for spread sl99cl49. For spread sl48cl06 only a greyscale dilation at maximum pixel value = 3.0 was applied. The watershed algorithm revealed the boundaries of the segmented chromosomes at minimum threshold values = 17.0 and 19.0 for spreads sl99cl49 and sl48cl06, respectively. Particle analysis was used to count and extract chromosome boundaries with corresponding x and y coordinates and to set measurement parameters. DAPI-banding patterns were detected and analysed using the ImageJ algorithms Plot Profile, Heatmap Histogram and Analyst (Pean, 2010; Rasband, 2010; Cheng et al., 2013; Jovanović, Perović & Djordjević, 2013; Zhuo et al., 2013).
Chromosome centromeres were localized based on concave points along each chromosome centreline (Soille, 2004; Landini, 2008, 2014a; Rasband, 2010; Mohammadi, 2012). Using ImageJ v. 1.460, individual DAPI counterstained chromosome images corresponding to the blue channel were converted to 8-bit greyscale, binarized and skeletonized. Branches were pruned using the “prune” program in ImageJ (Landini, 2008; Rasband, 2010). The chromosome centreline was established by creating point selections along the chromosome image skeleton, after which a segmented line was fitted to capture the x and y coordinates using a modified macro program (Weller, 2010). Each chromosome was straightened along the centreline and then h-concave transformed using the program of G. Landini (Vincent, 1993; Soille, 2004; Landini, 2008, 2014a; Rasband, 2010). Plot profiles along the centreline of the output images were made. The x and y coordinates of a point with the highest probable degree of concavity, based on the highest pixel value, were determined for each chromosome and were presumed to be the centromere location.
The chromosomes were grouped and paired according to the established centromere location. Chromosome measurements were made using the computer application MicroMeasure v. 3.3 (Reeves & Tear, 2000; Reeves, 2001). The chromosome nomenclature of Levan et al. (1964) was applied based on the respective p and q arm ratios.
FISH physical mapping of 18S rDNA, (GATA)n centromeric and (TTAGGG)n telomeric sequences
The RGB image outputs of FishView v. 4.5 were processed using ImageJ (Rasband, 2010). The images of chromosomal spreads were subjected to image colour thresholding using the RGB colourspace segmentation program Threshold Colour by G. Landini (a modification of Bob Dougherty's BandPass2 filtre) (Landini, 2014b) based on the corresponding relative chromaticity (RGB values) of each fluorescent probe used (Rasband, 2010; Abd-El-Haliem, 2012; Landini, 2014b). Corresponding RGB values (0, 255, 0), (255, 242, 0) and (242, 101, 34) were used for 18S rDNA probe labelled with Alexa 488, centromere (GATA)7 probe labelled with Alexa 532 and telomere (TTAGGG)5 probe labelled with Alexa 568, respectively. The 3D objects counter programme was used to establish particle surface and centres of mass distribution profiles of fluorescent-probe hybridization signals (Bolte & Cordelieres, 2006; Rasband, 2010). For better visualization, ImageJ 3D-surface plots of the yellow fluorescence (Alexa 532), green (Alexa 488) and red-orange (Alexa 568) were separately depicted in the DAPI counterstained chromosomes using the ImageJ plugin fire LUT.
18S rDNA hybridization signals were analysed. The intensity distributions of 18S-probe signals from both spreads sl48cl06 and sl99cl49 (separate and combined) were subjected to D'Agostino-Pearson (K2 omnibus method) and Shapiro-Wilk normality tests for Gaussian distribution. A registered GraphPad Prism v. 5.03 for Windows was used to conduct the normality test procedures. The distributional assumption was checked using a quantile-quantile (Q-Q) plot in R software and graphics (R Core Team, 2014). Using the combined data points from both spreads, the relative frequency distribution of 18S rDNA signals was analysed. Percentiles, sample quartiles [lower (Q1), median (Q2), upper (Q3) and interquartile range (IQR)] and mean absolute deviation (MAD) were computed (Kenney & Keeping, 1962; Vandervieren & Hubert, 2004; FSS International, 2005; Doane & Seward, 2012; Weisstein, 2014). Normalized averaged intensity values (normalized to the average of the major 18S rDNA FISH signals = 100%) were also calculated.
Model-based clustering procedures were conducted using R (R Core Team, 2014). A model for semiparametric estimation and maximization (EM) (Bordes, Mottelet & Vandekerkhove, 2006; Hunter, Wang & Hettmansperger, 2007; Benaglia et al., 2009) and hierarchical clustering and density estimation for parameterized Gaussian mixture models (Fraley & Raftery, 2002; Fraley et al., 2012) were applied in the approximation of the corresponding probability density function (pdf) of classes of 18S rDNA FISH signals.
Hybridization signals for the FISH-targeted 18S rDNA, (GATA)n and (TTAGGG)n sequences were scored and tabulated. FISH signals for (GATA)n and (TTAGGG)n were scored in the centromeric and noncentromeric, and the telomeric and nontelomeric regions, respectively, of both p and q arms of the chromosome pairs. Scoring the 18S rDNA FISH signals involved primarily the establishment of the count for doublets, triplets and multiplets, followed by establishing the scores for singlets in each homologue for each chromosome pair. Based on the number of occurrences, hybridization signal intensities that occurred once, twice and thrice were designated as singlets, doublets and triplets, respectively. Those with more than three occurrences were designated as multiplets.
The scores for 18S rDNA FISH signals were divided into six groups: (1) binding sites detected in each homologue; (2) 18S rDNA signals with positional similarities between two homologues; (3) cumulative scores of singlets with homologous counterparts (singlets; members of doublet/s or triplet/s or multiplet/s) in the other homologue or pair member (p′ and q′); (4) doublets; (5) triplets and (6) multiplets. The score for the 18S rDNA signals for each putative pair is cumulative and is at most a combination of three score components. These are the scores of homologous singlets with patterns (1) (p → p′) or (q → q′): singlets in the p chromosome arm (p-singlets) and q chromosome arm (q-singlets) that possess homologous counterparts with a singlet, or members of a doublet, triplet or multiplet in the other homologue or pair member (p′ and q′); (2) (p → q′) or (q → p′): a singlet found in the p arm of a chromosome is homologous to a singlet or a member of either a doublet, triplet or multiplet in the q arm (or vice versa) of the other homologue. A cumulative score calculated with this score component is labelled with a double dagger (‡); (3) (p → 2p′/3p′ … p′) or (q → 2q′/3q′ … p′): a singlet found in the p arm of a chromosome is homologous to a member of either a doublet, triplet or multiplet in the p arm of the other homologue (p → 2p′/3p′ … p′) or if a singlet found in the q arm of a chromosome is homologous to a member of a doublet, triplet and multiplet in the q arm of the other homologue (q → 2q′/3q′ … p′). The cumulative score calculated with this score component is labelled with a single dagger (†). The coefficient of (‡) and (†) specifies the number of occurrence of each corresponding pattern. A points system was based on full (equivalent scores) and partial (difference of 1 to 2) score agreements between putative homologous chromosome pairs. In 18S, telomeric and centromeric FISH-signal scoring, each chromosome with full score agreement was given a point value of 1.0 and each chromosome with partial score agreement was given a point value of 0.5. Chromosome homologues that agreed in the absence of FISH-probe signals were also given a full point of 1.0. In 18S rDNA singlet, doublet, triplet and multiplet scoring, only those homologues that got a score for homologous signal intensities were included in the points system. For chromosome pairs that do not have homologous singlets, their corresponding partial and full score agreements for doublets, triplets or multiplets were considered if present. A total of 32 points is equivalent to 100 percent. Percent average of positional similarities in p and q arms was obtained by dividing the sum of all scores by the total number of 18S rDNA binding sites (Supplementary material, Tables S3, S4).
These computations revealed the scores of homologous singlets in the p and q arms that confirm agreements in 18S rDNA profiles of putative homologous pairs of chromosomes (ch1 and ch2; ch3 and ch4).
RESULTS
Chromosome count
Using phase-contrast microscopy, the initial diploid chromosome count obtained for the in vitro gill cells of Conus magus was 2n = 32 (Fig. 1B). The initial chromosome count obtained for the in vivo and in vitro ovarian cells of C. magus was n = 16, confirming their haploid status (Fig. 1B, C, E, F).
Using the ImageJ particle analysis, 32 chromosomes were counted in both spreads (sl99cl49 and sl48cl06). Each region of interest (ROI) corresponded to individual chromosome images labelled as indicated in Supplementary material, Figure S1C, F. Profiles for the heterochromatic and euchromatic regions of a median-region chromosome 1 (sl99cl49) and subtelomeric chromosome 27 (sl48cl06) were plotted in an ideogram with corresponding heatmap histogram of intensity profiles, respectively (Supplementary material, Fig. S1).
Chromosome classification
Three major groupings were determined in both DAPI-counterstained chromosome spreads (sl99cl49 and sl48cl06). Twenty-two median-region (m, rlevan = 1.0–1.7), four submedian (sm, rlevan = 1.7–3.0) and six subterminal (st, rlevan = 3.0–7.0) (2n = 22m + 4sm + 6st) chromosomes with haploid counts n = 16 were grouped (Fig. 2A, C). (The r value is the range of ratios of p and q arms; Levan et al., 1964.)
FISH physical map of 18S rDNA, (GATA)n centromeric and (TTAGGG)n telomeric sequences
18S rDNA, (GATA)n and (TTAGGG)n FISH profiles are shown in Supplementary material, Figure S2E–H, Figures 2B, D and 3A, B, respectively. Based on the outcome of the two normality tests (P < 0.0001, α = 0.05), the distributions of 18S rDNA data from the two spreads were not consistent with a Gaussian distribution (i.e. null hypothesis rejected). An intensity distribution plot of combined data points of 18S rDNA FISH signals from spreads sl48cl06 (n = 343) and sl99cl49 (n = 312) was constructed using Doane's formula for non-normal data (Doane, 1976; Keen, 2010; Doane & Seward, 2011) (Supplementary material, Fig. S2A). Data points in the Q-Q plot vs normal theoretical quantiles showed a strongly nonlinear pattern (Supplementary material, Fig. S2A). The Q-Q plot of data points against the quantiles of an exponential distribution showed points lying along a line with slope equal to 1 (Becker, Chambers & Wilks, 1988; Wicklin, 2011; Tobias, 2013) (Supplementary material, Fig. S2A). 18S rDNA FISH signals that were large and strong were identified as ‘major’ while those that were small and weak were identified as ‘minor’, while intermediate signals were identified as ‘medium-to-minor’ or ‘medium’.
Beginning from the 75th percentile (x ≥ Q3 = ln(4)/λ) of the intensity distribution are x-values that indicate relatively large and strong major 18S FISH signals, whose number of occurences decreases with increasing values of signal intensities towards the right-tail exponential area. Values below the 75th percentile ranged from medium 18S FISH signals (x < Q3, x ≥ Q1 = ln(4/3)/λ) to relatively small and weak minor 18S FISH signals (x < Q1), whose number of occurrences increases towards the left-tail exponential area (Supplementary material, Fig. S2A and Table S1). Numerical dispersions were described by the boxplot as projected by the purple vertical broken lines (Supplementary material, Fig. S2A). The distance of the upper fence from Q3 was based on the skewness of the distribution (Vandervieren & Hubert, 2004; Hubert & Vandervieren, 2006). These three groups of 18S signal intensities were also designated as major, medium and minor classes, whose corresponding sample mean (µ) values are 9707.10, 2813.97 and 483.83, respectively (Supplementary material, Table S1). Normalized averaged intensities (normalized to the average of major 18S rDNA signals = 100%) of major, medium and minor 18S rDNA signals vs hybridization class numbered according to decreasing intensity are shown in Supplementary material, Figure S2B. The averaged intensity of the major 18S rDNA signals is about 2.4 and 3.2 mean absolute deviations (MAD = 2/eλ = 2862.84) from that of the medium and minor 18S rDNA signals, respectively. Model-based clustering of 18S rDNA signals suggested two sets of estimators, biased and unbiased, of the corresponding shape parameter (µ) of the major, medium and minor classes of 18S FISH signals. The density estimation parameterized for Gaussian mixture models using R mclust (Fraley & Raftery, 2002; Fraley et al., 2012) indicated that the 18S rDNA data (non-Gaussian) probability density function is approximated by that of three mixture components whose corresponding shape parameters (µ) are unbiased estimators for those of the 18S rDNA major, medium and minor classes (Supplementary material, Fig. S2CI, D and Table S2). The shape parameters of the three mixture components suggested by the semiparametric-EM model (Bordes et al., 2006; Hunter et al., 2007; Benaglia et al., 2009) are biased estimators of the true value parameter of each of the class of 18S rDNA signals (Supplementary material, Fig. S2CII and Table S2).
The scores of probe hybridization signals for 18S rDNA, (TTAGGG)n telomeric, (GATA)n centromeric sequences, nonterminal (TTAGGG)n, interstitial telomeric sites (ITS) and noncentromeric (GATA)n sequences were tabulated for spreads sl99cl49 (Supplementary material, Table S3) and sl48cl06 (Supplementary material, Table S4).
Both parametric and nonparametric paired-sample tests indicate significantly effective pairing between the mean scores of 18S rDNA signals of the two spreads (Supplementary material, Table S5). A paired samples t test failed to reveal a statistically significant difference between the mean (M) 18S FISH signal scores of sl48cl06 (M = 58.31, std dev = 10.79) and scores of sl99cl49 (M = 53.62, std dev = 15.8), t(12) = 1.652, P = 0.1245, α = .05 (Supplementary material, Table S5). Likewise, a Wilcoxon matched-pairs signed-rank test (a nonparametric test that does not assume Gaussian distribution) revealed no significant difference between the median scores of both spreads (Supplementary material, Table S5).
Percent full-score agreements profiled in p, q and centromere (c) of putative chromosome pairs in spread sl99cl49 were in the range of p[55-57%(18S); 53-88%(TTAGGG)n; 34%(GATA)n]; q[38-66%(18S); 41-59%(TTAGGG)n; 31%(GATA)n] and 63% c(GATA)n, respectively. In spread sl48cl076, scores were in the range of p[50-62%(18S); 56-69%(TTAGGG)n; 41%(GATA)n]; q[52-78%(18S); 53-63%(TTAGGG)n; 47%(GATA)n] and 75% c(GATA)n, respectively.
Full-score agreements were consistent in the p-arms of six median region and q-arms of four submedian chromosomes and q-arms of four subterminal chromosomes from both spreads. These are chromosomes 1, 2, 3, 4, 5 and 6 (Supplementary material, Tables S3, S4, p arm, 18S, column S2) and chromosomes 23, 24, 25, 26, 27, 28, 29 and 30 (Supplementary material, Tables S3, S4, q arm, 18S, column S2), respectively. An example is shown by the profiles of homologous signal intensities between median region chromosomes 1(m) and 2(m) of spread sl99cl49 that support the presumed pairing of these two chromosomes (Fig. 4C).
Similar chromosomal organization for centromeric (GATA)n and telomeric (TTAGGG)n sequences was detected in both C. magus chromosome spreads (Supplementary material, Table S6). Probe-hybridization signals for (GATA)n sequences were detected clustering in the centromeric regions of 23 (13 median region, 4 submedian region and 6 subtelomeric region) out of 32 chromosomes (Fig. 2B, D, Supplementary material, Table S6). Probe-hybridization signals for telomeric (TTAGGG)n sequences were detected in both p and q terminals of 15 (10 median region, 2 submedian region and 3 subtelomeric region) out of 32 chromosomes (Fig. 3A, B, Supplementary material, Table S6). Conus magus chromosomes analysed for centromere location and FISH physical mapping of the 18S rDNA, (GATA)n centromeric and (TTAGGG)n telomeric sequences are summarized in Figure 4. An algorithm was designed to plot the chromosome positions of 18S rDNA signals in an ideogram (Fig. 4). Comparison of 18S rDNA FISH signals from chromosomes 1(m) and 2(m) of spread sl99cl49 relative to their locations in either p or q arm, distance from the centromere and signal intensities, suggests presence of sequence agreements between p and q arms of each homologue. Direct chromosome-segment analysis based on the position of the 18S rDNA binding sites of specific signal intensities led to reconstruction of further agreements between chromosomes 1(m) and 2(m) (Fig. 4).
DISCUSSION
The haploid chromosome number (n = 16) that was established in gill cells and confirmed in the ovarian cells of Conus magus is close to that reported by Kuschakewitsch (1913) in C. mediterraneus (n = 14) (Kuschakewitsch, 1913; Nishikawa, 1962). It is also close to that of another neogastropod Nucella lapillus (n = 13–18) (Bantock & Cockayne, 1975; Page, 1988; Dixon et al., 1994; Pascoe & Dixon, 1994; Pascoe et al., 1996; Thiriot-Quiévreux, 2003). However, it is low compared with those of C. coronatus (n = 35; Vitturi and Catalano, 1984) and C. mediterraneus (n = 36; Ebied et al., 2000). A possible explanation of this finding is karyological divergence due to hypoploidization (Barsiene, 1994). In Nucella the haploid chromosome number of N. lapillus (n = 13–18) is lower than that of N. canaliculata, N. lima and N. lamellosa (n = 35). Molecular, morphological and palaeontological evidence suggest that the reduced number of chromosomes in N. lapillus is a derived state (Collins et al., 1996; Thiriot-Quiévreux, 2003). A similar evolutionary phenomenon could explain the variations found in the chromosome numbers of Conus, but requires testing with additional karyological data in combination with existing phylogenies of the genus (Vitturi & Catalano, 1984; Ebied et al., 2000; Thiriot-Quiévreux, 2003). Changes in environmental conditions such as temperature, salinity and pollutants may serve as possible selection forces that result in karyologic variation such as hypoploidy, hyperploidy and polyploidy (Barsiene, 1994).
Chromosomes from both in vitro spreads (sl99cl49 and sl48cl06) were grouped according to the categories of Levan et al. (1964) (2n = 22m + 4sm + 6st). Although both spreads represent the same species, animal and tissue, there are substantial morphological differences between them. Figure 2A of spread sl99cl49 shows extended chromosomes, while Figure 2C represents sl48cl06 and shows nonextended chromosomes. In Figure 2A, for example, the largest median region chromosome (chromosome 1) is larger than the largest submedian region chromosome (chromosome 12), which is larger than the largest subterminal region chromosome (chromosome 14). The sizes are reversed in Figure 2C (chromosome 14 is larger than chromosome 12, which is about the same size as chromosome 1). This may be an indication of a Robertsonian fusion or translocation of segments between the two nonhomologous chromosomes 12 and 14. A similar scenario was suggested in N. lapillus, in which Robertsonian polymorphism precluded unequivocal identification of individual chromosomes in the karyotype (Pascoe et al., 1996). In addition, chromosomal deletions and insertions are conceivable, as indicated by nearly uniform lengths of the median region chromosomes in Figure 2C, whereas in Figure 2A chromosomes 1 to 6 are significantly longer.
The system of karyotype designation developed in this work was appropriate for analysing the minute chromosomes of C. magus. The combination of Levan's system of nomenclature with Mohammadi's (2012) method of centromere location based on concave points and Landini's (2014a) h-concave transform method was effective. It should prove useful for derivation of the karyotypes of other Conus species in the future. DAPI-banding (C-like) patterns that needs further verification were visible in DAPI-counterstained chromosomes in both spreads (Supplementary material, Fig. S1A, B, D–E). Some locations of putative DAPI-bands approximate or are consistent with the region identified for the centromere position of the indicated chromosomes. It has been shown that the effect of chromosomal denaturation on DAPI banding leads to progressive occurrence of three different kinds of bands in human chromosomes (Heng & Tsui, 1993). First, DAPI multibanding (the equivalent of Q-banding) under short denaturation time, then partial C-banding including distamycin A (DA)/DAPI banding with a longer denaturation time (only some heterochromatic regions are visible), and finally a C-banding pattern after further denaturation (Heng & Tsui, 1993; Pieczarka et al., 2006). In Q-banding, the bright bands are primarily composed of DNA rich in adenine and thymine, while the dull bands are rich in guanine and cytosine (Strachan & Read, 1996). This is useful for examining chromosomal translocations (O'Connor, 2008). C-banding stains areas of heterochromatin, which is tightly packed and repetitive DNA. This is useful in staining centromeric regions of chromosomes and other regions containing constitutive heterochromatin (Strachan & Read, 1996). The combination of fluorescence in situ hybridization (FISH) with DAPI banding allows the simultaneous detection of signals from the DNA probes and the identification of the chromosomal band location of the probe (Heng & Tsui, 1993).
In C. magus there was a variable number of FISH signals for 18S rDNA, (GATA)n and (TTAGGG)n sequences as indicated by the FISH profiles (Supplementary material, Fig. S2E–H, Figs 2–3) and the fluorescence probe hybridization signals scored in two chromosome spreads (Supplementary material, Tables S3, S4). These repetitive sequences are molecular markers commonly used as chromosomal landmarks that can be of help in understanding genomic and karyotypic evolution (Kobayashi, 2011). The genome size remained constant in two cells, from the same tissue from the same individual with observed homologous repetitive-sequence FISH chromosome profiles. Statistical test (parametric and nonparametric paired-sample tests) results suggest that a relationship exists between the scores of 18S rDNA signals from the two spreads (Supplementary material, Table S5). If scores of agreement represent the fraction of the genome conserved after a genetic event that may have taken place, candidates from these presumed stable regions of specific chromosome types from both spreads may be selected as potential targets for chromosomal landmarks. Targeting homologous 18S rDNA binding sites from these putative pairs of chromosomes may resolve stable molecular landmarks. In addition to the system of classification of chromosomes described earlier, we have shown a comprehensive pairing of chromosomes based on their homologous signal intensities. This facilitated an inter-spread comparison of 18S rDNA profiles showing an average of three 18S rDNA signal intensities identical for corresponding chromosomes from the two spreads (data not shown). Conceivably, the discrepancies between the observed full-score agreements in 18S rDNA probe hybridization signals for the chromosome pairs in both spreads may indicate chromosome structural modifications (e.g. presumptive chromosome fusions). For example, a difference in the lengths of p and q arms was evident between chromosomes 1 and 2 (sl99cl49) (Fig. 4C). In addition, we observed variable clusters of 18S rDNA of specific signal intensities among 18S rDNA signals with identical intensities (i.e. homologous) and positional similarities can be observed on the two chromosomes. Further putative agreements between chromosomes 1 and 2 (sl99cl49) were reconstructed after being challenged by rearranging the positions of chromosome segments discerned from the 18S rDNA profile of the presumed homologue (Fig. 4C). Prominent differences between the lengths of p-arms and q-arms of C1 and C2 indicate possible events of insertion, addition, deletion, duplication and translocation of chromosome segments. If it is assumed that each chromosome contains the same amount of target DNA and that DNA-DNA in situ hybridization is quantitative, then the hybridization signals should have the same fluorescence intensity in the homologous spots (Celeda et al., 1994). However, in this case the correspondence is not always linear due to purported genetic modifications. Chromosomes that show similar (TTAGGG)n and (GATA)n chromosome organization in both spreads are summarized in Supplementary material, Table S6. The utility of these chromosomal regions as additional FISH landmarks may complement that of the presumed 18S-conserved regions. Perhaps in the future the position of heterochromatic and euchromatic regions derived from the DAPI-banding patterns of these putative pairs of chromosomes (Supplementary material, Fig. S1) may be correlated with their FISH profiles of repetitive sequences. Stringency conditions of FISH may be further modified (e.g. formamide concentration, temperature and length of hybridization time) to fully discriminate specific regions of hybridization (Celeda et al., 1994; Durm et al., 1996).
The 45S gene plays an important role in protein synthesis since it encodes the ribosomal RNAs (rRNAs) 18S, 5.8S and 28S. The ribosomal genes are clustered in tandemly repeated units or organized in an array. In every unit, each ribosomal gene is interspersed by spacer regions or primary transcripts, namely the external transcribed spacer (ETS) prior to the 18S gene, and internal transcribed spacers 1 and 2 (ITS1 and ITS2) located on either side of the 5.8S gene. In all eukaryotes, clusters of rRNA genes and intergenic spacers are located on certain chromosomes and form what are known as nucleolar-organizing regions (NORs) (Eickbush & Eickbush, 2007).
A portion of 18S rDNA from Paracentrotus lividus was used to target the 18S gene of C. magus. Under stringent conditions, a wide range of strong, medium and weak 18S rDNA FISH signal intensities was detected. The presence of signals was not exclusive to any single region—paracentromeric, centromeric, interstitial or telomeric—of the different chromosomes. In both spreads, these signals were observed in two to four different regions of each chromosome (Supplementary material, Fig. S2E–H). The ubiquity of these signals enabled discrimination of the major 18S rDNA hybridization sites from medium and minor ones (Islam-Faridi, Majid & Nelson, 2006). Each hybridization site (or spot) found in each chromosome possesses specific fluorescence intensity. Signal intensity gaps are supposed to facilitate the discrimination of major from minor binding sites (Durm et al., 1997). However, in our study the data were continuous, leading to difficulty in identifying the intensity gaps. The distribution was heavily skewed to the right, and was identified as exponential by means of Q-Q plots. So far two novel approaches were conceived to discriminate major from minor 18S rDNA FISH binding sites: (1) The use of computed measures of dispersion (quartiles and percentiles) of the exponentially distributed 18S rDNA signals provided an initial description of the confines of the three presumed classes (major, medium and minor) of 18S rDNA signals in the 18S data distribution (Supplementary material, Fig. S2A, B). (2) Model-based clustering of 18S rDNA signals led to the identification of unbiased estimators of the means (µ) of the three mixture components (non-Gaussian) whose classification outputs further substantiated the delineation of major, from medium and minor 18S rDNA signals (Fraley & Raftery, 2002; Verron, Tiplica & Kobi, 2010; Fraley et al., 2012) (Supplementary material, Fig. S2C, D).
The observed number of 18S rDNA probe hybridization signals in C. magus is relatively high compared to those found in Cantareus land snails, Scarabinae beetles, Symphysodon fishes, Triticeae grasses, squamates and lizards (Porter et al., 1991; Dubcovsky & Dvorak, 1995; Vitturi et al., 2005; Gross et al., 2010; Cabral-de-Mello et al., 2011). This could be indicative of a possible great demand for synthesis of proteins in this species. Although the 18S rDNA probe was not designed to be specific either to expressed ribosomal genes or to silent copies of the genes often highly compacted in dense chromatin, such a high genomic copy number of rDNA relative to other genes could be attributed to the fact that, unlike the protein-coding genes, rDNA loci that harbour 18S rDNA cannot undergo secondary rounds of amplification via translation when organisms require more rRNA transcripts (Prokopowich, Gregory & Crease, 2003). This possibly reflects the rapid synthesis of highly diversified conopeptides which are the ultimate by-products of post-translational modifications of prepropeptide gene secretions utilized by C. magus and other members of Conidae to immobilize their prey (Olivera, 2006). If the genome size remains fairly constant and the reduction in chromosome number in C. magus is due to chromosomal fusions as in Nucella, than one might expect to see twice as many 18S sites per chromosome as in a species that has n = 35 or 36.
The presence of major 18S rDNA probe hybridization signals within the indicated regions suggests possible presence of 18S rDNA (presumably with 5.8S and 28S) in tandem repeats—major rDNA loci. Conversely, minor 18S rDNA probe hybridization signals suggest the possible presence of a single or few transcription units in the indicated region, which could be inactive sites in the heterochromatin—minor rDNA loci (Supplementary material, Fig. S2E–H). Ribosomal genes are well known for their ubiquity and utmost degree of functional and sequence conservation (Prokopowich et al., 2003). However, their sequences can nevertheless change over time in a highly orchestrated manner, a phenomenon described as ‘concerted evolution’ (Eickbush & Eickbush, 2007). Scores of singlets revealed plenty of positional similarities and few hybridization sites of homologous signal intensities between two putative homologous chromosomes (Supplementary material, Tables S3, S4). Singlets with positional similarities may include major, medium or minor 18S rDNA hybridization sites. These sites of hybridization, although of similar position based on their centres of mass, were not reflective of homologous DNA sequences due to unique fluorescence intensities. It is possible that different numbers of rDNA units are present in each hybridization site. Partial and non-agreements between putative homologous chromosomes may perhaps be due to concerted evolution that results in rapid horizontal homogenization of a select variant through a number of molecular processes such as unequal crossing over, gene conversion, movement and multiplication of the major rDNA clusters without fusions or other chromosomal rearrangements that may generate one recombinant chromosome with more rDNA units and another chromosome with fewer units (Eickbush & Eickbush, 2007; Cabral-de-Mello et al., 2011).
Homologous singlets are chromosome regions that could be considered as FISH hotspots for molecular landmarks and are reflective of homologous DNA sequences in homologous chromosomes. Doublets and triplets, on the other hand, indicate presence of multiple copies of specific DNA sequences, perhaps for amplified synthesis of proteins. Homologous doublets further confirm agreements between two homologous chromosomes (Supplementary material, Tables S3, S4). Presence of potentially linked genes was discernible based on the distance profiles of these signal intensities of a specific type (data not shown).
The high frequency of minor 18S rDNA signals relative to that of major and medium 18S rDNA signals is noticeable (as indicated by its density in the left tail exponential area of 18S rDNA intensity distribution) (Supplementary material, Fig. S2A) and these signals are widely dispersed in the genome (Supplementary material, Fig. S2E–H). It is possible that rDNA loci change position by the same, albeit unknown, mechanism that results in the dispersion of heterochromatic sequences through a genome. If minor rDNA loci contain functional rDNA units, their copy numbers can potentially be magnified by unequal crossing-over to become major rDNA loci (Gross et al., 2010). Perhaps the presence of medium rDNA signals indicates such a transition process.
The genomes of most eukaryotes contain many families of repetitive DNA sequences. Many of these repeats, also known as microsatellites or tandem simple sequence repeats (SSRs), occur in the form of tandem arrays localized near centromeres or telomeres or, less frequently, at interstitial regions of whole chromosome arms (Meyne et al., 1990; Li et al., 2002). Two main observations can be made from the FISH profiles of (GATA)n and (TTAGGG)n in C. magus: the presence of hybridization signals (1) in the centromeric and noncentromeric regions, and (2) in the telomeric and nontelomeric regions. These involve both the p and q arms (Supplementary material, Tables S3, S4).
The observations of FISH hybridization profiles may provide clues as to how the chromosome number of C. magus may possibly have been reduced during karyological evolution. Patterns of hybridization signals were similar in both spreads sl99cl49 and sl48cl06. These were indicated by the FISH signals of (GATA)n found in the centromeric regions of 13 median, 4 submedian and 6 subterminal region chromosomes (Fig. 2B, D, Supplementary material, Table S6) and those of (TTAGGG)n at both p and q terminals of 10 median, 2 submedian and 3 subterminal region chromosomes (Fig. 3A, B, Supplementary material, Table S6). These similar patterns of hybridization signals may perhaps represent similar ancestral sequences that were retained by C. magus for telomeres and centromeres. Dissimilarities in chromosome morphology and FISH hybridization signals were also observed between the two spreads. Chromosome structural abnormalities brought about by genetic translocations may provide explanations for this. Two possible forms of chromosome structural rearrangements may have progessively occurred in C. magus in a sequential manner as follows. (1) The breakage, proximal to or within the centromere of two usually nonhomologous acro- or telocentric chromosomes, followed by reunion uniting the longer and shorter segments, is a characteristic of a Robertsonian translocation (Jones, 1998). It is suggested that a similar mechanism could have caused the reduction in chromosome number in the ancestors of C. magus. (2) Reciprocal translocation, a mutual exchange between terminal segments from the arms of two nonhomologous chromosomes, could have possibly occurred in the zygote, prior to embryogenesis, leading to cellular mosaicism, a mixture of karyologically different cells in the same organ of one individual (Huret, Leonard & Savage, 2000; Leshin, 2003). A mosaic individual is made of two (or more) cell populations characterized by differences in the chromosomes (Huret et al., 2000). In this case, structural changes occur within chromosomes themselves, not accompanied by numerical change (Huret et al., 2000). These cell populations, however, come from one, and only one, zygote (Huret et al., 2000). These could be substantiated by the same chromosome count in gill cells and ovarian cells, suggesting a chromosome anomaly that is constitutional in nature where the whole organism holds the same chromosome abnormalities as a mosaic (derived from the original cell where the anomaly first arose) discerned through chromosome morphology and hybridization signals, a phenomenon described as balanced reciprocal-translocation mosaicism (BRTM) (Opheim et al., 1995; Leegte et al., 1998; Huret et al., 2000; Vargas & Fernández-Novoa, 2001; Baptista et al., 2005; Akbas et al., 2012). Comparing the chromosome morphology, count and FISH hybridization signals from both spreads indicates structural polymorphism. However, the chromosome number and centromere indices were the same in both cells, so the change was apparently balanced, with no demonstrable loss or gain of genetic material. DNA sequencing data is required to test these suggestions. Further analysis of the fluorescence in situ hybridization signals herein may help detect the location of genes possibly disrupted in the breakpoints brought about by balanced reciprocal translocation indicating mosaicism distinguished by variable FISH signals in two or more cells (Akbas et al., 2012). This may also help in providing explanations for phenotypic abnormalities associated with gene expression brought about by cryptic imbalances (Baptista et al., 2005; Akbas et al., 2012).
(TTAGGG)n are noncoding DNA repeats that compose the chromosome telomere regions and protect the chromosome ends from degradation (Olovnikov, 1973; Morin, 1989; Meyne et al., 1990; Nomoto, Hirai & Ueshima, 2001). In the present study these repeats were found to occur in tandem arrays localized near both centromeres and telomeres and, less frequently, ininterstitial regions or whole chromosome arms. The rearrangement and amplification of these repeats have been hypothesized to allow greater flexibility for karyotype changes (Meyne et al., 1990). Furthermore, the distribution of terminal telomeric and interstitial telomeric (TTAGGG)n sequences in eukaryotic chromosomes may provide information on karyological evolution (Meyne et al., 1990; Traldi et al., 2013). Ancestral karyotypes have the (TTAGGG)n sequences only at the telomeres, while in the more derived condition the sequences are found in nontelomeric sites, as shown by comparative studies of mammals (Baker & Bickham, 1980; Meyne et al., 1990).
Interstitial telomeric (TTAGGG)n signals (ITSs) were detected, as exemplified by those in chromosomes 3, 6, 12 and 13 (Fig. 3A) and 3, 25, 27, 28 and 30 (Fig. 3B), as have been reported in vertebrates (Meyne et al., 1990). ITSs have also been observed in some chromosomes of the freshwater snail Biwamelania habei and indicate presumed telomeric chromosome fusion points of a Robertsonian translocation (Nomoto et al., 2001). Two possible mechanisms have been proposed to explain the generation of these interstitial telomeric sites: fusion of ancestral chromosomes at telomeric sites or amplification of endogenous, short (TTAGGG)n tandem repeats within the chromosome arms (Nomoto et al., 2001). This has also been proposed for Nucella lapillus (Pascoe et al., 1996), so a similar mechanism in Conus does not seem unlikely. Sequential fission and fusion may occur during the post-amplification of ITSs, especially in the pericentric regions of chromosomes (Meyne et al., 1990).
In the case of the FISH signals for noncentromeric (GATA)n, as detected in most chromosomes (Fig. 2B, D), similar findings have been reported in a range of organisms. In the tomato and chickpea genomes GATA signals are clustered in centromeric areas, with additional GATA signals along some chromosomes (Schmidt, Kubis & Heslop-Harrison, 1995; Gortner et al., 1998). In two land snails, Cantareus aspersus and C. mazzullii, the GATA probe hybridized with small regions dispersed along the chromosomal body (Vitturi et al., 2005). Consistent findings have also been reported for congeneric species of cyprinid fish and Pamphagus grasshoppers (Gold & Price, 1985; Vitturi, Mansueto & Di Paola Ficarella, 1993; Vitturi et al., 2005).
In Conus species, it has been proposed that there are special recombination mechanisms that are involved in the conservation of coding genes, particularly the conopeptide genes (Olivera et al., 1999). Although the mechanisms of gene rearrangements (i.e. slipped-strand mispairing) and the hypermutability of microsatellites are still under discussion (Levinson & Gutman, 1987; Rawlings et al., 2010; Grover & Sharma, 2011), the functional role of SSRs is substantiated by the effect of their genomic distribution in taxon-specific chromosomal organization, regulation of gene activity, recombination, DNA replication, the cell cycle and the mismatch-repair system (Li et al., 2002). Perhaps the wide distribution of these repetitive sequences in C. magus chromosomes may play a role in the rapid speciation of cone snails (Olivera et al., 1999).
To our knowledge, this is the first report on karyotyping in C. magus. It is also the first report on in vitro DAPI karyotyping and FISH physical mapping of chromosomal genomic distribution of 18S rDNA, (GATA)n and (TTAGGG)n repeats in Conus, and also the first report on the use of a system for karytoype derivation in this genus involving the use of primary cell culture and centromere localization techniques based on concave points, along with the application of Levan's chromosome nomenclature. The system will be further tested and used for the derivation of other Conus karyotypes. The C. magus DAPI karyotype and genome landmarks as revealed by the FISH physical map of 18S rDNA, (GATA)n and (TTAGGG)n sequences may help in the elucidation of the evolution of karyotypes in Conus species, as well as in the FISH detection and localization of Conus chromosomal genes.
SUPPLEMENTARY MATERIAL
Supplementary Material is available at Journal of Molluscan Studies online.
ACKNOWLEDGEMENTS
This work was supported by a PhD Dissertation Grant from the Philippine Council for Health Research and Development of the Department of Science and Technology of the National Government of the Republic of the Philippines. Specimen collection was carried out through the support of the PharmSeas Study Group (Dr G. Concepcion, PI) of the University of the Philippines (UP) Marine Science Institute and the initial identification of the C. magus specimen by Meljune Chicote. We are grateful to Vivian Ang of the Zoology Division of the Philippine National Museum for the verification of the identity of C. magus. For substantial comments and review of our manuscript we thank Dr Jerry Harasewych of the Department of Invertebrate Zoology at the National Museum of Natural History, Smithsonian Institution, Washington, D.C., Dr Alan J. Kohn of University of Washington, Dr Janice Voltzow and Dr David G Reid of Journal of Molluscan Studies. Thirdy Buno and Janina Lazo of the UP National Institute of Physics and UP Multi-Dimensional Imaging Center; and Rachel Ramirez of the Laboratory of Molecular Cell Biology (Dr C. Saloma, PI) of the UP National Institute of Molecular Biology and Biotechnology Laboratory, are thanked for their assistance in laboratory procedures. Dr Maria Corazon De Ungria and the UP DNA Analysis Laboratory Team of the UP Natural Sciences Research Institute kindly permitted the use of their laboratory and phase contrast microscope and the Department of Biochemistry and Molecular Biology, College of Medicine UP Manila provided their phase contrast microscope and photo-documentation unit for field use.
REFERENCES