Differential genome evolution and speciation of Coix lacryma-jobi L. and Coix aquatica Roxb. hybrid guangxi revealed by repetitive sequence analysis and fine karyotyping

Abstract

Background

Coix, Sorghum and Zea are closely related plant genera in the subtribe Maydeae. Coix comprises 9–11 species with different ploidy levels (2n = 10, 20, 30, and 40). The exclusively cultivated C. lacryma-jobi L. (2n = 20) is widely used in East and Southeast Asia for food and medicinal applications. Three fertile cytotypes (2n = 10, 20, and 40) have been reported for C. aquatica Roxb. One sterile cytotype (2n = 30) closely related to C. aquatica has been recently found in Guangxi of China. This putative hybrid has been named C. aquatica HG (Hybrid Guangxi). The genome composition and the evolutionary history of C. lacryma-jobi and C. aquatica HG are largely unclear.

Results

About 76% of the genome of C. lacryma-jobi and 73% of the genome of C. aquatica HG are repetitive DNA sequences as shown by low coverage genome sequencing followed by similarity-based cluster analysis. In addition, long terminal repeat (LTR) retrotransposable elements are dominant repetitive sequences in these two genomes, and the proportions of many repetitive sequences in whole genome varied greatly between the two species, indicating evolutionary divergence of them. We also found that a novel 102 bp variant of centromeric satellite repeat CentX and two other satellites only appeared in C. aquatica HG. The results from FISH analysis with repeat probe cocktails and the data from chromosomes pairing in meiosis metaphase showed that C. lacryma-jobi is likely a diploidized paleotetraploid species and C. aquatica HG is possibly a recently formed hybrid. Furthermore, C. lacryma-jobi and C. aquatica HG shared more co-existing repeat families and higher sequence similarity with Sorghum than with Zea.

Conclusions

The composition and abundance of repetitive sequences are divergent between the genomes of C. lacryma-jobi and C. aquatica HG. The results from fine karyotyping analysis and chromosome pairing suggested diploidization of C. lacryma-jobi during evolution and C. aquatica HG is a recently formed hybrid. The genome-wide comparison of repetitive sequences indicated that the repeats in Coix were more similar to those in Sorghum than to those in Zea, which is consistent with the phylogenetic relationship reported by previous work.

Electronic supplementary material

The online version of this article (doi:10.1186/1471-2164-15-1025) contains supplementary material, which is available to authorized users.     

Construction of bacterial artificial chromosome libraries from the parasitic nematode Brugia malayi and physical mapping of the genome of its Wolbachia endosymbiont

The parasitic nematode, Brugia malayi, causes lymphatic filariasis in humans, which in severe cases leads to the condition known as elephantiasis. The parasite contains an endosymbiotic alpha-proteobacterium of the genus Wolbachia that is required for normal worm development and fecundity and is also implicated in the pathology associated with infections by these filarial nematodes. Bacterial artificial chromosome libraries were constructed from B. malayi DNA and provide over 11-fold coverage of the nematode genome. Wolbachia genomic fragments were simultaneously cloned into the libraries giving over 5-fold coverage of the 1.1 Mb bacterial genome. A physical framework for the Wolbachia genome was developed by construction of a plasmid library enriched for Wolbachia DNA as a source of sequences to hybridise to high-density bacterial artificial chromosome colony filters. Bacterial artificial chromosome end sequencing provided additional Wolbachia probe sequences to facilitate assembly of a contig that spanned the entire genome. The Wolbachia sequences provided a marker approximately every 10 kb. Four rare-cutting restriction endonucleases were used to restriction map the genome to a resolution of approximately 60 kb and demonstrate concordance between the bacterial artificial chromosome clones and native Wolbachia genomic DNA. Comparison of Wolbachia sequences to public databases using BLAST algorithms under stringent conditions allowed confident prediction of 69 Wolbachia peptide functions and two rRNA genes. Comparison to closely related complete genomes revealed that while most sequences had orthologs in the genome of the Wolbachia endosymbiont from Drosophila melanogaster, there was no evidence for long-range synteny. Rather, there were a few cases of short-range conservation of gene order extending over regions of less than 10 kb. The molecular scaffold produced for the genome of the Wolbachia from B. malayi forms the basis of a genomic sequencing effort for this bacterium, circumventing the difficult challenge of purifying sufficient endosymbiont DNA from a tropical parasite for a whole genome shotgun sequencing strategy.     

Laboratory evolution and multi-platform genome re-sequencing of the cellulolytic actinobacterium Thermobifida fusca

Biological utilization of cellulose is a complex process involving the coordinated expression of different cellulases, often in a synergistic manner. One possible means of inducing an organism-level change in cellulase activity is to use laboratory adaptive evolution. In this study, evolved strains of the cellulolytic actinobacterium, Thermobifida fusca, were generated for two different scenarios: continuous exposure to cellobiose (strain muC) or alternating exposure to cellobiose and glucose (strain muS). These environmental conditions produced a phenotype specialized for growth on cellobiose (muC) and an adaptable, generalist phenotype (muS). Characterization of cellular phenotypes and whole genome re-sequencing were conducted for both the muC and muS strains. Phenotypically, the muC strain showed decreased cell yield over the course of evolution concurrent with decreased cellulase activity, increased intracellular ATP concentrations, and higher end-product secretions. The muS strain increased its cell yield for growth on glucose and exhibited a more generalist phenotype with higher cellulase activity and growth capabilities on different substrates. Whole genome re-sequencing identified 48 errors in the reference genome and 18 and 14 point mutations in the muC and muS strains, respectively. Among these mutations, the site mutation of Tfu_1867 was found to contribute the specialist phenotype and the site mutation of Tfu_0423 was found to contribute the generalist phenotype. By conducting and characterizing evolution experiments on Thermobifida fusca, we were able to show that evolutionary changes balance ATP energetic considerations with cellulase activity. Increased cellulase activity is achieved in stress environments (switching carbon sources), otherwise cellulase activity is minimized to conserve ATP.     

Characterization and potential evolutionary impact of transposable elements in the genome of Cochliobolus heterostrophus

Background

Cochliobolus heterostrophus is a dothideomycete that causes Southern Corn Leaf Blight disease. There are two races, race O and race T that differ by the absence (race O) and presence (race T) of ~ 1.2-Mb of DNA encoding genes responsible for the production of T-toxin, which makes race T much more virulent than race O. The presence of repetitive elements in fungal genomes is considered to be an important source of genetic variability between different species.

Results

A detailed analysis of class I and II TEs identified in the near complete genome sequence of race O was performed. In total in race O, 12 new families of transposons were identified. In silico evidence of recent activity was found for many of the transposons and analyses of expressed sequence tags (ESTs) demonstrated that these elements were actively transcribed. Various potentially active TEs were found near coding regions and may modify the expression and structure of these genes by acting as ectopic recombination sites. Transposons were found on scaffolds carrying polyketide synthase encoding genes, responsible for production of T-toxin in race T. Strong evidence of ectopic recombination was found, demonstrating that TEs can play an important role in the modulation of genome architecture of this species. The Repeat Induced Point mutation (RIP) silencing mechanism was shown to have high specificity in C. heterostrophus, acting only on transposons near coding regions.

Conclusions

New families of transposons were identified. In C. heterostrophus, the RIP silencing mechanism is efficient and selective. The co-localization of effector genes and TEs, therefore, exposes those genes to high rates of point mutations. This may accelerate the rate of evolution of these genes, providing a potential advantage for the host. Additionally, it was shown that ectopic recombination promoted by TEs appears to be the major event in the genome reorganization of this species and that a large number of elements are still potentially active. So, this study provides information about the potential impact of TEs on the evolution of C. heterostrophus.

Electronic supplementary material

The online version of this article (doi:10.1186/1471-2164-15-536) contains supplementary material, which is available to authorized users.     

Comparative genome analysis of pathogenic and non-pathogenic Clavibacter strains reveals adaptations to their lifestyle

Background

The genus Clavibacter harbors economically important plant pathogens infecting agricultural crops such as potato and tomato. Although the vast majority of Clavibacter strains are pathogenic, there is an increasing number of non-pathogenic isolates reported. Non-pathogenic Clavibacter strains isolated from tomato seeds are particularly problematic because they affect the current detection and identification tests for Clavibacter michiganensis subsp. michiganensis (Cmm), which is regulated with a zero tolerance in tomato seed. Their misidentification as pathogenic Cmm hampers a clear judgment on the seed quality and health.

Results

To get more insight in the genetic features linked to the lifestyle of these bacteria, a whole-genome sequence of the tomato seed-borne non-pathogenic Clavibacter LMG 26808 was determined. To gain a better understanding of the molecular determinants of pathogenicity, the genome sequence of LMG 26808 was compared with that of the pathogenic Cmm strain (NCPPB 382). The comparative analysis revealed that LMG 26808 does not contain plasmids pCM1 and pCM2 and also lacks the majority of important virulence factors described so far for pathogenic Cmm. This explains its apparent non-pathogenic nature in tomato plants. Moreover, the genome analysis of LMG 26808 detected sequences from a plasmid originating from a member of Enterobacteriaceae/Klebsiella relative. Genes received that way and coding for antibiotic resistance may provide a competitive advantage for survival of LMG 26808 in its ecological niche. Genetically, LMG 26808 was the most similar to the pathogenic Cmm NCPPB 382 but contained more mobile genetic elements. The genome of this non-pathogenic Clavibacter strain contained also a high number of transporters and regulatory genes.

Conclusions

The genome sequence of the non-pathogenic Clavibacter strain LMG 26808 and the comparative analyses with other pathogenic Clavibacter strains provided a better understanding of the genetic bases of virulence and adaptation mechanisms present in the genus Clavibacter.

Electronic supplementary material

The online version of this article (doi:10.1186/1471-2164-15-392) contains supplementary material, which is available to authorized users.     

Characterisation of the double genome structure of modern sugarcane cultivars (Saccharum spp.) by molecular cytogenetics

Cultivated sugarcane clones (Saccharum spp., 2n=100 to 130) are derived from complex interspecific hybridizations between the speciesS. officinarum andS. spontaneum. Using comparative genomic DNA in situ hybridization, we demonstrated that it is possible to distinguish the chromosomes contributed by these two species in an interspecific F1 hybrid and a cultivated clone, R570. In the interspecific F1 studied, we observed n+n transmission of the parental chromosomes instead of the peculiar 2n+n transmission usually described in such crosses. Among the chromosomes of cultivar R570 (2n=107–115) about 10% were identified as originating fromS. spontaneum and about 10% were identified as recombinant chromosomes between the two speciesS. officinarum andS. spontaneum. This demonstrated for the first time the occurrence of recombination between the chromosomes of these two species. The rDNA sites were located by in situ hybridization in these two species and the cultivar R570. This supported different basic chromosome numbers and chromosome structural differences between the two species and provided a first bridge between physical and genetical mapping in sugarcane.     

The repetitive component of the A genome of peanut (Arachis hypogaea) and its role in remodelling intergenic sequence space since its evolutionary divergence from the B genome

Background and Aims

Peanut (Arachis hypogaea) is an allotetraploid (AABB-type genome) of recent origin, with a genome of about 2·8 Gb and a high repetitive content. This study reports an analysis of the repetitive component of the peanut A genome using bacterial artificial chromosome (BAC) clones from A. duranensis, the most probable A genome donor, and the probable consequences of the activity of these elements since the divergence of the peanut A and B genomes.

Methods

The repetitive content of the A genome was analysed by using A. duranensis BAC clones as probes for fluorescence in situ hybridization (BAC-FISH), and by sequencing and characterization of 12 genomic regions. For the analysis of the evolutionary dynamics, two A genome regions are compared with their B genome homeologues.

Key Results

BAC-FISH using 27 A. duranensis BAC clones as probes gave dispersed and repetitive DNA characteristic signals, predominantly in interstitial regions of the peanut A chromosomes. The sequences of 14 BAC clones showed complete and truncated copies of ten abundant long terminal repeat (LTR) retrotransposons, characterized here. Almost all dateable transposition events occurred <3·5 million years ago, the estimated date of the divergence of A and B genomes. The most abundant retrotransposon is Feral, apparently parasitic on the retrotransposon FIDEL, followed by Pipa, also non-autonomous and probably parasitic on a retrotransposon we named Pipoka. The comparison of the A and B genome homeologous regions showed conserved segments of high sequence identity, punctuated by predominantly indel regions without significant similarity.

Conclusions

A substantial proportion of the highly repetitive component of the peanut A genome appears to be accounted for by relatively few LTR retrotransposons and their truncated copies or solo LTRs. The most abundant of the retrotransposons are non-autonomous. The activity of these retrotransposons has been a very significant driver of genome evolution since the evolutionary divergence of the A and B genomes.     

Acetone utilization by sulfate-reducing bacteria: draft genome sequence of Desulfococcus biacutus and a proteomic survey of acetone-inducible proteins

Background

The sulfate-reducing bacterium Desulfococcus biacutus is able to utilize acetone for growth by an inducible degradation pathway that involves a novel activation reaction for acetone with CO as a co-substrate. The mechanism, enzyme(s) and gene(s) involved in this acetone activation reaction are of great interest because they represent a novel and yet undefined type of activation reaction under strictly anoxic conditions.

Results

In this study, a draft genome sequence of D. biacutus was established. Sequencing, assembly and annotation resulted in 159 contigs with 5,242,029 base pairs and 4773 predicted genes; 4708 were predicted protein-encoding genes, and 3520 of these had a functional prediction. Proteins and genes were identified that are specifically induced during growth with acetone. A thiamine diphosphate-requiring enzyme appeared to be highly induced during growth with acetone and is probably involved in the activation reaction. Moreover, a coenzyme B₁₂- dependent enzyme and proteins that are involved in redox reactions were also induced during growth with acetone.

Conclusions

We present for the first time the genome of a sulfate reducer that is able to grow with acetone. The genome information of this organism represents an important tool for the elucidation of a novel reaction mechanism that is employed by a sulfate reducer in acetone activation.

Electronic supplementary material

The online version of this article (doi:10.1186/1471-2164-15-584) contains supplementary material, which is available to authorized users.     

Analysis of papaya BAC end sequences reveals first insights into the organization of a fruit tree genome

Papaya (Carica papaya L.) is a major tree fruit crop of tropical and subtropical regions with an estimated genome size of 372 Mbp. We present the analysis of 4.7% of the papaya genome based on BAC end sequences (BESs) representing 17 million high-quality bases. Microsatellites discovered in 5,452 BESs and flanking primer sequences are available to papaya breeding programs at . Sixteen percent of BESs contain plant repeat elements, the vast majority (83.3%) of which are class I retrotransposons. Several novel papaya-specific repeats were identified. Approximately 19.1% of the BESs have homology to Arabidopsis cDNA. Increasing numbers of completely sequenced plant genomes and BES projects enable novel approaches to comparative plant genomics. Paired BESs of Carica, Arabidopsis, Populus, Brassica and Lycopersicon were mapped onto the completed genomes of Arabidopsis and Populus. In general the level of microsynteny was highest between closely related organisms. However, papaya revealed a higher degree of apparent synteny with the more distantly related poplar than with the more closely related Arabidopsis. This, as well as significant colinearity observed between peach and poplar genome sequences, support recent observations of frequent genome rearrangements in the Arabidopsis lineage and suggest that the poplar genome sequence may be more useful for elucidating the papaya and other rosid genomes. These insights will play a critical role in selecting species and sequencing strategies that will optimally represent crop genomes in sequence databases.Electronic Supplementary Material Supplementary material is available for this article at and is accessible for authorized users.Chun Wan J. Lai and Qingyi Yu have contributed equally to this work.     

Insights into the evolutionary process of genome degradation

Studies of noncoding and pseudogene sequence diversity, particularly in Rickettsia, have begun to reveal the basic principles of genome degradation in microorganisms. Increasingly, studies of genes and genomes suggest that there has been an extensive amount of horizontal gene transfer among microorganisms. As this inflow of genetic material does not seem generally to have resulted in genome size expansions, however, degenerative processes must be at the very least as widespread as horizontal gene transfer. The basic principles of gene degradation and elimination that are being explored in Rickettsia are likely to be of major importance for our understanding of how microbial genomes evolve.     

RiTE database: a resource database for genus-wide rice genomics and evolutionary biology

Background

Comparative evolutionary analysis of whole genomes requires not only accurate annotation of gene space, but also proper annotation of the repetitive fraction which is often the largest component of most if not all genomes larger than 50 kb in size.

Results

Here we present the Rice TE database (RiTE-db) - a genus-wide collection of transposable elements and repeated sequences across 11 diploid species of the genus Oryza and the closely-related out-group Leersia perrieri. The database consists of more than 170,000 entries divided into three main types: (i) a classified and curated set of publicly-available repeated sequences, (ii) a set of consensus assemblies of highly-repetitive sequences obtained from genome sequencing surveys of 12 species; and (iii) a set of full-length TEs, identified and extracted from 12 whole genome assemblies.

Conclusions

This is the first report of a repeat dataset that spans the majority of repeat variability within an entire genus, and one that includes complete elements as well as unassembled repeats. The database allows sequence browsing, downloading, and similarity searches. Because of the strategy adopted, the RiTE-db opens a new path to unprecedented direct comparative studies that span the entire nuclear repeat content of 15 million years of Oryza diversity.

Electronic supplementary material

The online version of this article (doi:10.1186/s12864-015-1762-3) contains supplementary material, which is available to authorized users.     

The evolutionary landscape of the Mycobacterium tuberculosis genome

Mycobacterium tuberculosis is one of the most deadly human pathogens. The major mechanism for the adaptations of M. tuberculosis is nucleotide substitution. Previous studies have relied on the nonsynonymous-to-synonymous substitution rate (d_N/d_S) ratio as a measurement of selective constraint based on the assumed selective neutrality of synonymous substitutions. However, this assumption has been shown to be untrue in many cases. In this study, we used the substitution rate in intergenic regions (d_i) of the M. tuberculosis genome as the neutral reference, and conducted a genome-wide profiling for d_i, d_S, and the rate of insertions/deletions (indel rate) as compared with the genome of M. canettii using a 50 kb sliding window. We demonstrate significant variations in all of the three evolutionary measurements across the M. tuberculosis genome, even for regions in close vicinity. Furthermore, we identified a total of 233 genes with their d_S deviating significantly from d_i within the same window. Interestingly, d_S also varies significantly in some of the windows, indicating drastic changes in mutation rate and/or selection pressure within relatively short distances in the M. tuberculosis genome. Importantly, our results indicate that selection on synonymous substitutions is common in the M. tuberculosis genome. Therefore, the d_N/d_S ratio test must be applied carefully for measuring selection pressure on M. tuberculosis genes.     

Estimation of the Spontaneous Mutation Rate per Nucleotide Site in a Drosophila melanogaster Full-Sib Family

We employed deep genome sequencing of two parents and 12 of their offspring to estimate the mutation rate per site per generation in a full-sib family of Drosophila melanogaster recently sampled from a natural population. Sites that were homozygous for the same allele in the parents and heterozygous in one or more offspring were categorized as candidate mutations and subjected to detailed analysis. In 1.23 × 10⁹ callable sites from 12 individuals, we confirmed six single nucleotide mutations. We estimated the false negative rate in the experiment by generating synthetic mutations using the empirical distributions of numbers of nonreference bases at heterozygous sites in the offspring. The proportion of synthetic mutations at callable sites that we failed to detect was <1%, implying that the false negative rate was extremely low. Our estimate of the point mutation rate is 2.8 × 10⁻⁹ (95% confidence interval = 1.0 × 10⁻⁹ − 6.1 × 10⁻⁹) per site per generation, which is at the low end of the range of previous estimates, and suggests an effective population size for the species of ∼1.4 × 10⁶. At one site, point mutations were present in two individuals, indicating that there had been a premeiotic mutation cluster, although surprisingly one individual had a G→A transition and the other a G→T transversion, possibly associated with error-prone mismatch repair. We also detected three short deletion mutations and no insertions, giving a deletion mutation rate of 1.2 × 10⁻⁹ (95% confidence interval = 0.7 × 10⁻⁹ − 11 × 10⁻⁹).     

A comprehensive genetic map of sugarcane that provides enhanced map coverage and integrates high-throughput Diversity Array Technology (DArT) markers

Background

Sugarcane genetic mapping has lagged behind other crops due to its complex autopolyploid genome structure. Modern sugarcane cultivars have from 110-120 chromosomes and are in general interspecific hybrids between two species with different basic chromosome numbers: Saccharum officinarum (2n = 80) with a basic chromosome number of 10 and S. spontaneum (2n = 40-128) with a basic chromosome number of 8. The first maps that were constructed utilised the single dose (SD) markers generated using RFLP, more recent maps generated using AFLP and SSRs provided at most 60% genome coverage. Diversity Array Technology (DArT) markers are high throughput allowing greater numbers of markers to be generated.

Results

Progeny from a cross between a sugarcane variety Q165 and a S. officinarum accession IJ76-514 were used to generate 2467 SD markers. A genetic map of Q165 was generated containing 2267 markers, These markers formed 160 linkage groups (LGs) of which 147 could be placed using allelic information into the eight basic homology groups (HGs) of sugarcane. The HGs contained from 13 to 23 LGs and from 204 to 475 markers with a total map length of 9774.4 cM and an average density of one marker every 4.3 cM. Each homology group contained on average 280 markers of which 43% were DArT markers 31% AFLP, 16% SSRs and 6% SNP markers. The multi-allelic SSR and SNP markers were used to place the LGs into HGs.

Conclusions

The DArT array has allowed us to generate and map a larger number of markers than ever before and consequently to map a larger portion of the sugarcane genome. This larger number of markers has enabled 92% of the LGs to be placed into the 8 HGs that represent the basic chromosome number of the ancestral species, S. spontaneum. There were two HGs (HG2 and 8) that contained larger numbers of LGs verifying the alignment of two sets of S. officinarum chromosomes with one set of S. spontaneum chromosomes and explaining the difference in basic chromosome number between the two ancestral species. There was also evidence of more complex structural differences between the two ancestral species.

Electronic supplementary material

The online version of this article (doi:10.1186/1471-2164-15-152) contains supplementary material, which is available to authorized users.     

Full genome SNP-based phylogenetic analysis reveals the origin and global spread of Brucella melitensis

Background

Brucellosis is an important zoonotic disease that affects both humans and animals. We sequenced the full genome and characterised the genetic diversity of two Brucella melitensis isolates from Malaysia and the Philippines. In addition, we performed a comparative whole-genome single nucleotide polymorphism (SNP) analysis of B. melitensis strains collected from around the world, to investigate the potential origin and the history of the global spread of B. melitensis.

Results

Single sequencing runs of each genome resulted in draft genome sequences of MY1483/09 and Phil1136/12, which covered 99.85% and 99.92% of the complete genome sequences, respectively. The B. melitensis genome sequences, and two B. abortus strains used as the outgroup strains, yielded a total of 13,728 SNP sites. Phylogenetic analysis using whole-genome SNPs and geographical distribution of the isolates revealed spatial clustering of the B. melitensis isolates into five genotypes, I, II, III, IV and V. The Mediterranean strains, identified as genotype I, occupied the basal node of the phylogenetic tree, suggesting that B. melitensis may have originated from the Mediterranean regions. All of the Asian B. melitensis strains clustered into genotype II with the SEA strains, including the two isolates sequenced in this study, forming a distinct clade denoted here as genotype IId. Genotypes III, IV and V of B. melitensis demonstrated a restricted geographical distribution, with genotype III representing the African lineage, genotype IV representing the European lineage and genotype V representing the American lineage.

Conclusion

We showed that SNPs retrieved from the B. melitensis draft full genomes were sufficient to resolve the interspecies relationships between B. melitensis strains and to discriminate between the vaccine and endemic strains. Phylogeographic reconstruction of the history of B. melitensis global spread at a finer scale by using whole-genome SNP analyses supported the origin of all B. melitensis strains from the Mediterranean region. The possible global distribution of B. melitensis following the ancient trade routes was also consistent with whole-genome SNP phylogeny. The whole genome SNP phylogenetics analysis, hence is a powerful tool for intraspecies discrimination of closely related species.

Electronic supplementary material

The online version of this article (doi:10.1186/s12864-015-1294-x) contains supplementary material, which is available to authorized users.     

Floral development of Hydrocera and Impatiens reveals evolutionary trends in the most early diverged lineages of the Balsaminaceae

Background and Aims

Balsaminaceae consist of two genera, the monospecific Hydrocera and its species-rich sister Impatiens. Although both genera are seemingly rather similar in overall appearance, they differ in ecology, distribution range, habitat preference and morphology. Because morphological support for the current molecular phylogenetic hypothesis of Impatiens is low, a developmental study is necessary in order to obtain better insights into the evolutionary history of the family. Therefore, the floral development of H. triflora and I. omeiana was investigated, representing the most early-diverged lineage of Impatiens, and the observations were compared with the literature.

Methods

Flowers at all developmental stages were examined using scanning electron microscopy and light microscopy.

Key results

In Hydrocera, two whorls of five free perianth primordia develop into a less zygomorphic perianth compared with its sister genus. The androecial cap originates from five individual stamen primordia. Post-genital fusion of the upper parts of the filaments result in a filament ring below the anthers. The anthers fuse forming connivent anther-like units. The gynoecium of Hydrocera is pentamerous; it is largely synascidiate in early development. Only then is a symplicate zone formed resulting in style and stigmas. In I. omeiana, the perianth is formed as in Hydrocera. Five individual stamen primordia develop into five stamens, of which the upper part of the filaments converge with each other. The gynoecium of I. omeiana is tetramerous; it appears annular in early development.

Conclusions

Comparison of the present results with developmental data from the literature confirms the perianth morphocline hypothesis in which a congenital fusion of the parts of the perianth results in a shift from pentasepalous to trisepalous flowers. In addition, the development of the androecial cap and the gynoecium follows several distinct ontogenetic sequences within the family.