Genotyping of European Toxoplasma gondii strains by a new high-resolution next-generation sequencing-based method.

PURPOSE: A new high-resolution next-generation sequencing (NGS)-based method was established to type closely related European type II Toxoplasma gondii strains. METHODS: T. gondii field isolates were collected from different parts of Europe and assessed by whole genome sequencing (WGS). In comparison to ME49 (a type II reference strain), highly polymorphic regions (HPRs) were identified, showing a considerable number of single nucleotide polymorphisms (SNPs). After confirmation by Sanger sequencing, 18 HPRs were used to design a primer panel for multiplex PCR to establish a multilocus Ion AmpliSeq typing method. Toxoplasma gondii isolates and T. gondii present in clinical samples were typed with the new method. The sensitivity of the method was tested with serially diluted reference DNA samples. RESULTS: Among type II specimens, the method could differentiate the same number of haplotypes as the reference standard, microsatellite (MS) typing. Passages of the same isolates and specimens originating from abortion outbreaks were identified as identical. In addition, seven different genotypes, two atypical and two recombinant specimens were clearly distinguished from each other by the method. Furthermore, almost all SNPs detected by the Ion AmpliSeq method corresponded to those expected based on WGS. By testing serially diluted DNA samples, the method exhibited a similar analytical sensitivity as MS typing. CONCLUSION: The new method can distinguish different T. gondii genotypes and detect intra-genotype variability among European type II T. gondii strains. Furthermore, with WGS data additional target regions can be added to the method to potentially increase typing resolution.

A comparison of sequence and length polymorphism for genotyping Cryptosporidium isolates.

Simple sequence repeat markers have played an important role in elucidating the epidemiology of human and animal cryptosporidiosis. The drawback of sequence length polymorphisms is that nucleotide substitutions remain undetected. As some laboratories have opted for using length polymorphisms, while others have relied on sequencing, there is a need to compare both methods. We used a diversified set of unique length polymorphisms and matching nucleotide sequences to assess the ability of each genotyping protocol to discern clusters of related Cryptosporidium parvum isolates. We found a weak correlation between the two distance measures for individual markers. This analysis was extended to four-locus genotypes based on sequence length data or concatenated sequences from the same loci. We interrogated these data to assess whether one would reach the same conclusions regardless of the genotyping method. Clusters of isolates generated with the concatenated sequences were not observed with amplicon length, indicating that inferences on the structure of a Cryptosporidium population depend on the genotyping method. Moreover, isolate clusters derived from concatenated sequences were dependent on the algorithm used to calculate distances. These results emphasize the need for harmonizing genotyping tools, not only by selecting informative markers, but also by standardizing the entire genotyping method.

Molecular typing of Giardia duodenalis isolates from German travellers.

Giardia duodenalis isolates from German travellers returning from tropical areas were characterised by PCR amplification and sequencing of fragments of the beta-giardin (bg), glutamate dehydrogenase (gdh) and triose phosphate isomerase (tpi) genes. Assignment of isolates to specific G. duodenalis assemblages was found to differ according to the marker used. Indeed, at the bg locus, assemblages A and B were identified, with a higher prevalence of the former over the latter, whereas at the tpi and gdh loci, most samples were classified as assemblage B. In agreement with previous studies, sequence analysis showed that assemblage B isolates have a higher genetic polymorphism than assemblage A isolates, and novel variants were described. The degree of polymorphism was shown in a graphical representation of the polymorphic sites generating a novel sequence, the heterogeneous positions common to assemblages A and B (double peaks), that may represent mixed assemblage infection and the heterogeneous positions detected at random sites. Notably, assemblage D, which is considered to be adapted to dogs, was found at the gdh locus in two samples originating from southern Asia, as novel genotypes. By comparing the geographical origin of the infected cases and the number of German travellers visiting the areas considered, India and west Africa appeared to be the areas associated to the highest risk of acquiring Giardia infection. The analysis of the geographical distribution of the genotypes did not suggest any particular geographical clustering pattern, but it may be useful to evaluate these results with a higher number of isolates. Most of the samples typed at the three markers could not be assigned unequivocally to either assemblage A or B, and this was confirmed also by a real-time PCR assay, using a set of assemblage-specific primers. The results of this study reinforce the notion that genetic exchanges and allelic sequence heterogeneity represent major obstacles towards understanding the epidemiology of giardiasis and that exposure to Giardia parasites in endemic areas often results in mixed infections in returning travellers.

Prevalence and molecular identification of Cryptosporidium isolates from pet lizards and snakes in Italy.

In order to acquire prevalence and genetic data on Cryptosporidium infections in captive lizards and snakes kept as pets, a survey was conducted on 150 individual reptiles from southern Italy. Fecal samples were preserved in 5% formalin and analyzed using a commercial immunofluorescence assay (IFA) for the detection of Cryptosporidium oocysts and Giardia cysts. IFA revealed the presence of Cryptosporidium oocysts in nine of the 150 samples examined (6.0%), precisely in 6/125 snakes (4.8%) and in 3/25 lizards (12.0%); all fecal samples tested negative for the presence of Giardia cysts. Molecular characterization based on nested PCR amplification and sequencing of the SSU-rRNA gene, revealed the presence of Cryptosporidium serpentis in three samples from snakes (Boa constrictor constrictor, Elapheguttata guttata guttata and Python molurus).

Ready-to-eat salads and berry fruits purchased in Italy contaminated by Cryptosporidium spp., Giardia duodenalis, and Entamoeba histolytica.

Ready-to-eat (RTE) salads and berries are increasingly consumed in industrialized countries. These products can be contaminated by pathogenic parasites that have been responsible for foodborne outbreaks worldwide. In Italy, there are few data on contamination of RTE salads and berries with parasite transmission stages and this requires more-in-depth investigations. To estimate the prevalence of contamination with Cryptosporidium spp. and Giardia duodenalis in these fresh products, a total of 324 packages of local RTE mixed salads - belonging to three different industrial brands - and 324 packages of berries - blueberries from Peru, blackberries from Mexico, raspberries from Italy - were bought from supermarkets located in the Provinces of Bari and Foggia, Apulia, Italy. A pool size of nine packages was chosen and a total of 72 pools were processed in the whole year. After washing, the pellets were examined by microscopy (FLOTAC) and tested using conventional simplex PCR, targeting Cryptosporidium spp., Giardia duodenalis, and Entamoeba spp., and sequencing. Several Cryptosporidium species and Giardia duodenalis assemblages, some of which are of potential zoonotic relevance, as well as Entamoeba spp., were identified in both matrices. By microscopy, Giardia-like cysts in local raspberries and Entamoeba-like cysts in imported blueberries were detected. Giardia duodenalis (Assemblages A, B and E) and Entamoeba histolytica were molecularly confirmed with overall prevalences of 4.6% (95% C.I. 3.0-6.8) and 1% (95% C.I. 0.3-2.1), respectively. Molecular methods identified Cryptosporidium ryanae, Cryptosporidium bovis, Cryptosporidium xiaoi, and Cryptosporidium ubiquitum in both matrices, with a prevalence of 5.1% (95% C.I. 3.3-7.3). A distinct seasonality in prevalence was observed for G. duodenalis, with most positives occurring in spring, whereas Cryptosporidium showed no significant seasonal variations. These results highlight that inadequate management of fresh produce, both locally produced and imported, along the food chain may have the potential for consequences on human health.

Cases of cryptosporidiosis co-infections in AIDS patients: a correlation between clinical presentation and GP60 subgenotype lineages from aged formalin-fixed stool samples.

Nine cases of cryptosporidiosis co-infections in AIDS patients were clinically categorised into severe (patients 1, 3, 8 and 9), moderate (patients 4 and 5) and mild (patients 2, 6 and 7). Formalin-fixed faecal specimens from these patients were treated to obtain high quality DNA competent for amplification and sequencing of the 60-kDa glycoprotein (GP60) gene. Sequence analysis revealed that one patient was infected with Cryptosporidium hominis whereas the remaining eight patients were infected with C. parvum. Interestingly, the patients showing severe cryptosporidiosis harboured two subtypes within the C. parvum allelic family IIc (IIcA5G3 and IIcA5G3R2), whereas patients with moderate or mild infections showed various subtypes of the C. parvum allelic family IIa (IIaA14G2R1, IIaA15G2R1, IIaA17G3R1 and IIaA18G3R1). DNA extraction and genotyping of Cryptosporidium spp. is a challenging task on formalin-fixed stool samples, whose diagnostic outcome is age-dependent. The method herein reported represents a step forward routine diagnosis and improves epidemiology of HIV-related clinical cases. Due to the need to elucidate genetic richness of Cryptosporidium human isolates, this approach represents a useful tool to correlate individual differences in symptoms to subgenotyping lineages.

A validated method to identify Echinococcus granulosus sensu lato at species level.

The zoonotic tapeworm Echinococcus granulosus sensu lato (s.l.) represents a species complex encompassing multiple causative agents of cystic echinococcosis, a neglected tropical disease affecting more than one million people in the world. At least eight genotypes, grouped in five species, are currently recognized within this species complex, and they differ in terms of relative public health impact. Here we present a molecular method that first identifies the common E. granulosus sensu stricto (s.s.) (genotypes G1 and G3) based on a PCR-RFLP assay, and can further identify the remaining species based on a multiplex PCR assay. We demonstrate the applicability of the method to DNA extracted from parasitic cyst material of human and animal origin, preserved in ethanol or frozen. The method has been developed and validated at the European Union Reference Laboratory for Parasites (EURLP), according to the ISO/IE 17025.

Occurrence and Molecular Typing of Giardia psittaci in Parakeets in Germany-A Case Study.

A grey-hooded parakeet (Psilopsiagon aymara) and two budgerigars (Melopsittacus undulatus) from different owners presented with decreased activity, vomitus, and diarrhea. A microscopic examination of feces showed trophozoites of the protozoan flagellate Giardia. A commercial immunochromatographic dipstick test for Giardia sp. antigens confirmed the infection. These findings were assured by PCR of the small subunit ribosomal RNA (SSU rRNA) gene and coproantigen ELISA. Sequencing of PCR products of the SSU rRNA (292 bp) and ß-giardin genes (511 bp) identified Giardia psittaci as the species involved. Therefore, our results show that a GSA 65-based coproantigen ELISA, which was established for diagnosis of Giardia duodenalis is applicable for the detection of G. psittaci. A treatment with ronidazole was started. Additionally, fecal examination and dissection of the dead birds revealed coinfection with the fungal pathogen Macrorhabdus ornithogaster. One budgerigar survived and repeatedly tested negative after treatment with ronidazole. The described cases indicate that a single infection with G. psittaci has a good prognosis, whereas the prognosis is poor when coinfections occur, especially with M. ornithogaster.

Reporte de caso- Presentación y tipificación molecular de Giardia psittaci en periquitos en Alemania: Un estudio de caso. Un periquito catita aimará (Psilopsiagon aymara) y dos periquitos australianos (Melopsittacus undulatus) de diferentes propietarios presentaron actividad disminuida, vómito y diarrea. El examen microscópico de las heces mostró trofozoitos del protozoo flagelado Giardia. Una prueba de tira reactiva inmunocromatográfica comercial para antígenos de Giardia sp. confirmó la infección. Estos resultados fueron confirmados por PCR para el gene de ARN de la subunidad pequeña ribosomal (SSU rRNA) y por ELISA de coproantígeno. La secuenciación de los productos de PCR del ARNr de SSU (292 pb) y los genes de ß-giardina (511 pb) identificaron a Giardia psittaci como la especie involucrada. Por lo tanto, estos resultados muestran que el método de ELISA de coproantígeno basado en GSA 65, que se estableció para el diagnóstico de Giardia duodenalis, es aplicable para la detección de G. psittaci. Se inició un tratamiento con ronidazol. Además, el examen fecal y la disección de las aves muertas revelaron coinfección con el patógeno fúngico Macrorhabdus ornithogaster. Un periquito australiano sobrevivió y dio negativo repetidamente después del tratamiento con ronidazol. Los casos descritos indican que la infección única con G. psittaci tiene un buen pronóstico, mientras que el pronóstico es malo cuando ocurren coinfecciones, especialmente con M. ornithogaster. Abbreviations: GSA = Giardia-specific antigen; OD = optical density; rRNA = ribosomal ribonucleic acid; SSU = small subunit.

Multilocus genotyping of Cryptosporidium and Giardia in non-outbreak related cases of diarrhoea in human patients in Belgium.

Stool samples from Belgian patients suffering from abdominal pain and/or diarrhoea were examined for Cryptosporidium and Giardia. Cryptosporidium-positive samples were genotyped using the 70 kDa heat shock protein and the 60 kDa glycoprotein (GP60) genes: C. hominis was identified in 54.2% and C. parvum in 45.8% of the samples. Sequencing at the GP60 locus indicated that subgenotype IbA10G2 of C. hominis and subgenotype IIaA15G2R1 of C. parvum were the most prevalent, although several other subgenotypes were identified. For Giardia, sequencing at the beta-giardin, triose phosphate isomerase (TPI) and glutamate dehydrogenase (GDH) genes revealed assemblage B as the most prevalent (74.4%) in human patients. A high degree of heterogeneity was found, especially on the beta-giardin gene, and to a lesser extent on the GDH gene. Furthermore, using a novel species-specific PCR based on the TPI gene, mixed infections with both assemblage A and B were detected in a large number (32.4%) of human patients, which might have important epidemiological implications.

Distribution of Cryptosporidium parvum subtypes in calves in Germany.

Cryptosporidium DNA was extracted from 134 faecal specimens from pre-weaned calves from different German Federal States (age range, 3-15 days old), which tested positive for oocysts by microscopic analysis. The 18S rDNA gene and the oocyst wall protein gene (COWP) were used as targets for PCR and RFLP techniques. Cryptosporidium species were identified by using SspI, MboII and RsaI endonucleases for the digestion of 18S rDNA and COWP amplified fragments, respectively. In all samples, restriction patterns corresponding to Cryptosporidium parvum were identified, which is in agreement with abundant literature data indicating C. parvum as the most common species in pre-weaned calves. In order to estimate the genetic heterogeneity among C. parvum calf isolates, 53 samples chosen to represent different German Federal States were successfully subtyped by sequence analysis of the highly polymorphic 60-kDa glycoprotein gene. All isolates belonged to the allele IIa (with seven subtypes), with the exception of one isolate that belonged to the allele IId. Moreover, three novel subtypes of the allele family IIa have been found. This study confirms the utility of genotyping and subtyping tools in characterizing the transmission of Cryptosporidium spp. This is the first molecular epidemiological report about subtyping of Cryptosporidium bovine isolates in Germany.

Cryptosporidium and Giardia as foodborne zoonoses.

Cryptosporidium and Giardia are major causes of diarrhoeal disease in humans, worldwide and are major causes of protozoan waterborne diseases. Both Cryptosporidium and Giardia have life cycles which are suited to waterborne and foodborne transmission. There are 16 'valid'Cryptosporidium species and a further 33+ genotypes described. Parasites which infect humans belong to the Giardia duodenalis "type", and at least seven G. duodenalis assemblages are recognised. Cryptosporidium parvum is the major zoonotic Cryptosporidium species, while G. duodenalis assemblages A and B have been found in humans and most mammalian orders. In depth studies to determine the role of non-human hosts in the transmission of Cryptosporidium and Giardia to humans are required. The use of harmonised methodology and standardised and validated molecular markers, together with sampling strategies that provide sufficient information about all contributors to the environmental (oo)cyst pool that cause contamination of food and water, are recommended. Standardised methods for detecting (oo)cysts in water are available, as are optimised, validated methods for detecting Cryptosporidium in soft fruit and salad vegetables. These provide valuable data on (oo)cyst occurrence, and can be used for species and subspecies typing using appropriate molecular tools. Given the zoonotic potential of these organisms, epidemiological, source and disease tracking investigations involve multidisciplinary teams. Here, the role of the veterinarian is paramount, particularly in understanding the requirement for adopting comprehensive sampling strategies for analysing both sporadic and outbreak samples from all potential non-human contributors. Comprehensive sampling strategies increase our understanding of parasite population biology and structure and this knowledge can be used to determine what level of discrimination is required between isolates. Genetic exchange is frequent in C. parvum populations, leading to recombination between alleles at different loci, the generation of a very large number of different genotypes and a high level of resolution between isolates. In contrast, genetic exchange appears rare in Cryptosporidium hominis and populations are essentially clonal with far fewer combinations of alleles at different loci, resulting in a much lower resolution between isolates with many being of the same genotype. Clearly, more markers provide more resolution and high throughput sequencing of a variety of genes, as in multilocus sequence typing, is a way forward. Sub-genotyping tools offer increased discrimination, specificity and sensitivity, which can be exploited for investigating the epidemiology of disease, the role of asymptomatic carriers and contaminated fomites and for source and disease tracking for food and water contaminated with small numbers of (oo)cysts.

A novel Giardia duodenalis assemblage A subtype in fallow deer.

The molecular identification of species and genotypes of Giardia spp. infecting wild mammals represents the most reliable tool to understand the role played by these animals as reservoirs of cysts infectious for human and other animals. Of 139 fecal samples collected from fallow deer (Dama dama L.) hunted in a Natural Reserve of northern Italy, the prevalence of Giardia sp. was 11.5% (16 of 139 animals), and it was higher in fawns than in older animals. Fragments of the betagiardin and triose phosphate isomerase (tpi) genes were successfully polymerase chain reaction amplified and sequenced from 8 isolates. No sequence variation was observed between isolates at the 2 genetic loci. Sequence and phylogenetic analyses identified a Giardia duodenalis subtype that clusters with assemblage A isolates and that shows homologies of 98 and 97% at the beta-giardin and tpi loci, respectively, compared with the A1 subtype. Because the G. duodenalis subtype found in fecal samples of fallow deer has never been detected previously, its role as a pathogen for humans and domestic animals is unknown, but, considering its genetic distinctiveness, it is likely to be low.

Human cryptosporidiosis in Europe.

Cryptosporidium has emerged as a significant cause of diarrhoeal disease worldwide, with severe health consequences for very young, malnourished children living in endemic areas and for individuals with highly impaired T-cell functions. In Europe, as elsewhere, the burden of disease has been difficult to measure as a result of the lack of appropriate, standardized surveillance and monitoring systems. The recent occurrence of large water- and foodborne outbreaks in several EU countries, as well as the results of many surveys of human and animal cryptosporidiosis, indicate that this parasite is widespread. Specific subtypes of the zoonotic Cryptosporidium parvum and the anthroponotic C. hominis are responsible for the majority of human cases in Europe. No treatment is currently available to clear the infection, but recent progress in genetic engineering of the parasite, coupled with advances in genomics, have opened important avenues for future research. Here we explore the possible reasons for underascertainment of cryptosporidiosis and the importance of accurate diagnosis in clinical management, the epidemiology of human cryptosporidiosis and key messages from recent outbreaks to highlight important interventions and emerging public health issues.

Molecular epidemiology of human cryptosporidiosis.

Species within the genus Cryptosporidium are protozoan parasites that infect a wide range of vertebrates, and represent a significant cause of morbidity and mortality in those animals. In humans, cryptosporidiosis is a common cause of diarrhoeal disease with a global distribution. Unravelling the epidemiology of human infection has proven to be difficult, due to the existence of multiple transmission routes (person-to-person, animal-to-person, waterborne, foodborne and airborne transmission), and to the difficulties in identifying the different species using conventional criteria, such as oocyst morphology. The advent of molecular techniques has had a remarkable impact on the way the epidemiology of cryptosporidiosis can be studied. Molecular investigations have shown that the vast majority of human cases are caused by C. hominis and C. parvum. Interestingly, differences in geographical and temporal distribution, disease presentations and risk factors for infection have been identified for both C. hominis and C. parvum. Further, molecular analyses have revealed that other species, including C. meleagridis, C. felis, C. canis, C. suis, C. muris and two Cryptosporidium genotypes, can infect humans and may be linked to clinical disease, not only in immunocompromised but also in immunocompetent individuals.

Dientamoeba fragilis is more prevalent than Giardia duodenalis in children and adults attending a day care centre in Central Italy.

Giardia duodenalis is a well recognised enteropathogen, while Dientamoeba fragilis is rarely detected and consequently it is not recognised as an important human pathogen. In 2002-2003, a survey has been carried out on enteroparasites in faecal samples of outpatients attending a day care centre in the town of Perugia (Central Italy). To improve the detection level, at least three samples from each patient were collected at different days and within two hours from defecation. The coproparasitological examination has been carried out by direct microscopic examination, faecal concentration, and Giemsa and modified Ziehl-Nielsen stainings of faecal smears. The genotypes of Giardia duodenalis isolates were determined by PCR of the beta-giardin gene. Of 1,989 enrolled people (966 children, 1,023 adults), 165 persons (8.3%; 153 adults, 15.0%; 12 children, 1.2%), were positive for parasites, but only 1 12 adults (73.2% of those infected) and eight children (66.7% of those infected) harboured D. fragilis and G. duodenalis. Both the Assemblages A and B were detected in 18 G. duodenalis isolates examined at the beta-giardin gene. The higher prevalence of D. fragilis infections than that of G. duodenalis is probably related to the method used, a procedure, which is rarely followed in laboratories for the diagnosis of enteric parasites. These epidemiological data suggest that when faecal samples are examined after a period of time and without Giemsa staining, most D. fragilis infections goes undetected.

Cryptosporidium parvum Caused a Large Outbreak Linked to Frisée Salad in Finland, 2012.

Over 250 individuals fell ill in five outbreaks caused by Cryptosporidium parvum in Finland, October-November 2012. The cases were connected by lunch meals at restaurants in four different cities. In two outbreaks, the same C. parvumIIdA17G1 subtype was found in patients' stool samples which supports a single source of infection. Frisée salad was the only common food item served at the restaurants, and consumption of lunch salad containing the frisée salad was associated with the illness. Lunch customers who responded that they had eaten lunch salad were three times more likely to have become ill than those who had not answered whether they had eaten the salad or not (RR 2.66; 95% Cl 1.02-6.9, P-value <0.01). Cryptosporidiosis should be considered as a causal agent in long-lasting watery diarrhoea combined with abdominal cramps, and clinical samples should be tested for Cryptosporidium at the same time bacteria and viruses are tested. Measures to prevent contamination of 'ready-to-eat vegetables' with Cryptosporidium oocysts and methods to test frozen food samples should be developed.

Establishing the Cryptosporidium parvum karyotype by NotI and SfiI restriction analysis and Southern hybridization.

The molecular karyotype of the coccidian parasite Cryptosporidium parvum has proven difficult to study because chromosomes of similar sizes migrate together when submitted to pulsed-field gel electrophoresis (PFGE). In the present work, the karyotype was studied by restriction of chromosome-sized DNA with the rare-cutting enzymes NotI and SfiI, followed by PFGE separation of the restriction fragments and Southern hybridization. These experiments showed that the C. parvum karyotype is formed by eight chromosomes, ranging in size from approximately 0.95 to 1.45 million base pairs (Mbp), accounting for a genome size of 9.6Mbp. As a first step towards the construction of a physical map of the C. parvum genome, a total of 20 probes, including 16 genes and the ribosomal DNA (rDNA) sequence, was mapped to intact chromosomes and to their restriction fragments. In this way, all chromosomes, but one, were identified by specific markers. A comparison of mapping data of homologous genes from different species belonging to the phylum Apicomplexa showed differences in the distribution of rDNA sequences and in the chromosomal localization of alpha- and beta-tubulin genes. The variation in genome size among these parasites is also discussed.

Identification of the gene-richest bands in human chromosomes.

The human genome is a mosaic of isochores, long DNA segments which are compositionally homogeneous and which can be partitioned into five families, L1, L2, H1, H2 and H3, characterized by increasing GC levels and by increasing gene concentrations. Previous investigations showed that in situ hybridization with a DNA fraction derived from the GC-richest and gene-richest isochores of the H3 family produced the highest concentration of signals on 25 R(everse) bands that include the 22 most thermal-denaturation-resistant T(elomeric) bands, a subset of R bands. Using an improved protocol for in situ hybridization and cloned H3 isochore DNA, we have now shown (i) that the number of bands which are characterized by strong hybridization signals, and which are here called T or H3+, is 28; (ii) that 31 additional R bands, here called T'or H3* bands, also contain H3 isochores, although at a lower concentration than H3+ bands; and (iii) that the remaining R bands (about 140 out of 200, at a resolution of 400 bands), here called R" or H3- bands, do not contain any detectable H3 isochores. H3+ and H3* bands contain all the gene-richest isochores of the human genome. The existence of three distinct sets of R bands is further supported (i) by the different compositional features of genes located in them; (ii) by the very low gene density of chromosomes 13 and 18, in which all R bands are H3- bands; (iii) by the compositional map of a H3* band, Xq28; (iv) by the overwhelming presence of GC-rich and GC-poor long (> 50 kb) DNA sequences in H3+/H3* and in H3-/G bands, respectively; and (v) by the large degree of coincidence of H3+ and H3* bands with CpG island-positive bands. These observations have implications for our understanding of the causes of chromosome banding and provide a classification of chromosomal bands that is related to GC level (and to gene concentration).

Evolutionary changes in CpG and methylation levels in the genome of vertebrates.

We have analysed the levels of 5-methylcytosine (5mC) in DNAs from 42 vertebrates, and compiled, including data from literature, a table of genomic 5mC and GC levels (as well as the available c-values, i.e., the haploid genome sizes) of 87 species from all vertebrate classes. An analysis of the data indicates that (i) two positive correlations hold between the 5mC and GC levels of the genomes of fishes/amphibians and mammals/birds, respectively; (ii) the genomes of fishes and amphibians are, on average, about twice as methylated as those of mammals, birds and reptiles, this difference being unrelated to the amounts of repetitive DNA sequences; (iii) the 5mC and CpG observed/expected values show no overlap between the two groups of vertebrates and suggest the existence of two equilibria. The transition separating the two equilibria appears to have taken place at the time of appearance of reptiles. Its possible cause(s) and its implications are discussed.