Theor Appl Genet (2016) 129:1985–2001 DOI 10.1007/s00122-016-2754-7 ORIGINAL ARTICLE Allelic diversity of S‑RNase alleles in diploid potato species Daniel K. Dzidzienyo1,2,3 · Glenn J. Bryan2 · Gail Wilde2 · Timothy P. Robbins1 Received: 22 March 2016 / Accepted: 15 July 2016 / Published online: 6 August 2016 © The Author(s) 2016. This article is published with open access at Springerlink.com Abstract S-RNase sequences were obtained from pistil RNA by Key message The S‑ribonuclease sequences of 16 RT-PCR or 3′RACE (Rapid Amplification of cDNA Ends) S‑alleles derived from diploid types of Solanum are pre‑ using a degenerate primer. Full-length sequences were sented. A phylogenetic analysis and partial phenotypic obtained for two alleles by 5′RACE. Database searches analysis support the conclusion that these are functional with these sequences identified 16 S-RNases in total, all S‑alleles. of which are novel. The sequence analysis revealed all the Abstract S-Ribonucleases (S-RNases) control the pistil expected features of functional S-RNases. Phylogenetic specificity of the self-incompatibility (SI) response in the analysis with selected published S-RNase and S-like- genus Solanum and several other members of the Solan- RNase sequences from the Solanaceae revealed extensive aceae. The nucleotide sequences of S-RNases correspond- trans-generic evolution of the S-RNases and a clear dis- ing to a large number of S-alleles or S-haplotypes have tinction from S-like-RNases. Pollination tests were used to been characterised. However, surprisingly, few S-RNase confirm the self-incompatibility status and cross-compati- sequences are available for potato species. The identifica- bility relationships of the S. okadae accessions. All the S. tion of new S-alleles in diploid potato species is desirable okadae accessions were found to be self-incompatible as as these stocks are important sources of traits such as biotic expected with crosses amongst them exhibiting both cross- and abiotic resistance. S-RNase sequences are reported compatibility and semi-compatibility consistent with the here from three distinct diploid types of potato: cultivated S-genotypes determined from the S-RNase sequence data. Solanum tuberosum Group Phureja, S. tuberosum Group The progeny analysis of four semi-compatible crosses Stenotomum, and the wild species Solanum okadae. Partial examined by allele-specific PCR provided further confir- mation that these are functional S-RNases. Communicated by Y. Xue. Electronic supplementary material The online version of this Introduction article (doi:10.1007/s00122-016-2754-7) contains supplementary material, which is available to authorized users. The majority of flowering plants are hermaphrodite, and Timothy P. Robbins their reproductive organs are located in close proximity, a * tim.robbins@nottingham.ac.uk feature which might have imposed self-pollination and, sub- sequently, self-fertilisation on angiosperms. However, due 1 Plant and Crop Sciences Division, School of Biosciences, to the deleterious effect of self-fertilisation and inbreeding, University of Nottingham, Sutton Bonington Campus, Loughborough LE12 5RD, UK flowering plants have evolved several strategies to avoid 2 this, the most widespread of which is self-incompatibil- Cell and Molecular Sciences, The James Hutton Institute, Invergowrie, Dundee DD2 5DA, Scotland, UK ity (de Nettancourt 1997, 2001). Self-incompatibility is a 3 prezygotic barrier that enables the pistil, to distinguish self- Present Address: Biotechnology Centre, College of Basic and Applied Sciences, University of Ghana, P.O. Box LG 68, pollen from non-self-pollen, leading to the arrest of self-pol- Legon-Accra, Ghana len and thus blocking self-fertilisation (Kao and McCubbin 1 3 1 986 Theor Appl Genet (2016) 129:1985–2001 1996; Takayama and Isogai 2005). Depending on the and characterisation of these additional S-RNases in potato genetic control of self-incompatibility in the pollen, plants give an indication of the diversity of S-alleles in these taxa, are classified as having either a sporophytic or gametophytic thereby contributing significantly to the existing knowledge mechanism (de Nettancourt 1977; Hiscock and McInnis of the diversity of S-RNases in the Solanaceae family gener- 2003). The evolution of mechanisms to prevent inbreeding ally, and the genus Solanum subsection Petota, in particular. in flowering plants is partly responsible for their evolution- ary success, thereby making them one of the most success- ful terrestrial groups of plants (Silva and Goring 2001). Materials and methods Gametophytic self-incompatibility (GSI) represents the most prevalent form of self-incompatibility found in Plant materials more than 60 flowering plant families. The Solanaceae, Rosaceae, Plantaginaceae, Leguminoceae, Onagraceae, We use the classification of Dodds (1962) for describing Papaveraceae and Poaceae are amongst the plant families primitive diploid cultivated potato germplasm. This scheme exhibiting this form of self-incompatibility (de Nettancourt places the landrace germplasm studied here into two groups 1977). The extensive study of GSI at the molecular level within S. tuberosum, Group Phureja and Stenotomum . has revealed that it operates by two different mechanisms to Genebank accessions and breeding clones of Solanum achieve self-pollen recognition and rejection. One of these tuberosum Group Phureja, Solanum tuberosum Group Sten- is the stylar ribonuclease (S-RNase) mechanism which has otomum, and Solanum okadae (Table 1) are maintained at been initially identified and characterised in members of The James Hutton Institute (JHI, Invergowrie, Scotland) as the Solanaceae, and later in the Rosaceae, Plantaginaceae part of the Commonwealth Potato Collection (CPC), and and most recently in the Rubiaceae (Kao and McCubbin also duplicate clones were maintained during the course of 1996; Nowak et al. 2011; Asquini et al. 2011). A distinct these studies at the University of Nottingham. The plants mechanism involving a pollen receptor is found in the were grown under controlled glasshouse conditions of 16 h Papaveraceae, in particular, Papaver rhoeas (Franklin-Tong photoperiod and 25 °C day/18 °C night temperatures dur- and Franklin 2003; Wheeler et al. 2009). The pistil speci- ing winter and natural day lengths during summer with ficity of plant families exhibiting the S-RNase-based GSI supplementary lighting where necessary. system is controlled by polymorphic glycoproteins which are ribonucleases (S-RNases) and that have confirmed Controlled pollinations ribonuclease activity (Bredemeijer and Blass 1981; Ander- son et al. 1986; McClure et al. 1989). Transgenic experi- Controlled self- and cross-pollinations were carried out ments in both petunia and tobacco have established that the to confirm the SI status of the potato stocks and also to S-RNase is the sole determinant of pistil specificity (Lee et al. 1994; Murfett et al. 1994). Table 1 Diploid potato species and clone designations used in this Most of the diploid tuber-bearing Solanum species have study a gametophytic system of self-incompatibility which is controlled by a single multi-allelic S-locus (Pushkarnath Potato species Plant ID 1942; Pandey 1962; Cipar et al. 1964). Although phyloge- S. okadae OKA 7129-1 netic studies of S-RNase diversity have been reported for OKA 7129-3 several solanaceous species (Richman et al. 1996; Igic and OKA 7129-5 Kohn 2001), this is yet to be conducted comprehensively in OKA 7129-7 potato (Solanum subsection Petota) partly due to the rela- OKA 7129-9 tive paucity of S-RNase sequences available (see Table 4). S. stenotomum STN 4679 Our aim in the present study was to identify and charac- STN 4679-68 terise additional S-RNase sequences in both cultivated STN 4679-72 and wild diploid potatoes based on breeding objectives as STN 4711-61 well as for the potential development of specialised genetic STN 4741 resources, such as populations of recombinant inbred lines STN 4741-119 (RILs). We have characterised S-alleles both phenotypically, STN 4741-135 using pollination tests, and genotypically, using an RT-PCR STN 4786-80 approach with degenerate primers. These led to the iden- S. phureja DB 226-70 tification of new S-alleles in the primitive cultivated Sola- DB 337-37 num tuberosum Groups Phureja and Stenotomum, and the DB 536-102 wild diploid species Solanum okadae. The identification 1 3 Theor Appl Genet (2016) 129:1985–2001 1987 determine their possible compatibility relationships. The Reverse Transcription (RT)‑PCR reaction: 3′RACE anthers of potatoes are hollow tubes (anther cones) that open by small apical pores. To collect pollen from these Reverse Transcription (RT) reactions were carried out using tubular anthers, a buzzer was used to vibrate the anther an oligo-dT primer (NotI d(T)18) (5′-AACTGGAAGAA cone and the pollen collected into a microcentrifuge tube. TTCGCGGCCGCAGGAA(T)18-3′) consisting of a 27-bp The pollen was then deposited onto the stigma using a paint anchor part with a NotI restriction site (included at the 5′ brush after the stigma had reached maturity. Crosses were end) and 18 thymidine (T) nucleotides at the 3′end. First- scored approximately 4 weeks after pollination as either: strand cDNA synthesis was achieved using Omniscript® self-incompatible if there were no berries (or no seed set) Reverse Transcriptase (Qiagen). The reaction comprised formed or fully self-compatible if berries (or seed set) 1X RT Buffer, 0.5 mM dNTPs, 1 µM NotI d(T)18 primers, could be seen following pollination. 10 U of RNase inhibitor (RNaseOut Recombinant Ribo- nuclease Inhibitor, Invitrogen), 4 U of Omniscript Reverse DNA extraction Transcriptase and approximately 1–2 µg of total RNA tem- plate denatured at 65 °C for 5 min in RNase-free water. The Genomic DNA was extracted for allele-specific PCR geno- final RT reaction mixture was incubated at 37 °C for 1 h for typing from progenies obtained from S. okadae crosses that cDNA synthesis. segregated for different S-alleles. Two leaf discs weighing A degenerate primer, SolC2-F1.3 (5′-TTTACNRTN- approximately 10–100 mg of young leaves were processed CATGGNCTNTGGCC-3′) was designed based on the C2 for DNA extraction using the DNeasy Plant Mini kit (Qia- conserved domain of Solanaceae S-RNases (Ioerger et al. gen, UK). 1991). This was used to amplify partial S-RNase sequences from pistil RNA using the RT-PCR-based 3′ RACE (Rapid Allele‑specific PCR genotyping Amplification of Complementary DNA Ends) technique. The degenerate primer (SolC2-F1.3) was used together The genotyping primers were designed to allow allele- with another primer (NotI-anchor primer) (5′-AACTG- specific amplification from plants segregating for known GAAGAATTCGCGG-3′) which has the same recognition S-RNases. Specific primers used were as follows: So1-RNase sequence as the anchor part of the oligo-dT primer (NotI (So1-F (5′ GGATAAGGAGGGATCACAGC 3′) and So1- d(T)18) used in the first-strand cDNA synthesis. RT-PCR R (5′ TGTTGGCTTTGTATTTTGTAGCA 3′), So2-RNase amplification was performed in a 25-µl total reaction mix (So2-F (5′ TGCGAGTCCGAAGACAAGTA 3′) and So2- comprising a 2.5-µl aliquot of RT reaction (cDNA reac- R (5′ AAGGGAAAGAAAACGGAAGC 3′)), So4-RNase tion), 20 mM Tris–HCl (pH 8.3), 50 mM KCl, 1.5 mM (So4-F (5′ TCGATTGGAGTTCTGCACTG 3′) and So4-R MgCl2, 0.2 mM dNTPs (Bioline), 0.4 µM of SolC2-F1.3 (5′TTTCATCGCATGTGTTACCC3′)) and So5-RNase (So5- (forward), 0.2 µM NotI-anchor (reverse) primers and 2 U F (5′ TGGTCGAAAGGAACAACCTT 3′) and So5-R (5′ of Taq DNA polymerase (Bioline). Amplification was per- TTCCAACCTGGTCATTCAAAG 3′)). The primers were formed as described previously with an annealing tempera- designed to have optimal melting temperature of 60 °C. ture of 55 °C and a final extension of 5 min. PCR reactions were performed in a 25-μl reaction volume comprising 1X PCR buffer, 3 mM MgCl2, 0.2 mM dNTPs Reverse transcription (RT)‑PCR reaction: 5′RACE (Bioline), 0.4 μM of forward and reverse primers each and 2 U of Taq DNA polymerase (Bioline, London, UK). PCR Gene-specific primers were designed from the partial amplification was performed in a PTC-200 Thermal Cycler sequences isolated from the 3′RACE cloning for the identified (MJ Research, Watertown MA, USA) under the following S-RNases and used for full-length cDNA cloning using the cycling conditions: an initial 3-min denaturation at 94 °C, 5′RACE-PCR technique. Three gene-specific primers (GSPs): followed by 35 cycles of 30 s at 94 °C, 30 s of annealing at So2-GSP1 (5′-ATTATACCATGATTTCGGAGAGC-3′), So2- a temperature depending on the Tm (melting temperature) of GSP2 (5′-AGATCGATACTACACGTTCCATG-3′) and So2- the primer, 1 min at 72 °C and a final extension of 7 min at GSP3 (5′-CAGTGATACTCCAGTTGTTTGC-3′) were used 72 °C. for cloning full-length So2-RNase. For the Ss2-RNase, Ss2- GSP1 (5′-AGAAGTAATACCATTCTTTCCGAG-3′), Ss2- RNA extraction GSP2 (5′-AACACGTTCCATGCTTAATG-3′) and Ss2- GSP3 (5′-CCAGATGTATTCCAGAGCTTC-3′) gene-spe- RNA was extracted from pistil tissues using the RNeasy cific primers were used. First-strand cDNA synthesis and Plant Mini kit (Qiagen). Approximately 10–100 mg of tis- 5′RACE-PCR technique was carried out using the 5′RACE sue, pre-chilled in liquid nitrogen (N2), was ground to a System for Rapid Amplification of cDNA Ends, Version 2.0 fine powder and processed for the RNA extraction. (Invitrogen) following the manufacturers’ recommendations. 1 3 1988 Theor Appl Genet (2016) 129:1985–2001 PCR products were obtained using AAP (abridged anchor basis of the alignment using ClustalW method as imple- primer) or AUAP (abridged universal anchor primer) primers mented in MegAlign™ (DNASTAR). Phylogenetic analy- (supplied in the kit) and the gene-specific primers (GSPs) for sis of the cloned S-RNase sequences together with selected each allele. solanaceous S-RNases and S-like-RNases retrieved from the EBI database (http://srs.ebi.ac.uk) was performed using the Cloning into pCR®2.1 vector neighbour-joining tree method as implemented in MEGA5 software (Tamura et al. 2011). The evolutionary distances RACE-PCR products were cloned using the TA cloning kit used to infer the phylogenetic tree were calculated using a (Invitrogen). The concentration of the PCR product (less Poisson correction method (Zuckerkandl and Pauling 1965) than 1 day old) needed to ligate with 50 ng (20 fmoles) as implemented in MEGA5. To show how well the topology of pCR®2.1 vector was determined, and a 10-μl ligation of the tree was supported, bootstrap analysis (Felsenstein reaction was set up with 1 µl of 10× ligation buffer, 2 µl 1985) was performed using 1000 replicates. Three Antir- of 25 ng pCR®2.1 vector and 1 µl of 4.0 Weiss units T4 rhinum S-RNase amino acid sequences (accession num- DNA ligase. The ligation mixture was incubated at 14 °C bers X96464, X96465 and X96466) were also included in overnight, transformed into One Shot competent cells the phylogenetic tree as an ‘out-group’ to root the tree, and (TOP10F’ Invitrogen) and plated on LB plates containing in other cases, T2-RNase of Aspergillus oryzae (accession 40 mg/ml of X-Gal, 100 mM of IPTG and either 50 µg/ml number CAA43400.1) was used to root the phylogenetic of kanamycin or 100 µg/ml of ampicillin. tree. Positions which contain gaps and missing data were eliminated from the phylogenetic analysis. All S-RNase Sequencing of plasmid inserts sequence data have been submitted to the NCBI database with accession numbers listed in Supplementary Table S4. Colonies were screened using M13 universal primers, and plasmid DNA was extracted from putative positive clones using the QIAprep Spin Miniprep kit (Qiagen). The Results sequence of the inserted DNA was determined using for- ward and reverse M13 universal primers at either the Qia- Confirming the SI status and the compatibility gen Genomic Services/Sequencing Services (Qiagen) or relationships of potato germplasm The James Hutton Institute sequencing facility (JHI, Inver- gowrie, Scotland). To confirm the SI status and the compatibility relation- ships among some of the potato germplasm used in this Analysis of S‑RNase sequences study, pollinations were performed following a diallel cross design. The results from clones within the S. okadae acces- The alignment, comparison and phylogenetic analysis of sion OKA7129 (Table 2) showed that all self-pollinations the putative S-RNase sequences were performed using the resulted in no berries set (self-incompatible). From the sequence data analysis tools listed below. DNASTAR soft- crosses, it was observed that two of the OKA7129 clones ware (DNASTAR Inc., Madison, USA) was used to deduce are likely to be harbouring the same pair of S-alleles based the amino acid sequence of the cloned putative S-RNases. on incompatible cross-pollinations. Crosses between OKA1 The deduced amino acid sequences were then subjected to and OKA3 consistently failed to produce any berries in the NCBI database BLAST (http://blast.ncbi.nlm.nih.gov) searches, and the accession numbers of the best hits were noted. The accession numbers were then entered into the Table 2 Pollination data for S. okadae EBI database (http://srs.ebi.ac.uk) for the retrieval of the Female parent Male parent amino acid sequence of four of these best hits where nec- OKA1 OKA3 OKA5 OKA7 OKA9 essary. The cloned S-RNases were aligned together with one of the selected S-RNase sequences from the database OKA 1 * * 106 125 125 to act as a reference gene. Alignments were performed by OKA 3 * * 50 71 82 ClustalW method (Thompson et al. 1994) using BioEdit OKA 5 19 60 * 37 87 (http://www.mbio.ncsu.edu/BioEdit/bioedit) and also MegA- OKA 7 52 50 106 * 90 lign™ as implemented in DNASTAR using the default set- OKA 9 113 115 100 79 * tings and edited manually. The percentage similarities of the deduced amino acid sequence of the cloned S-RNases were The figures in the table are the average numbers of seed set per berry (seeds/berry) for the various crosses. The asterisks (*) represent the also determined by calculating the sequence distances using failed (incompatible) crosses. Accession names in Table 1 are abbre- the neighbour-joining method (Saitou and Nei 1987) on the viated e.g. OKA1 = OKA7129-1 1 3 Theor Appl Genet (2016) 129:1985–2001 1989 HP +RT +RT -RT -RT and used in combination with an oligo-dT NotI-anchor primer. This enabled the amplification by RT-PCR of puta- tive pistil S-RNases from S. tuberosum Group Stenotomum, S. tuberosum Group Phureja and S. okadae resulting in the expected amplicon size of ~ 850 bp for all three taxa (data ~850 bp not shown). An example of such an RT-PCR for S. okadae OKA9 is given in Fig. 1. The first-round RT-PCR gave a discrete size of PCR product for almost all accessions tested and, hence, were used directly for cloning without the need for a nested PCR step. To enable the selection of colonies having the expected insert size, colony PCR using M13 universal forward and reverse primers was performed. Fig. 1 Representative example of RT-PCR of pistil S-RNases in An example of the colony PCR results obtained with acces- Solanum okadae. RT-PCR of OKA 9 diploid potato pistils using a sion OKA9 is shown in Supplementary Fig. 1. Three such consensus S-RNase C2 domain primer. Lane +RT represents dupli- cate samples with RT (Reverse Transcriptase), −RT represents dupli- colonies were selected for each S-RNase cloning experi- cate samples without RT. Lane HP represents Hyperladder II size ment, and plasmid DNA was purified for sequencing. marker (Bioline) Sequencing and alignment of the 3′RACE products either direction indicating that these clones share the same The sequencing results produced a total of 17 putative pair of S-alleles. Crosses involving the other parents were S-RNases from genotypes of the three taxa studied following observed to set seed in either direction and could represent database searches. These S-RNases were provisionally called either a compatible or a semi-compatible cross (see sup- So1 to So5 for those from S. okadae, Ss1 to Ss10 for those from plementary list S1 for the full details of the crosses). The Group Stenotomum and Sp1 and Sp2 for those from Group compatibility relationship of other potato plants used in this Phureja (Table 3). The S-RNase designation is based on the study (Group Stenotomum and Group Phureja) could not be order in which they were identified in each of the species. checked using pollination tests due to the limited number of Results from the NCBI database searches revealed that our flowers available or lack of flowering time synchronisation. cloned S-RNases shared sequence similarity with S-RNases from other solanaceous species (data not shown). When the RT‑PCR cloning of pistil S‑RNases S-RNases were initially cloned, only four S-RNases from Solanum chacoense could be found that were similar (S2, A degenerate primer was designed for the C2 conserved S3, S11 and S14). However, during the later stages of this domain based on the alignment of solanaceous S-RNases research, additional sequences for potato species (Solanum Table 3 Summary of the Species Plant ID Cloned S-allele Deduced S-genotype sixteen deduced S-genotypes of Solanum accessions used in Allele Freq* Allele Freq* the study Solanum okadae OKA 1 So1 3 So2 4 So1So2 Solanum okadae OKA 3 So1 2 So2 3 So1So2 Solanum okadae OKA 5 So1 13 So5 2 So1So5 Solanum okadae OKA 7 So3 10 – – So3So? Solanum okadae OKA 9 So1 7 So4 2 So1So4 Solanum stenotomum STN 4679 Ss1 3 Ss9 2 Ss1Ss9 Solanum stenotomum STN 4679-72 Ss5 4 – – Ss5Ss? Solanum stenotomum STN 4711-61 Ss2 3 Ss3 2 Ss2Ss3 Solanum stenotomum STN 4741 Ss4 8 Ss10 1 Ss4Ss10 Solanum stenotomum STN 4741-135 Ss8 4 – – Ss8Ss? Solanum stenotomum STN 4786-80 Ss6 3 Ss7 4 Ss6Ss7 Solanum phureja DB 226 Sp1 4 – – Sp1Sp? Solanum phureja DB 337 Sp2 4 – – Sp2Sp? Solanum phureja DB 536 Sp2 4 – – Sp2Sp? * Freq cloned frequency of each allele observed from the total number of colonies examined 1 3 1990 Theor Appl Genet (2016) 129:1985–2001 Table 4 S-RNases for potato Potato species S-allele Accession # References retrieved from NCBI database S. chacoense S2 X56896.1 Xu et al. (1990) S. chacoense S3 X56897.1 Xu et al. (1990) S. chacoense S11 S69589.1/L36464.1 Saba-el-Leil et al. (1994) S. chacoense S12 AF176533.1/AF191732.1 Qi et al. (2001) S. chacoense S13 L36667.1 Despres et al. (1994) S. chacoense S14 AF232304 O’Brien et al. (2002) S. chacoense S16 DQ007316 Marcellan et al. (2006) S. stenotomum S3 HM446648 Kear (2010), unpublished S. stenotomum 60_A AEN02425.1 Kear and Malinski (2010), unpublished S. stenotomum 60_B AEN02426.1 Kear and Malinski (2010), unpublished S. stenotomum 60_C AEN02427.1 Kear and Malinski (2010), unpublished S. stenotomum 60_D AEN02428.1 Kear and Malinski (2010), unpublished S. stenotomum 60_E AEN02429.1 Kear and Malinski (2010), unpublished S. stenotomum 47_A AEN02423.1 Kear and Malinski (2010), unpublished S. stenotomum 47_D AEN02424.1 Kear and Malinski (2010), unpublished S. phureja S36 HM446649 Kear (2010), unpublished S. tuberosum S2 X62727 Kaufmann et al. (1991) subsection Petota) were submitted to NCBI and found to be accession. For example the So1-RNase was found to be similar to the cloned alleles. Database mining for S-RNases present in four of the S. okadae genotypes from accession published in potato to date (December 2015) has revealed OKA7129, i.e. OKA1, OKA3, OKA5 and OKA9. Simi- a total of 17 S-RNase sequences (Table 4). However, there larly, the So2-RNase was also found to be present in both were no exact matches of our cloned S-RNases to the previ- OKA1 and OKA3. Interestingly, Ss9 and Sp1 were found to ously published alleles indicating that they are all novel. be present in both the Stenotomum accession STN 4679 Some of the plants from which we have isolated and Phureja clone DB226(70), respectively. This allele S-RNases harbour the expected two S-alleles, whilst only was initially cloned from Stenotomum and named Ss9 and one allele could be found in others, notably the Phureja was later cloned from Phureja and named Sp1 (see Table 3). group genotypes (Table 3). Attempts were made to identify However, sequence comparison showed that, these two the second allele of those in which only one could be found S-RNases shared exactly the same deduced amino acid by further rounds of colony PCR which proved success- sequence (data not shown) and, hence, were provisionally ful for some but not all genotypes studied. For instance, a renamed Ss9/Sp1 (or Ss9p1) to represent both alleles although second allele was identified for OKA9, but examination of they could theoretically represent two functionally distinct a similar total number of clones examined for OKA7 was alleles. This observation is not unexpected as Stenotomum not successful in identifying a second allele. In an attempt and Phureja are closely related evolutionarily and are con- to identify the second allele of these accessions, a nega- sidered members of the same Solanum tuberosum species tive screening approach was taken. Allele-specific primers group (Spooner et al. 2014). Allowing for this identity, the were designed and used to screen all colonies revealed by total number of novel S-RNase sequences identified in this M13 primers to have inserts with the expected sizes. Those study is 16 comprising 5 from S. okadae and 11 from S. that did not amplify with the allele-specific primers were tuberosum Groups Stenotomum and Phureja. thought to be candidates for the putative second allele of The alignment of the deduced partial amino acid the accessions screened and were sent for sequencing. sequence of the 16 novel putative Solanum S-RNases is However, after screening large numbers of colonies (typi- shown in Fig. 2. Three of the conserved domains (C3–C5) cally ~50–100), only a few of these putative second alleles and the two hypervariable domains (HVa and HVb) are could be identified. The previously cloned allele from highlighted and are part of the primary structural features of which the primers were designed was present in almost all solanaceous S-RNases as defined by Ioerger et al. (1991). the colonies screened for the accessions with just one allele One of the two catalytic histidine residues (His) known (data not shown), suggesting that the allele-specific primers to be involved in the ribonuclease activity of S-RNases is did not amplify the original allele efficiently in every case. indicated in the alignment of the C3 region. Six out of the From the sequencing results summarised in Table 3, eight conserved cysteine residues found in selected sola- it is apparent that some alleles occur in more than one naceous S-RNases (Ioerger et al. 1991) are also indicated. 1 3 Theor Appl Genet (2016) 129:1985–2001 1991 Fig. 2 Alignment of the deduced amino acid sequence of the 16 ties among the 16 sequences with reference to the S11-RNase. Gaps novel S-RNases cloned from S. okadae, S. phureja and S. stenotomum in the alignment are indicated by (−). The conserved histidine residue with one published S-RNase, S11 from Solanum chacoense (Acc. No: in the C3 region involved in the ribonuclease activity of S-RNases S69589.1). The hypervariable regions (HVa and HVb) and conserved is marked with an arrow head. Six conserved cysteine residues are regions (C3–C5) are boxed. The dots in the alignment indicate identi- marked with asterisks (*) under the alignment These cysteine residues are important for determining the just after the C5 region). This particular cysteine residue is tertiary structure of S-RNases (Ishimizu et al. 1996). The also missing in four other solanaceous S-RNases published remaining two were not shown because they are located in in public databases (accession numbers AAB26702.1, regions which are outside of this alignment (i.e. between CAA53666.1 and BAC00933.1 and AAV69976.1). C1 and C2). However, one of the cysteine residues shown The percentage amino acid sequence similarity among in this alignment was not perfectly conserved. Thus, all six- the 16 S-RNases ranged from 32.9 to 94.5 % (Table 5). teen cloned S-RNases contained all six cysteine residues This low level of sequence similarity is consistent with expected with the exception of the So2-RNase which lacks a the high level of polymorphism known to exist between residue at the 3′ end (i.e. the last conserved cysteine located S-RNase alleles in the Solanaceae (Ioerger et al. 1990). 1 3 1992 Theor Appl Genet (2016) 129:1985–2001 Table 5 Percentage amino acid S-allele S S S S S S S S S S S S S S S S similarity of the sixteen cloned o1 o2 o3 o4 o5 s1 s2 s3 s4 s5 s6 s7 s8 s9p1 s10 p2 S-RNases So1 – 32.9 42.2 52.3 73.9 43.7 33.5 50.6 41.9 72.0 38.3 41.1 75.8 76.4 51.3 39.6 So2 – 47.5 38.0 34.8 44.6 48.4 34.4 45.3 37.4 44.1 42.0 35.5 35.5 34.4 45.3 So3 – 42.3 44.8 44.2 37.7 39.2 43.8 44.8 42.1 43.6 44.2 44.2 39.9 44.7 So4 – 53.5 40.3 38.0 70.8 40.5 54.8 44.6 38.3 56.8 56.1 72.7 44.6 So5 – 39.7 33.5 50.6 39.4 84.7 43.5 40.4 86.0 86.6 50.0 44.2 Ss1 – 42.9 35.9 39.5 39.1 37.2 74.5 39.7 38.4 36.5 37.2 Ss2 – 33.1 43.5 34.8 36.2 45.5 33.5 33.5 35.0 38.8 Ss3 – 40.6 53.8 38.4 34.0 53.2 52.6 94.5 40.9 Ss4 – 40.6 43.8 40.1 41.9 41.3 42.5 48.8 Ss5 – 43.5 37.7 82.8 84.1 53.2 42.9 Ss6 – 37.8 41.6 42.2 40.3 80.4 Ss7 – 39.7 38.4 36.5 39.1 Ss8 – 87.9 53.8 42.9 Ss9p1 – 53.8 44.2 Ss10 – 42.1 Sp2 – Species of origin for S-alleles are abbreviated as follows: So = S. okadae, Ss = S. stenotomum, Sp = S. phureja From the table, it could be observed that amino acid simi- HP +RT1 -RT1 W +RT2 - RT2 W larity within a species could be as low as 32.9 % for the S. okadae S-RNases (So1 vs So2), 33.1 % for the Stenoto- mum S-RNases (Ss2 vs Ss3) and 44.2 % for the two Phureja S-RNases. The isoelectric point (pI) value was calculated for the fully and partially deduced amino acid sequences using DNASTAR software, and the values revealed that all ~400 bp the S-RNases are basic proteins and have a pI value in the range of 8.6–9.6 (data not shown). 5′RACE of selected S‑RNases An attempt was made to clone the full-length sequence of Fig. 3 5′RACE-PCR of So2-RNase from OKA 1. Lane HP rep- resents Hyperladder II (bioline), +RT1 represents sample with RT two selected alleles, i.e. So2-RNase from S. okadae and Ss2- (reverse transcriptase) from 1st round PCR, −RT1 represents sample RNase from Stenotomum, using 5′ RACE. Three antisense without RT from 1st round PCR, +RT2 represents sample with RT gene-specific primers (GSPs) were designed from the par- from 2nd round (nested) PCR, sample −RT2 represents sample with- tial sequence obtained through the 3′RACE cloning and out RT from 2nd round PCR, W represents water control used in 5′RACE to enable the full-length S-RNase gene sequence to be determined. The RT-PCR products gave the expected amplicon in both cases as indicated for the So2- The amino acid sequence alignment of the two full-length RNase in Fig. 3. Following transformation, three colonies cloned S-RNases (So2 from S. okadae and Ss2 from Group having the expected insert size as revealed by colony PCR Stenotomum) with three of the published full-length S. (Supplementary Fig. 2) were selected for plasmid DNA chacoense S-RNase sequences from the database is shown extraction and sequencing of inserts. in Fig. 4. The two hypervariable regions (HVa and HVb) and the five conserved domains (C1, C2, C3, C4 and C5) Sequencing and alignment of the 5′RACE products found in all solanaceous S-RNases are indicated. The two conserved histidine residues which are located on the C2 The 5′RACE cloning allowed the addition of the C1 and and C3 domains and known to be involved in the ribonucle- C2 domains to the 5′ region of the two selected S-RNases. ase activity of S-RNases are also indicated with an arrow- The C1 and C2 domains were not part of the initial par- head in the alignment. The eight conserved cysteine resi- tial sequence alignment of the cloned S-RNases (Fig. 2). dues found in solanaceous S-RNases (Ioerger et al. 1991) 1 3 Theor Appl Genet (2016) 129:1985–2001 1993 Fig. 4 Alignment of the deduced amino acid sequence of the full conserved histidine residues involved in the ribonuclease activity of gene sequence of So2-RNase (S. okaSo2) cloned from S. okadae S-RNases are marked with arrow heads. The eight conserved cysteine and Ss2-RNase (S. steSs2) from S. stenotomum with three published residues are marked with an asterisk (*) under the alignment. The full-length S-RNases from S. chacoense. The hypervariable regions only conserved potential N-glycosylation site found in solanaceous (HVa and HVb) and conserved regions (C1–C5) are boxed. The dots S-RNases is marked with a hash symbol (#) under the alignment in the alignment indicate similarities between the five sequences. The are also indicated, with the exception of So2-RNase which 2011), and type I and II groups are non-S-RNases and are has only seven cysteine residues as explained previously. generally referred to as S-like I and S-like II, respectively. Also, the only conserved single potential N-glycosyla- In an attempt to confirm that our cloned S-RNases belong tion site found within C2 in most solanaceous S-RNases to the class III RNase group and are genuine S-RNases and (Ioerger et al. 1991) is shown in the alignment. not S-like-RNases (as suggested from the Blast search), a phylogenetic tree was constructed based on an alignment Phylogenetic analysis of S‑RNases using our cloned S-RNases and selected class I and II RNase members from the Solanaceae (Fig. 5). The result- Phylogenetic trees were constructed using the neighbour- ing phylogenetic tree showed that the putative S-RNases joining method as implemented in MEGA5 (Tamura et al. cluster together with typical class III S-RNases used for 2011). Bootstrap values of the tree were calculated based reference and are distinct from the clades of S-like-RNases. on 1000 replicates to allow an estimate of how well the The fungal RNase (T2-RNase) which shares homology topology at a branch on the tree is supported. Plant T2-type with S-RNases (and from which the name T2-type RNase RNases have been classified into three categories: class I, was derived) was used as an out-group to root the phyloge- II and III. All S-RNase genes identified to date belong to netic tree (see supplementary list S2 for the accession num- the class III type/group (Igic and Kohn 2001; Nowak et al. bers of the selected S-like-RNases used). Although there 1 3 1 994 Theor Appl Genet (2016) 129:1985–2001 53 S.okaSo5 S.steSs8 S.phuSp1 99 S.steSs9 99 S.steSs5 S.okaSo1 89 100 S.chaS11 S.okaSo4 P.infS2 94 61 S.steSs3 S-RNases 100 S.steSs10 S.phuSp2 100 S.steSs6 S.okaSo3 S.okaSo2 97 99 S.steSs1 S.steSs7 66 S.steSs2 82 N.alaS3 S.steSs4 99 S.lyc, RNS2 100 S.lyc, LER S-like II RNases N.glu, NGR2 90 N.glu, NGR3 S.lyc, LE 76 S-like I RNases 86 N.ala, NE 86 N.tab, NK1 RNase T2 Fungal RNase T2 0.2 Fig. 5 Phylogenetic tree of S-RNases and S-like RNases. A phylo- and only those exceeding 50 % are shown. Bootstrap values were genetic tree of the 16 cloned S-RNases and selected S-like RNases based on 1000 replicates. The phylogenetic tree was drawn using from Solanaceae. Three published S-RNases, one from each of MEGA5 software. N.ala = Nicotiana alata, N.glu = Nicotiana glu- Petunia, Nicotiana and Solanum chaoense (highlighted with a tri- tinosa, N.tab = Nicotiana tabacum, S.lyc = Solanum lycopersicon, angle) were included as reference sequences. Fungal RNase T2 of S.oka = Solanum okadae, S.phu = Solanum phureja, S.ste = Sola- Aspergillus oryzae was included as an out-group to root the phylo- num stenotomum genetic tree. Numbers are bootstrap values expressed as a percentage are rare examples of sequences that group with S-RNases Fig. 6 Phylogenetic tree of selected S-RNases from the Solanaceae ▸ in phylogenenetic analysis that have been shown to be non- and the S-RNases reported in this study. The 16 cloned S-RNases are indicated with red diamonds. Three Antirrhinum S-RNases polymorphic and unlinked to the S-locus (e.g. Lee et al. are included as an out-group to root the phylogenetic tree. Other 1992), this is exceptional, and, generally, this association is sequences are selected solanaceous S-RNases retrieved from data- considered as good evidence for the identification of genu- bases. Numbers are bootstrap values expressed as a percentage and ine S-RNases (Igic and Kohn 2001). only those exceeding 50 % are shown. Bootstrap values were based on 1000 replicates. Phylogenetic tree was drawn using MEGA5 An expanded phylogenetic analysis to cover a more software. Solanum (potato) S-RNases are labelled in red, Solanum comprehensive sample of S-RNases in the Solanaceae (tomato) S-RNases are in green, Solanum carolinense S-RNases are is shown in Fig. 6. This phylogenetic tree is based on an in black, Petunia S-RNases are in blue, Lycium in fuchsia, Nicotiana alignment involving the partially deduced amino acid S-RNases in purple, Witheringia solanacae S-RNases are in aqua, Physalis S-RNases are in lime and Antirrhinum S-RNases are in sequence of the 16 cloned S-RNases and a sample of 70 maroon. A.his = Antirrhinum hispanicum, L.par = Lycium parishii, selected solanaceous S-RNases retrieved from public data- N.glu = Nicotiana glutinosa, P.inf = Petunia inflata, P.lon = Phys- bases (see supplementary list S3). This sample included 10 alis longifolia, S.car = Solanum carolinense, S.chac = Solanum S-alleles from each of seven representative species from chacoense, S.chil = Solanum chilense, S.oka = Solanum okadae, S.ste = Solanum stenotomum, W.sol = Witheringia solanacae 1 3 Theor Appl Genet (2016) 129:1985–2001 1995 W.solS5 P.lonS6 67 W.solS6 53 W.solS4 99 P.lonS9 W.solS3 N.alaSc10 96 N.alaS3 S.carF-SC 68 L.parS10 N.alaS2 N.alaS6 56 S.chiS4 59 S.steSs1 96 S.chiS2 S.steSs7 100 S.carH-SC S.chiS6 75 99 S.chiS8 S.chiS9 S.steSs2 99 S.chaS12 P.infS10 L.parS8 S.chaS3 75 N.alaS5 73 N.alaS75 99 S.okaSo2 S.carD-SC L.parS3 L.parS1 100 56 L.parS2 N.alaS70 P.infS11 96 68 S.carA-SC 89 S.okaSo3 93 L.parS5 50 L.parS9 96 S.chaS2 S.chiS10 100 L.parS6 L.parS7 96 S.carB-SC 100 S.carK-SC 71 S.steSs6 S.chiS7 69 S.chiS11 S.phuSp2 68 99 S.chaS14 S.carC-SC N.alaS210 S.steSs4 95 74 W.solS7 80 100 W.solS11 N.alaS27 65 P.infS2 P.infS8 P.infS3 53 P.infS12 P.infS1 98 S.okaSo4 N.alaS107 66 85 S.steSs3 100 S.steSs10 100 P.infS6 PinlfS9 91 S.carG-SC S.carJ-SC 71 56 P.lonS2 P.lonS4 78 P.lonS1 83 P.lonS3 98 P.lonS10 80 P.lonS5 P.lonS7 84 99 P.lonS8 S.carE-SC 85 W.solS1 93 W.solS9 99 W.solS8 W.solS2 S.chaS11 S.okaSo1 99 79 S.chaS13 L.parS4 P.infS7 S.chiS1 S.steSs5 SchiS3 75 S.steSs8 S.okaSo5 S.phuSp1 A.hisS4 A.hisS2 99 A.hisS5 0.2 1 3 1996 Theor Appl Genet (2016) 129:1985–2001 Table 6 Analysis of progenies resulting from S. okadae crosses by PCR using allele-specific primers Cross name Number of progeny Parental genotypes Expected Number of individuals evaluated (♀) × (♂) genotypes obtained genotypes OKA 1 × OKA 5 30 (S1S2) × (S1S5) S1S5 13 S2S5 17 OKA 1 × OKA 9 29 (S1S2) × (S1S4) S1S4 18 S2S4 11 OKA 3 × OKA 5 29 (S1S2) × (S1S5) S1S5 16 S2S5 13 OKA 3 × OKA 9 28 (S1S2) × (S1S4) S1S4 14 S2S4 14 Sample 8 Sample 9 Sample 19 Sample 20 H So1 So2 So4 So5 So1 So2 So4 So5 So1 So2 So4 So5 So1 So2 So4 So5 H Fig. 7 Progeny analysis of semi-compatible crosses with allele-spe- tive for So1 and So5. The reference band indicated with a red arrow at cific primers. S-allele genotyping of a selection of four progeny (sam- approximately 230 bp is the result of an internal control PCR. The ples 8, 9, 19 and 20) from the cross OKA1 (So1So2) × OKA5 (So1So5). full genotype analysis of this family based on the genotype of 30 Each DNA sample was amplified with allele-specific primers for the individuals is shown in Table 6 So1, So2, So4 and So5 alleles, respectively, as indicated. Samples 8 and 9 were positive for So2 and So5, whilst samples 19 and 20 were posi- the genera: Solanum, Nicotiana, Petunia, Physalis, Lycium Witheringia and Nicotiana but no Solanum alleles identi- and Witheringia. Three Antirrhinum S-RNase sequences fied to date. The clustering of the S-RNases of one species/ are also included as an ‘out-group’ to root the phyloge- genus with members of another species/genus rather than netic tree (Xue et al. 1996). The topology of the phylo- the same species/genus is a commonly observed feature genetic tree showed that the cloned potato S-RNases are of solanaceous S-RNases reflecting the ancient origin and dispersed across many solanaceous lineages. Also, inter- diversification of the S-RNases in the common ancestors of specific similarities rather than intraspecific similarities the Solanaceae (Ioerger et al. 1990). are often observed among the S-RNases, i.e. the cloned S-RNases do not often cluster into species-specific clades Progeny analysis with allele‑specific primers but rather form clades with alleles from other members of the Solanaceae that are supported by high bootstrap values To test S-allele function, a selection of semi-compati- in most cases. For example, S.okaSo2 is part of a strongly ble crosses was performed in Solanum okadae. Progeny supported clade with two alleles from Nicotiana alata, one plants of four crosses between genotypes OKA1, OKA3, from Solanum carolinense, and none from Solanum oka- OKA5 and OKA9 were analysed for the presence of the dae. In some cases, the S-RNases can be observed to be expected S-RNase alleles. Allele-specific primers were clustered with S-RNases from exclusively other genera of designed for each of the four alleles represented in these the Solanaceae rather than forming genus-specific clades. genotypes (So1, So2, So4 and So5) and used in PCR ampli- For example, S.steSs4 is part of a clade with alleles of fications from each of 28–30 progeny plants per cross. 1 3 Theor Appl Genet (2016) 129:1985–2001 1997 The allelic constitution of these plants is summarised in inferred. The first is a compatible reaction (cross) where Table 6, and an example showing the PCR genotyping of the parents involved harbour different pairs of S-haplotypes S-RNases from four such individuals is shown in Fig. 7. In and, therefore, should not lead to the arrest of the growing each case, the individual plant genotypes were consistent pollen tube. Alternatively, the parents could differ by at with a ‘semi-compatible’ reaction, whereby pollen contain- least one S-allele in a semi-compatible cross where one of ing an allele held in common with the pistillate parent did the alleles go through to effect fertilisation, whilst the one not fertilise the female parent. All progeny plants contained shared by both parents will be arrested. It should be noted the ‘non-common’ S-allele from the pollen donor parent that all of the self-pollinated S. okadae plants did not yield in a heterozygous state. For example, in the first cross in any berries/seeds, thereby confirming that the accessions Table 6 (OKA1 × OKA5), the S5 allele of the pollen parent are, indeed, self-incompatible. is inherited by all progeny. Conversely, the S1 allele of the The S-genotypes inferred from the S-RNase sequence pollen parent is efficiently rejected as no progeny of S1S1 or analysis in S. okadae (Table 3) are entirely consistent with S1S2 genotype was identified. All 116 genotyped progenies the pollination data. OKA1 and OKA3 are shown to have are consistent with the transmission of the expected single the same genotype So1So2 consistent with the reciprocal staminate S-RNase allele. This provides good evidence that cross-incompatibility observed between these two lines. the cloned sequences are equivalent to functional S-alleles The S-genotyping further suggests that the crosses between showing gametophytic control of the pollen phenotype OKA1 and OKA3 and the other three lines studied should such that matching alleles are efficiently rejected during be either semi-compatible (OKA5 and OKA9) or possi- pollination and do not appear in the zygotes formed. bly fully compatible (OKA7). The RT-PCR cloning could not identify a second S-RNase allele in OKA7, possibly due to a divergence from the consensus C2 sequence, in Discussion which case the crosses would be fully compatible. The seed set observed in all these crosses in Table 2 supports this SI status confirmation and the compatibility interpretation, and it was not possible to distinguish semi- relationships of potato germplasm compatibility and full compatibility based on the amount of seed set. This suggests that pollen is available in excess The SI mechanism in angiosperms enables the female such that all possible ovules are fertilised even in a semi- reproductive organ of a flower to recognise and distin- compatible pollination. guish between self-pollen and non-self-pollen and, hence, to allow only the non-self-pollen to effect fertilisation. In Progeny analysis of semi‑compatible pollinations the GSI system, depending on the nature of the S-haplo- types expressed in both the pollen and the pistil, differ- To further test the above interpretation, a series of four ent compatibility relationships could be observed. Thus, a Solanum okadae semi-compatible progenies were geno- compatible or semi-compatible reaction will occur when typed at the S-RNase locus using allele-specific primers. at least one of the haplotypes expressed in the pollen and The four allele-specific primers were designed to amplify pistils do not match. If the two parents differ by just one all four possible S-alleles expected to be segregating in haplotype, then a semi-compatible reaction will occur, and the progenies of the respective crosses. In all the crosses, all the pollen tubes carrying the shared haplotype will be a semi-compatible reaction was expected since a common arrested, whilst those carrying the unique haplotype will be S-allele, So1-RNase, was present in all the parents involved accepted. Alternatively, when both of the expressed S-hap- in the crosses (Table 6). The S-genotyping results for a lotypes in the pollen and pistil are matched, it will result total of 116 progeny are consistent with the pollination data in an incompatible reaction leading to the arrest of all the (Table 2) and the deduced S-RNase analysis sequence data growing pollen tubes. (Table 3), such that the expected segregation of S-alleles The diallel cross design used here has established the was observed in the progenies of the respective crosses. compatibility relationships among the five S. okadae gen- The observations from the allele-specific PCR genotyping otypes studied (Table 2). For instance, crosses between are consistent with the expectation that all the pollen tubes OKA1 and OKA3 did not yield any berries or seeds in carrying the shared S-haplotype are arrested, whilst those either direction. This implies that both parents harbour the carrying the unique S-haplotype are accepted. For example, same pair of S-haplotypes, hence, the arrest of all growing in the first cross in Table 6 (S1S2 × S1S5), there were no pollen tubes results in an incompatible reaction. Crosses progeny of S1S2 or S1S1 genotype identified, consistent with among all the other S. okadae genotypes resulted in the pro- the rejection of the S1 haplotype in pollen. These crosses duction of berries and seeds in either direction. From these provide good evidence that the cloned S-RNase sequences crosses, two plausible compatibility relationships could be in S. okadae represent functional S-alleles. 1 3 1998 Theor Appl Genet (2016) 129:1985–2001 Primary structural features of S‑RNases in selected from some published functional solanaceous S-RNases, diploid potato plants notably the four best matches for the So2-RNase from the NCBI database searches (i.e. three S-RNases from S. peru- Relatively few S-RNase sequences are available for potato vianum with database accession numbers AAB26702.1, species (Solanum subsection Petota) compared to other CAA53666.1 and BAC00933.1 and one from N. glauca members of the Solanaceae. As of December 2015, the with database accession number AAV69976.1). The forma- NCBI database mining for S-RNases revealed only 17 tion of disulphide bridges by the cysteine residues is con- S-RNases for potato (Table 4), most of which were cloned sidered vital for forming and stabilising the tertiary struc- from either S. chacoense or S. tuberosum Group Stenoto- ture of the proteins (Ishimizu et al. 1996; Ida et al. 2001) mum. There are some additional S-alleles from S. tubero- although it appears that at least one disulphide bridge is sum mentioned in the literature (Kirch et al. 1989; Kauf- dispensible. All of the cloned putative S-RNase proteins mann et al. 1991), however, not all of them are in public (full and partial sequences) were predicted to have strong databases. The use of the C2 domain degenerate primer basic isoelectric point values (8.6–9.6), which is consist- and the 3′ RACE strategy described here has enabled the ent with observations made with functional S-RNases cloning of an additional 16 novel putative S-RNases from involved in the self-incompatibility reaction, i.e. functional accessions of three diploid potato plants; S. okadae, S. S-RNases are basic proteins having an isoelectric point (pI) tuberosum Group Stenotomum and S. tuberosum Group value of >7.5 (Nowak et al. 2011), and usually between ~8 Phureja. Plants exhibiting the S-RNase-based GSI sys- and 10 (Roalson and McCubbin 2003). tem are expected to be heterozygotes bearing two differ- An attempt to clone the full length of selected S-RNases ent S-alleles. However, the observation that only one allele using the partial sequences obtained through the initial could be cloned from some of the accessions could be 3′RACE cloning has been successful for two alleles. The explained by the differential amplification of the S-RNases full-length cloning of the So2-RNase from S. okadae and by the degenerate primer. Alternatively, the unidentified Ss2-RNase from S. tuberosum Group Stenotomum allowed S-RNases may have very low transcript levels relative to the sequences of the 5′ regions to be determined. The full- the other allele and, hence, could not be detected or ampli- length sequences revealed the two conserved catalytic fied. Alleles at the S-locus can show significant variation histidine residues involved in the ribonuclease activity of in their transcript levels or their stability (Roldan et al. S-RNases and located in the C2 and C3 regions (Fig. 4). 2010). All the identified putative S-RNases are novel, and Also, the eight conserved cysteine residues could be no cDNA sequences were available for them prior to this observed in Ss2-RNase and only seven for So2-RNase which study. The 16 alleles reported here are a significant addi- lacks one such residue as described earlier. A variable num- tion to the existing number of S-RNase sequences reported ber of potential N-glycosylation sites can be identified in for tuber-bearing members of Solanum (subsection Petota), S-RNase sequences (e.g. Oxley et al. 1998; Qi et al. 2001). almost doubling the number available in public databases. However, the analysis of solanaceous S-RNase sequence Analysis of the partially deduced amino acid sequence has identified one single conserved potential N-glycosyla- obtained here (Fig. 2) showed that all the cloned sequences tion site found in the C2 conserved region (Ioerger et al. have the primary structural features of solanaceous 1991). This was observed to be conserved in the full-length S-RNases as originally defined by Ioerger et al. (1991). sequence of the two cloned putative S-RNases reported Also, the cloned partial S-RNase sequences contain one here. The presence of this conserved glycosylation site of the active site histidine (His) residues located in the C3 could possibly be responsible for modulating the S-RNase region which are involved in the ribonuclease activity of ribonuclease activity (Ioerger et al. 1991). However, stud- the S-RNase gene. The other histidine which is located in ies have shown that, the removal of the glycan-side chains the C2 region is not part of the partial sequences obtained did not alter the enzymatic activity of the S-RNase gene for most of the S-RNases cloned here, because the degener- in vitro (Broothaerts et al. 1991) or its function in self- ate primer used for the cloning is from the C2 region and, incompatibility in transgenic Petunia inflata (Karunana- hence, was removed from all partial sequences due to nucle- ndaa et al. 1994). otide sequence ambiguity in that region. The cloned partial S-RNase sequences contain six out of the eight conserved Solanaceous S‑RNases exhibit allelic diversity cysteine residues found in functional S-RNases (with the and intraspecific sequence polymorphism exception of So2-RNase) which can form potential disul- phide bonds (Ioerger et al. 1991). The partial So2-RNase An early report of the extreme level of polymorphism at sequence was observed to contain only five of the expected the S-locus of a narrow endemic species, Oenothera organ- six cysteine residues. This observation was not unprec- ensis, which was estimated to contain ca. 500 individuals edented since this particular cysteine residue is absent (Emerson 1939), generated interest in the understanding of 1 3 Theor Appl Genet (2016) 129:1985–2001 1999 the population genetics of gametophytic SI. The low level S-like-RNases, a phylogenetic tree was constructed using of amino acid sequence similarity observed in the par- an alignment of the cloned S-RNases and selected S-like- tial sequences of the S-RNases reported here (32.9 %) is RNases from the Solanaceae (Fig. 5). The observation from consistent with earlier observations made for solanaceous the phylogenetic analysis that the non-S-RNases (S-like I S-RNases (Ioerger et al. 1990; McCubbin and Kao 2000). and S-like II-RNases) fall outside the S-RNase clade is as Ioerger et al. (1990) analysed the S-locus of three species anticipated. S-like-RNases are known to be unlinked to the of the Solanaceae and observed a low level of amino acid S-locus (and hence are unlikely to be involved in GSI) and sequence similarity within a species as low as 40 %. This is may be involved in pathogen defence mechanisms or induced consistent with this study, where the amino acid sequence in response to phosphate starvation (Kao and McCubbin similarity within a species was as low as 32.9–44.2 %. Pol- 1996; Dodds et al. 1996; Hugot et al. 2002). The evolution- ymorphism at the S-locus and the high level of sequence ary relationship between S-RNases and S-like-RNases still divergence in solanaceous S-RNases could partly account remains uncertain, the S-like-RNases may be the ancestral for these observations. Solanaceous S-allele polymorphism genes involved in defence against pathogen attack in the has led to some unexpected observations where S-alleles style that were recruited with modifications to function in of one species or genus were found to be more closely the self-incompatibility reaction (Kao and McCubbin 1996). related to alleles from another species or genus; a case of Although rare examples of S-unlinked sequences, such as trans-specific or trans-generic evolution of these S-alleles RNase X2 in Petunia inflata, are known to cluster in S-RNase (Ioerger et al. 1990). It is proposed that the S-RNase alleles clades (e.g. Lee et al. 1992), all the S-RNases cloned from are exceptionally old and have been inherited from a com- this study appear to represent genuine S-alleles and are mon ancestor and passed down to multiple descendant clearly distinct from and distantly related to S-like-RNases. taxa, i.e. the S-allele lineage origins predate their current The use of phylogenetic tools has enabled us to compare species origin. More recent work has extended the initial the cloned S-RNases with other alleles from the Solanaceae observations in the Solanaceae to the Rosaceae family, to study the diversity of these S-RNases. A general obser- specifically in the genus Prunus (Sutherland et al. 2008). vation is that the S-RNases do not form species-specific Polymorphism at the S-locus in angiosperms is a result of clades (Fig. 6) but rather that interspecific clades (showing diversifying selection, the age of S-alleles and the absence interspecific similarities) are predominant. Similar results of recombination at the S-locus which acts to preserve and were obtained by Ioerger et al. (1990) when they analysed maintain allelic variations at the S-locus (Ioerger et al. the S-locus of three species of the Solanaceae and observed 1991; Richman and Kohn 2000; Igic et al. 2003). a higher interspecies similarity rather than intraspecies sim- ilarity. They concluded that polymorphism at the S-locus Phylogenetic analysis of solanaceous S‑RNases predates the divergence of the species in the Solanaceae, and this polymorphism has been maintained to the present The use of molecular phylogenetic analysis tools has ena- time. This extreme level of polymorphism in S-proteins bled robust phylogenetic trees to be constructed leading to indicates an unusual aspect of balancing selection which the identification of the most basal lineages of angiosperms. operates at the S-locus (Ioerger et al. 1990). With the help of phylogenetic studies, S-genes have been Furthermore, from the phylogenetic tree (Fig. 6), characterised allowing conclusions to be drawn about the some trans-generic clades can be observed. For instance, evolutionary history of their sequences (Allen and Hiscock S-RNases from Solanum (potato, tomato), S. carolinense, 2008). For instance, the use of phylogenetic tools has ena- Lycium, Petunia and Nicotiana can be observed to form bled Igic and Kohn (2001) to make predictions that GSI is extensive trans-generic clades indicating extensive diversi- ancestral to ~75 % of eudicots and that RNase-based self- fication of S-alleles in all these genera. In contrast, reduced incompatibility of the GSI system was the ancestral state or very limited trans-generic lineages could be observed of self-incompatibility system existing in the majority of for S-RNases from Witheringia and Physalis. This is con- dicots. sistent with previous studies that have revealed a cluster- All the identified S-RNases known to be involved in the ing together of the Physalis alleles in just three clades of GSI reaction in plants belong to the T2-type Class III gene the extensive Solanaceae S-allele phylogenetic tree (Rich- subfamily (Igic and Kohn 2001). Other non-S-RNases man et al. 1996; Richman and Kohn 1999). Richman et al. referred to as S-like-RNases have also been identified in (1996) proposed that the loss of trans-generic lineages in many plant species. These S-like-RNases share domain struc- Physalis crassifolia was an outcome of the effect of severe ture with S-RNases and appear in phylogenetic analyses to be population bottlenecks imposed on this genus. Similar related to S-RNases but have no role in the self-incompati- explanations have been proposed to account for the reduced bility reaction. In an initial attempt to show that the cloned or very limited trans-generic lineages observed in the Solanum putative S-RNases are genuine S-RNases and not closely related genus Witheringia (Stone and Pierce 2005). 1 3 2 000 Theor Appl Genet (2016) 129:1985–2001 Following the identification of the first S-protein associated with expression of self-incompatibility in Nicotiana sequence from Nicotiana (Anderson et al. 1986), a large alata. Nature 321:38–44 Asquini E, Gerdol M, Gasperini D, Igic B, Graziosi G, Pallavicini A number of S-RNase sequences have been isolated from (2011) S-RNase-like sequences in styles of Coffea (Rubiaceae). other species of the Solanaceae. However, unlike some Evidence for S-RNase based gametophytic self-incompatibility? other members of the Solanaceae, relatively few S-RNase Trop Plant Biol 4(3):237–249 gene sequences are available for potato. The relatively large Bredemeijer GM, Blass J (1981) S-specific proteins in styles of self- incompatible Nicotiana alata. Theor Appl Genet 59:185–190 number of putative S-RNases identified from the relatively Broothaerts W, Vanvinckenroye P, Decock B, Van Damme J, Vendrig J small number of potato genotypes from this current study (1991) Petunia hybrida S-proteins: ribonuclease activity and the implies that there is a high level of S-RNase gene vari- role of their glycan side chains in self-incompatibility. Sex Plant ability and diversity in potato. However, this S-gene vari- Reprod 4:258–266 Cipar MS, Peloquin SJ, Hougas RW (1964) Inheritance of incompat- ability has not been well exploited and characterised com- ibility in hybrids between Solanum tuberosum haploids and dip- pared to other species in the Solanaceae. It is worth noting loid species. Euphytica 13:163–172 that although functional assays (e.g. transgenic) have not de Nettancourt D (1977) Incompatibility in angiosperms. Springer, been performed for the cloned S-RNases, their sequence Berlin de Nettancourt D (1997) Incompatibility in angiosperms. Sex Plant characterisation coupled with selected pollination tests in Reprod 10:185–199 Solanum okadae indicates that they are likely to be genu- de Nettancourt D (2001) Incompatibility and incongruity in wild and cul- ine S-RNases and in the case of those tested by pollina- tivated plants. Springer, Berlin tion clearly function in the SI reaction. The S-RNase genes Despres C, Saba-El-Leil MK, Rivard SR, Morse D, Cappadocia M (1994) Molecular cloning of two Solanum chacoense S-alleles reported here represent unique and useful additions to the and a hypothesis concerning their evolution. Sex Plant Reprod limited available potato S-RNase gene sequence database. 7:169–176 The identified alleles can be used for further studying the Dodds KS (1962) Classification of cultivated potatoes. In: Correll DS diversity and phylogenetic relationship of S-alleles, particu- (ed) The potato and its wild relatives. Texas Research Founda- tion, Renner, pp 517–539 larly in tuber-bearing Solanum (subsection Petota). These Dodds PN, Clarke AE, Newbigin E (1996) A molecular perspective findings may also have application for the maintenance and on pollination in flowering plants. Cell 85:141–144 application of potato germplasm for crop improvement. Emerson S (1939) A preliminary survey of the Oenothera organensis population. Genetics 24:524–537 Author contribution statement Experiments were devised by TPR, Felsenstein J (1985) Confidence limits on phylogenies: an approach GJB and DKD. Experiments were executed by DKD and GW. The using the bootstrap. Evolution 39:783–791 manuscript was written by DKD, TPR and GJB. Franklin-Tong VE, Franklin FCH (2003) Gametophytic self-incom- patibility inhibits pollen tube growth using different mecha- Compliance with ethical standards nisms. Trends Plant Sci 12:598–605 Hiscock SJ, McInnis SM (2003) The diversity of self-incompatibility Funding This study was funded through a PhD studentship for systems in flowering plants. Plant Biol 5:23–32 D.K.D. provided by The James Hutton Institute and the University of Hugot K, Ponchet M, Marais A, Ricci P, Galiana E (2002) A tobacco Nottingham. S-like RNase inhibits hyphal elongation of plant pathogens. Mol Plant Microbe Interact 15:243–250 Conflict of interest The authors declare that they have no conflict of Ida K, Norioka S, Yamamoto M, Kumasaka T, Yamashita E, Newbigin interest. E, Clarke AE, Sakiyama F, Sato M (2001) The 1.55 A resolution structure of Nicotiana alata SF11-RNase associated with game- tophytic self-incompatibility. J Mol Biol 314:103–112 Open Access This article is distributed under the terms of the Crea- Igic B, Kohn JR (2001) Evolutionary relationships among self-incom- tive Commons Attribution 4.0 International License (http://crea- patibility RNases. Proc Natl Acad Sci USA 98:13167–13171 tivecommons.org/licenses/by/4.0/), which permits unrestricted use, Igic B, Bohs L, Kohn JR (2003) Historical inferences from the self- distribution, and reproduction in any medium, provided you give incompatibility locus. New Phytol 161:97–105 appropriate credit to the original author(s) and the source, provide a Ioerger TR, Clark AG, Kao TH (1990) Polymorphism at the self- link to the Creative Commons license, and indicate if changes were incompatibility locus in Solanaceae predates speciation. Proc made. Natl Acad Sci USA 87:9732–9735 Ioerger TR, Gohlke JR, Xu B, Kao TH (1991) Primary structural fea- tures of the self-incompatibility protein in Solanaceae. Sex Plant Reprod 4:81–87 References Ishimizu T, Norioka S, Kanai M, Clarke AE, Sakiyama F (1996) Location of cysteine and cystine residues in S-ribonucleases Allen AM, Hiscock SJ (2008) Evolution and phylogeny of self- associated with gametophytic self-incompatibility. Eur J Bio- incompatibility systems in angiosperms. In: Franklin-Tong VE chem 242:627–635 (ed) Self-incompatibility in flowering plants: evolution, diversity Kao TH, McCubbin AG (1996) How flowering plants discriminate and mechanisms. Springer, Berlin, pp 73–101 between self and non-self pollen to prevent inbreeding. Proc Natl Anderson MA, Cornish EC, Mau SL, Williams EG, Hoggart R, Atkin- Acad Sci USA 93:12059–12065 son A, Bonig I, Grego B, Simpson R, Roche PJ, Haley JD, Pen- Karunanandaa B, Huang S, Kao T (1994) Carbohydrate moiety of the schow JD, Niall HD, Tregear GW, Coughlan JP, Crawford RJ, Petunia inflata S3 protein is not required for self-incompatibility Clarke AE (1986) Cloning of cDNA for a stylar glycoprotein interactions between pollen and pistil. Plant Cell 6:1933–1940 1 3 Theor Appl Genet (2016) 129:1985–2001 2001 Kaufmann H, Salamini F, Thompson RD (1991) Sequence variability Roalson EH, McCubbin AG (2003) S-RNases and sexual incompat- and gene structure at the self-incompatibility locus of Solanum ibility: structure, functions, and evolutionary perspectives. Mol tuberosum. Mol Gen Genet 226:457–466 Phylogenet Evol 29:490–506 Kirch HH, Uhrig H, Lottspeich F, Salamini F, Thompson RD (1989) Roldan JA, Quiroga R, Goldraij A (2010) Molecular and genetic Characterization of proteins with self-incompatibility in Sola- characterization of novel S-RNases from a natural population of num tuberosum. Theor Appl Genet 78:581–588 Nicotiana alata. Plant Cell Rep 29:735–746 Lee HS, Singh A, Kao TH (1992) RNase X2, a pistil specific ribonu- Saba-el-Leil MK, Rivard S, Morse D, Cappadocia M (1994) The S11 clease from Petunia inflata, shares sequence similarity with sola- and S13 self incompatibility alleles in Solanum chacoense Bitt. naceous S proteins. Plant Mol Biol 20:1131–1141 are remarkably similar. Plant Mol Biol 24(4):571–583 Lee HS, Huang S, Kao TH (1994) S proteins control rejection of Saitou N, Nei M (1987) The neighbor-joining method: a new method incompatible pollen in Petunia inflata. Nature 367:560–563 for reconstructing phylogenetic trees. Mol Biol Evol 4:406–425 Marcellan ON, Acevedo A, Camadro EL (2006) S16, a novel S-RNase Silva NF, Goring DR (2001) Mechanisms of self-incompatibility in allele in the diploid species Solanum chacoense. Genome flowering plants. Cell Mol Life Sci 58:1988–2007 49(8):1052–1054 Spooner DM, Ghislain M, Simon R, Jansky SH, Gavrilenko T (2014) McClure BA, Haring V, Ebert PR, Anderson MA, Simpson RJ, Sakiy- Systematics, diversity, genetics, and evolution of wild and culti- ama F, Clarke AE (1989) Style self-incompatibility gene prod- vated potatoes. Bot Rev 80:283–383 ucts of Nicotiana alata are ribonucleases. Nature 342:955–957 Stone JL, Pierce SE (2005) Rapid recent radiation of S-RNase McCubbin AG, Kao TH (2000) Molecular recognition and response lineages in Witheringia solanacea (Solanaceae). Heredity in pollen and pistil interactions. Annu Rev Cell Dev Biol 94:547–555 16:333–364 Sutherland BG, Tobutt K, Robbins TP (2008) Trans-specific S-RNase Murfett J, Atherton T, Mou B, Gasser C, McClure B (1994) S-RNase and SFB alleles in Prunus self-incompatibility haplotypes. Mol expressed in transgenic Nicotiana causes S-allele-specific pollen Genet Genom 279:95–106 rejection. Nature 367:563–566 Takayama S, Isogai A (2005) Self-incompatibility in plants. Ann Rev Nowak MD, Davis AP, Anthony F, Yoder AD (2011) Expression and Plant Biol 56:467–489 trans-specific polymorphism of self-incompatibility RNases in Tamura K, Peterson D, Peterson N, Stecher G, Nei M, Kumar S Coffea (Rubiaceae). PLoS ONE 6:e21019. doi:10.1371/journal. (2011) MEGA5: molecular evolutionary genetics analysis using pone.0021019 maximum likelihood, evolutionary distance, and maximum par- O’Brien M, Kapfer C, Major G, Laurin M, Bertrand C, Kondo K, simony methods. Mol Biol Evol 28:2731–2739 Kowyama Y, Matton DP (2002) Molecular analysis of the sty- Thompson J, Higgins D, Gibson T, Thompson JD, Higgins DG, Gib- lar-expressed Solanum chacoense small asparagine-rich protein son TJ (1994) CLUSTAL W: improving the sensitivity of pro- family related to the HT modifier of gametophytic self-incom- gressive multiple sequence alignment through sequence weight- patibility in Nicotiana. Plant J 32(6):985–996 ing, position-specific gap penalties and weight matrix choice. Oxley D, Munro SL, Craik DJ, Bacic A (1998) Structure and distri- Nucleic Acids Res 22:4673–4680 bution of N-glycans on the S7-allele stylar self-incompatibility Wheeler MJ, de Graaf BHJ, Hadjiosif NE, Perry RM, Poulter NS, ribonuclease of Nicotiana alata. J Biochem 123:978–983 Osman K, Vatovec S, Harper A, Franklin FCH, Franklin-Tong Pandey KK (1962) Interspecific incompatibility in Solanum species. VE (2009) Identification of the pollen self-incompatibility deter- Am J Bot 49:874–882 minant in Papaver rhoeas. Nature 459:992–995 Pushkarnath (1942) Studies on sterility in potatoes. 1. The genetics Xu BB, Mu JH, Nevins DL, Grun P, T-h Kao (1990) Cloning and of self- and cross-incompatibilities. Indian J Genet Plant Breed sequencing of cDNAs encoding two self-incompatibility 2:11–36 associated proteins in Solanum chacoense. Mol Gen Genet Qi X, Luu D, Yang Q, Maes O, Matton D, Morse D, Cappadocia M 224(3):341–346 (2001) Genotype-dependent differences in S12-RNase expression Xue Y, Carpenter R, Dickinson HG, Coen ES (1996) Origin of lead to sporadic self-compatibility in Solanum chacoense. Plant allelic diversity in Antirrhinum S Locus RNases. Plant Cell Mol Biol 45:295–305 8(5):805–814 Richman AD, Kohn JR (1999) Self-incompatibility alleles from Phys- Zuckerkandl E, Pauling L (1965) Evolutionary divergence and con- alis: implications for historical inference from balanced genetic vergence in proteins. In: Bryson V, Vogel HJ (eds) Evolving polymorphisms. Proc Natl Acad Sci USA 96:168–172 genes and proteins. Academic Press, New York, pp 97–166 Richman AD, Kohn JR (2000) Evolutionary genetics of self-incom- patibility in the Solanaceae. Plant Mol Biol 42:169–179 Richman AD, Uyenoyama MK, Kohn JR (1996) Allelic diversity and gene genealogy at the self-incompatibility locus in the Solan- aceae. Science 273:1212–1216 1 3