Female H. lupulus cv Saaz hop (Osvald’s clone 72) and male H. lupulus Lib male (15181) (2n = 18 + XX/XY) were provided by the Hop Research Institute Co. Ltd. in Žatec (Czech Republic). Female and male H. japonicus (2n = 14 + XX/XY₁Y₂) and C. sativa cv Kompolti (2n = 18 + XX/XY) plants were grown from seeds obtained from W. Legutko (Poland) and SEMO a.s. (Czech Republic), respectively. All plants were grown in a greenhouse under controlled conditions (16h daylight/8h dark photoperiod) at the Department of Plant Developmental Genetics in Brno (Czech Republic). The sex of H. japonicus plants was determined based on floral morphology and chromosome number. In Cannabis sativa and H. lupulus, sex was determined using PCR with male-specific molecular markers: MADC2 for C. sativa (Mandolino et al., 1999) and OPJ9 for H. lupulus (Polley et al., 1997). Both male-specific markers were amplified using Taq polymerase (Top Bio) according to the manufacturer’s instructions (for primers, see Supplementary Table S1 ). PCR cycling conditions followed protocols described by Razumova et al. (2016) and Patzak et al. (2002). Male and female plants of each species were selected for further analysis based on the PCR results ( Supplementary Figure S1 ).

Genomic DNA was extracted from young leaves of male C. sativa, H. lupulus, and H. japonicus, using the NucleoSpin Plant II (Macherey-Nagel GmbH and Co. KG., Germany), following the manufacturer’s protocol. Libraries were prepared using the NEBNext^® Ultra™ II DNA Library Prep Kit. Genomic DNA from male plants of C. sativa and H. japonicus was low-coverage sequenced using an Illumina MiSeq sequencer, generating 300 bp paired-end reads at the Centre of Plant Structural and Functional Genomics (Olomouc, Czech Republic). Additionally, we used the library of H. lupulus described in Horáková et al. (2025).

The FastQC tool (available at http://www.bioinformatics.babraham.ac.uk/projects/fastqc) was used to assess the quality of sequencing reads. Furthermore, reads were pre-processed based on quality (Q30) with subsequent adaptor trimming, filtering out short or unpaired sequences, and all reads were trimmed to a uniform length of 200 bp using Trimmomatic 0.32 (Bolger et al., 2014). To identify repetitive DNA composition and tandem repeats, these datasets were analyzed using RepeatExplorer2 (Novák et al., 2013, 2020) and Tandem Repeat Analyzer (TAREAN) pipelines (Novák et al., 2017). The clustering of randomly selected 2 x 500–000 reads was performed by default setting.

The major satellites, including Cl5 (CS-1), Cl75 (HSR1), and Cl57 (HJSR) were amplified by PCR using specific primers ( Supplementary Table S1 ). Primers were designed in Geneious Prime (version 2023.1.1) based on the consensus sequences generated by the TAREAN pipeline. PCR amplification was performed using Q5 High-Fidelity DNA polymerase (M0491S; NEB) according to the manufacturer’s instructions. The cycling conditions were as follows: 95°C for 4 min followed by 35 cycles of 94°C for 30 s, 52 - 57°C for 35 s, 72°C for 30 s, and a final extension step at 72°C for 10 min. The annealing temperature was optimized for each primer pair. Selected units for each satellite ( Supplementary Figure S2 ) were extracted from agarose gel and purified using the QIAquick Gel Extraction Kit (QIAGEN GmbH - Hilden, Germany).

The purified fragments were individually cloned into the pJET1.2 plasmid vector using the CloneJET PCR Cloning Kit (K1231; ThermoFisher), following the manufacturer’s instructions. Ligation reactions were incubated overnight at 16°C and subsequently desalted. Competent E. coli cells were transformed by electroporation using a MicroPulser Electroporator (Bio-Rad). After 30-minute incubation, transformed cells were plated onto LB medium containing ampicillin (100 mg/L) and incubated overnight at 37°C. The next day, colony PCR was performed using pJET1.2-specific primers to identify positive clones. Individual bacterial colonies were used as templates. From positive PCR products, residual primers and PCR dNTPs were removed using ExoSAP reaction. PCR amplicons were sequenced in Macrogen (Amsterdam, Netherlands).

Ribosomal DNA (5S and 45S rDNA) was amplified by PCR using specific primers ( Supplementary Table S1 ). The primers were designed in Geneious Prime (version 2023.1.) based on the sequences from GenBank database (accession numbers: MN537579 (Easterling et al., 2020) and AF223066.1). PCR amplification was performed in a 20 µl reaction mixture containing 1x PCR buffer, 0.0001 M dNTPs, 0.0001 M of each primer, 0.5 U Taq polymerase (Top Bio), and 10–15 ng of template DNA. The PCR cycling conditions were as follows: initial denaturation at 95°C for 4 min, 35 cycles at 94°C for 30 s, 55°C for 35 s, 72°C for 30 s, followed by a final extension at 72°C for 10 min. The telomeric repeat sequence (Arabidopsis-type, TTTAGGG) was amplified without template DNA according to Ljdo et al. (1991). PCR cycling conditions were: first denaturation at 94°C for 1 min, 10 cycles of 94°C for 1 min, 55°C for 30 s, and 72°C for 1 min, followed by 30 cycles of 94°C for 1 min, 60°C for 30 s, 72°C for 90 s, and final extension step at 72°C for 5 min.

PCR products were separated on a 1% agarose gel with EtBr staining and purified using the QIAquick PCR Purification Kit (28104; QIAGEN) according to the manufacturer’s instructions. Purified PCR products (1µg) of each satellite and species and 5S, 45S rDNA, and telomere obtained as described above, were labeled using Nick Translation Labelling kits (Jena Bioscience, Germany) with Atto488 NT (PP-305L-488), Atto550 NT (PP-305L-550), and Cy5 (PP-305L-647N), following the manufacturer’s instructions. Labeling reactions were incubated at 15°C for 90 min. The labeled products were analyzed on a 1% agarose gel with EtBr staining. Reactions of well-labeled DNA probes were stopped by adding 0.5M EDTA and incubating at 85°C. The labeled DNA probes were then used directly in the hybridization mixture for FISH.

Mitotic chromosomes were prepared from young leaves (2–5 mm in length) of C. sativa, H. lupulus, and H. japonicus as described in Horáková et al. (2025). FISH was performed on mitotic metaphase chromosomes according to Sacchi et al. (2024) under two stringency conditions (77% and 68%; Supplementary Table S2 ). FISH with 77% stringency was used to confirm the subtelomeric localization of CS-1 (Cl5) in C. sativa, HSR1 (Cl75) in H. lupulus, HJSR (Cl57) in H. japonicus, and 45S rDNA. Low-stringency FISH (68%) was used to determine the interspecific localization of satellites. Chromosomes were counterstained with 4´,6´-diamidino-2-phenylindole (DAPI) in Vectashield Antifade Mounting Medium. Images were captured using an Olympus AX70 epifluorescence microscope equipped with a CCD camera and processed using Adobe Photoshop. FISH experiments were performed in triplicate, with at least ten metaphases analyzed per experiment.

The sequences were aligned in MAFFT in two rounds. At first, the adjusted direction option was used to ensure the correct orientation of sequences. The alignment was further improved in MAFFT using an iterative refinement method incorporating local pairwise alignment (mafft-linsi). The phylogenetic tree was constructed using maximum-likelihood approach using IQ-TREE. The substitution model (HKY+F+G4) was chosen in IQ-TREE based on the BIC criterion. The support values were obtained using the ultrafast bootstrap method (with 1000 replicates). To obtain sequence logos for each satellite-species combination, we mapped the reads from a given species to the reference, which was represented by a 60% consensus sequence prepared from sequenced clones of PCR products obtained in the species in which that satellite was originally described (for primers, see Supplementary Table S1 ). In several satellite-species combinations, low-coverage sequencing did not allow sufficient coverage of the reference, so the sequence logo was prepared based on the cloned PCR products. DNA reads were trimmed using Trimmomatic (Bolger et al., 2014) and mapped to the reference using BWA-MEM (Li, 2013). SAM files were converted to fasta alignments using a simple Python script (sam2fasta.py; https://sourceforge.net/projects/sam2fasta/). Sequence logos were generated from fasta alignments using a locally installed version of the Weblogo 3 program (Crooks et al., 2004) with sequence-type set to ‘dna’.

Repeatome analysis revealed major satellites candidates in Cannabis and Humulus species. Specifically, Cl5 (CS-1) in C. sativa, Cl75 (HSR1) in H. lupulus, and Cl57 (HJSR) in H. japonicus. We extracted consensus sequence for each repeat and their abundance in the genomes based on sequencing data ( Supplementary Table S3 ). These major satellites - Cl5 (CS-1), Cl75 (HSR1), and Cl57 (HJSR) show high sequence similarity to previously described sequences in GenBank: CS-1 (JX402748.2; Divashuk et al., 2014), HSR1 (Humulus subtelomeric repeat 1, GU831574.1; Divashuk et al., 2011), HJSR (Humulus japonicus subtelomeric repeat, GU831573.1; Alexandrov et al., 2012). For clarity and consistency, we refer to these repeats throughout the text by their established GenBank names: CS-1, HSR1, and HJSR, respectively. The monomer unit lengths for these satellites range from 370 bp for CS-1, 380 bp for HJSR, to 383 bp for HSR1. Sequence comparison revealed relatively higher sequence similarity between HSR1 and CS-1 (58.82%) and HSR1 and HJSR (65.05%; Supplementary Figure S3 ).

To analyze the interspecific distribution of major satellites, CS-1, HSR1, and HJSR, we performed FISH on metaphase chromosomes of C. sativa, H. lupulus, and H. japonicus ( Figure 1 ; Table 1 ). Additionally, Arabidopsis-type telomeric (TTTAGGG) and 45S rDNA probes were used to label chromosome ends and enable the identification of previously described chromosomes ( Figure 1A ). Telomeric signals are present at the terminal regions of all chromosomes, without any interstitial telomeric signals ( Figure 1A ; Supplementary Figure S4A ). The distribution of 45S rDNA varies among studied species, being localized on chromosome 9 in C. sativa, chromosome 8 in H. lupulus, and on two pairs of autosomes, specifically on chromosomes 5 and 7 in H. japonicus ( Figure 1A ). Chromosome nomenclature in H. lupulus is based on the pseudomolecule assembly by Akagi et al. (2025) with chromosomes numbered according to their size, from largest to smallest. Whereas in C. sativa and H. japonicus, it follows the systems established by Kim et al. (2008); Alexandrov et al. (2012), and Divashuk et al. (2014).

We observed chromosome-specific differences in the localization of all three studied satellites.

In Cannabis sativa, CS-1 is predominantly localized in the subtelomeric regions of both arms of all autosomes, except for the q-arm of chromosome 4 and the p-arm of chromosome 9, the latter of which possess 45S rDNA. Notably, the Y chromosome lacks CS-1 on the q-arm ( Figure 1B ). This distinct distribution of CS-1, combined with the Y chromosome size and DAPI banding pattern, enables its precise identification ( Figures 1B, C ). However, the identification of the X chromosome is still difficult, based solely on CS-1 distribution and factors mentioned above ( Figures 1B, C ). In Humulus lupulus, the CS-1 satellite is localized in the (peri)centromeric region of chromosome 2 ( Figure 1B ) while 5S rDNA is simultaneously detected in its subtelomeric region ( Supplementary Figure S5 ). Although the CS-1 satellite was amplified with the designed primers in H. japonicus ( Supplementary Figure S2 ), we did not observe any clear signals on metaphase chromosomes of either sex ( Figure 1B and Supplementary Figure S4B ), presumably resulting from its low abundance on chromosomes and in the genome.

The Humulus lupulus-specific repeat, HSR1 satellite, is localized in the (peri)centromeric regions of ten chromosomes in males ( Figure 1C ) and eleven chromosomes in females C. sativa ( Supplementary Figure S4C ). This structural chromosomal heterozygosity may indicate a hybrid origin of the female C. sativa cv Kompolti or suggest the presence of HSR1 on one of the X chromosomes. Simultaneous hybridization of 45S rDNA, HSR1, and CS-1 satellites shows that HSR1 is absent from chromosomes 4, 9, and Y ( Supplementary Figure S6 ). In contrast, the HSR1 satellite is localized in the subtelomeric regions of almost all autosomes of H. lupulus, except for the p-arm of chromosome 8, which carries 45S rDNA, the p-arm of X chromosome, and the q-arm of chromosome Y ( Figure 1C ). Notably, the X chromosome exhibits a strong HSR1 signal in the pericentromeric region of its p-arm. In Humulus japonicus, the HSR1 probe is distributed in the subtelomeric regions of almost all chromosomes ( Figure 1C ), showing similarities to the HJSR probe ( Figure 1D ). However, HSR loci on one chromosome pair and the chromosome Y₂ are absent or underrepresented.

In Cannabis sativa, HJSR is found mainly in subtelomeric regions, although, some chromosomes lack the HJSR signal ( Figure 1D ). The distribution of the HJSR satellite in H. lupulus was identical to that of the HSR1 satellite, including its localization on the sex chromosomes ( Figure 1D ; Supplementary Figure S4D ). In Humulus japonicus, the HJSR satellite is present in subtelomeric regions, except for one arm of X chromosome, p-arm of chromosome 5, and both arms of chromosome 7, which carry 45S rDNA. No HJSR signal was observed on the Y₂ chromosome ( Figure 1D ). The distribution of 45S rDNA across all analyzed species aligns with previous studies. The distribution of CS-1, HSR1 and HJSR satellites in female plants are shown in Supplementary Figure S4 . We observed no sex-specific differences in the chromosomal distribution of these satellites between studied accessions ( Figure 1 ; Supplementary Figure S4 ). However, the number of HSR1 signals varied in C. sativa ( Supplementary Figures S4C , S6 ). The overall comparative distribution of CS-1, HSR1, and HJSR satellite further suggests intergenomic changes during the species divergence ( Figure 2 ).

Using the primers sets designed for each major satellite across the three species, we obtained consistent amplification pattern for HSR1 and HJSR, with monomer unit length of approximately 380 bp. In contrast, the CS-1 satellite exhibited a species-specific ladder pattern in H. japonicus. The expected monomer band of 370 bp for CS-1 was obtained exclusively in C. sativa and H. lupulus ( Supplementary Figure S2 ). Selected units for each satellite and species were cloned and sequenced. We analyzed a total of 35 CS-1 sequences (23 from C. sativa, 5 from H. lupulus, and 7 from H. japonicus), 46 HSR1 sequences (15 from C. sativa, 14 from H. lupulus, and 17 from H. japonicus), and 36 HJSR sequences (11 from C. sativa, 14 from H. lupulus, and 11 from H. japonicus). A comparison of satellite sequences cloned from each species supported low interspecific variability and their classification into one supercluster family, in line with physical localization on metaphase chromosomes ( Supplementary Figure S7 ). To determine the relationships among these satellites in studied species, we conducted a phylogenetic analysis of individual sequence clusters ( Figure 3 ; Supplementary Figures S8 , S9 ). The CS-1 amplified in C. sativa and H. japonicus grouped together, while those amplified in H. lupulus formed a separate clade ( Figure 3 ). Although the sequences from C. sativa and H. japonicus are closely related, no visible CS-1 signal was detected on the chromosomes of H. japonicus ( Figure 1B ), suggesting a low enrichment of these satellite units insufficient for detection by FISH. CS-1 was clearly localized to the subtelomeric regions of C. sativa chromosomes, in contrast to its (peri)centromeric localization on chromosome 2 in H. lupulus. The HSR1 and HJSR sequences tended to cluster together, with no unique species-specific clades identified between the species ( Supplementary Figures S8 , S9 ), indicating low divergence among individual subunits. This is consistent with their shared localization in subtelomeric regions. In Cannabis sativa, HSR1 was found in pericentromeric regions on five autosome pairs ( Figure 1C ). Among the three satellite arrays, only the HJSR satellite exhibited a conserved subtelomeric localization across all three species ( Figure 1D ), lacking clear species-specific clade and suggesting common origin for all three species.