Isolation, genetic variation and expression of TIR-NBS-LRR resistance gene analogs from western white pine ( Pinus monticola Dougl. ex. D. Don.)

Western white pine ( Pinus monticola Dougl. ex. D. Don., WWP) shows genetic variation in disease resistance to white pine blister rust ( Cronartium ribicola). Most plant disease resistance (R) genes encode proteins that belong to a superfamily with nucleotide-binding site domains (NBS) and C-terminal leucine-rich repeats (LRR). In this work a PCR strategy was used to clone R gene analogs (RGAs) from WWP using oligonucleotide primers based on the conserved sequence motifs in the NBS domain of angiosperm NBS-LRR genes. Sixty-seven NBS sequences were cloned from disease-resistant trees. BLAST searches in GenBank revealed that they shared significant identity to well-characterized R genes from angiosperms, including L and M genes from flax, the tobacco N gene and the soybean gene LM6. Sequence alignments revealed that the RGAs from WWP contained the conserved motifs identified in angiosperm NBS domains, especially those motifs specific for TIR-NBS-LRR proteins. Phylogenic analysis of plant R genes and RGAs indicated that all cloned WWP RGAs can be grouped into one major branch together with well-known R proteins carrying a TIR domain, suggesting they belong to the subfamily of TIR-NBS-LRR genes. In one phylogenic tree, WWP RGAs were further subdivided into fourteen clusters with an amino acid sequence identity threshold of 75%. cDNA cloning and RT-PCR analysis with gene-specific primers demonstrated that members of 10 of the 14 RGA classes were expressed in foliage tissues, suggesting that a large and diverse NBS-LRR gene family may be functional in conifers. These results provide evidence for the hypothesis that conifer RGAs share a common origin with R genes from angiosperms, and some of them may play important roles in defense mechanisms that confer disease resistance in western white pine. Ratios of non-synonymous to synonymous nucleotide substitutions (Ka/Ks) in the WWP NBS domains were greater than 1 or close to 1, indicating that diversifying selection and/or neutral selection operate on the NBS domains of the WWP RGA family.     

Isolation of a family of resistance gene analogue sequences of the nucleotide binding site (NBS) type from Lens species.

Most known plant disease-resistance genes (R genes) include in their encoded products domains such as a nucleotide-binding site (NBS) or leucine-rich repeats (LRRs). Sequences with unknown function, but encoding these conserved domains, have been defined as resistance gene analogues (RGAs). The conserved motifs within plant NBS domains make it possible to use degenerate primers and PCR to isolate RGAs. We used degenerate primers deduced from conserved motifs in the NBS domain of NBS-LRR resistance proteins to amplify genomic sequences from Lens species. Fragments from approximately 500-850 bp were obtained. The nucleotide sequence analysis of these fragments revealed 32 different RGA sequences in Lens species with a high similarity (up to 91%) to RGAs from other plants. The predicted amino acid sequences showed that lentil sequences contain all the conserved motifs (P-loop, kinase-2, kinase-3a, GLPL, and MHD) present in the majority of other known plant NBS-LRR resistance genes. Phylogenetic analyses grouped the Lens NBS sequences with the Toll and interleukin-1 receptor (TIR) subclass of NBS-LRR genes, as well as with RGA sequences isolated from other legume species. Using inverse PCR on one putative RGA of lentil, we were able to amplify the flanking regions of this sequence, which contained features found in R proteins.     

Survey of resistance gene analogs in <Emphasis Type="Italic">Solanum caripense</Emphasis>, a relative of potato and tomato,and update on <Emphasis Type="Italic">R</Emphasis> gene genealogy

The resistance (R) proteins of the TIR- and non-TIR (or CC-) superfamilies possess a nucleotide binding site (NBS) domain. Within an R gene, the NBS is the region of highest conservation, suggesting an essential role in triggering R protein activity. We compared the NBS domain of functional R genes and resistance gene analogs (RGA) amplified from S. caripense genomic DNA via PCR using specific and degenerate primers with its counterpart from other plants. An overall high degree of sequence conservation was apparent throughout the P-loop, kinase-2 and kinase-3a motifs of NBS fragments from all plants. Within the non-TIR class of R genes a prominent sub-class similar to the potato R1 gene conferring resistance to late blight, was detected. All non-TIR-R1-like R gene fragments that were sequenced possessed an intact open reading frame, whereas 22% of all non-TIR-non-R1-like fragments and 59% of all TIR-NBS RGA fragments had an interrupted reading frame or contained transposon-specific sequence. The non-TIR-R1-like fragments had high similarity to Solanaceae R genes and low similarity to RGAs of other plant species including A. thaliana and the cereals. It is concluded that appearance of the non-TIR-R1-like NBS domain represents a relatively recent evolutionary development. Electronic supplementary material Supplementary material is available in the online version of this article at and is accessible for authorized users.     

Isolation,Cloning and Characterization of Resistance Gene Analogues in Pearl Millet Based on Conserved Nucleotide‐binding Sites

Sequence analysis of plant disease resistance genes shows similarity among themselves, with the presence of conserved motifs common to the nucleotide‐binding site (NBS). Oligonucleotide degenerate primers designed from the conserved NBS motifs encoded by several plant disease resistance genes were used to amplify resistance gene analogues (RGAs) corresponding to the NBS sequences from the genomic DNA of various plant species. Using specific primers designed from the conserved NBS regions, 22 RGAs were cloned and sequenced from pearl millet (Pennisetum glaucum L. Br.). Phylogenetic analysis of the predicted amino acid sequences grouped the RGAs into nine distinct classes. GenBank database searches with the consensus protein sequences of each of the nine classes revealed their conserved NBS domains and similarity to other known R genes of various crop species. One RGA 213 was mapped onto LG1 and LG7 in the pearl millet linkage map. This is the first report of the isolation and characterization of RGAs from pearl millet, which will facilitate the improvement of marker‐assisted breeding strategies.     

Cloning and characterisation of a family of disease resistance gene analogs from wheat and barley

 The most common class of plant disease resistance (R) genes cloned so far belong to the NBS-LRR group which contain nucleotide-binding sites (NBS) and a leucine-rich repeat (LRR). Specific primer sequences derived from a previously isolated NBS-LRR sequence at the Cre3 locus, which confers resistance to cereal cyst nematode (CCN) in wheat (Triticum aestivum L.) were used in isolating a family of resistance gene analogs (RGA) through a polymerase chain reaction (PCR) cloning approach. The cloning, analysis and genetic mapping of a family of RGAs from wheat (cv ‘Chinese Spring’) and barley (Hordeum vulgare L. cvs ‘Chebec’ and ‘Harrington’) are presented. The wheat and barley RGAs contain other conserved motifs present in known R genes from other plants and share between 55–99% amino acid sequence identity to the NBS-LRR sequence at the Cre3 locus. Phylogenetic analysis of the RGAs with other cloned R genes and RGAs from various plant species indicate that they belong to a superfamily of NBS-containing genes. Two of the barley derived RGAs were mapped onto loci on chromosomes 2H (2), 5H (7) and 7H (1) using barley doubled haploid (DH) mapping populations. Some of these loci identified are associated with regions carrying resistance to CCN and corn leaf aphid. Received: 6 January 1998 / Accepted: 1 April 1998     

PCR-based approach for mining of resistant gene analogues in taro (Colocasia esculenta)

Primers based on the conserved motifs were used to isolate nucleotide-binding sites (NBS) type sequences in taro (Colocasia esculenta). Cloning and sequencing identified three taro NBS-type sequences called resistance gene analogues (RGAs) that depicted similarity to other cloned RGA sequences. The deduced amino acid sequences of the RGAs detected the presence of conserved domains, viz. P-loop, categorising them with the NBS–leucine-rich repeat class gene family. Phylogenetic characterisation of the taro RGAs along with RGAs of other plant species grouped them with the non-toll interleukin receptor subclasses of the NBS sequences. The isolation and characterisation of taro RGAs have been reported for the first time in this study. This will provide a starting point towards characterisation of candidate resistance genes in taro and can act as a reference guide for future studies.     

Efficient targeting of plant disease resistance loci using NBS profiling

The conserved sequences in the nucleotide-binding sites of the nucleotide-binding site-leucine-rich repeat (NBS-LRR) class of disease resistance (R) genes have been used for PCR-based R-gene isolation and subsequent development of molecular markers. Here we present a PCR-based approach (NBS profiling) that efficiently targets R genes and R-gene analogs (RGAs) and, at the same time, produces polymorphic markers in these genes. In NBS profiling, genomic DNA is digested with a restriction enzyme, and an NBS-specific (degenerate) primer is used in a PCR reaction towards an adapter linked to the resulting DNA fragments. The NBS profiling protocol generates a reproducible polymorphic multilocus marker profile on a sequencing gel that is highly enriched for R genes and RGAs. NBS profiling was successfully used in potato with several restriction enzymes, and several primers targeted to different conserved motifs in the NBS. Across primers and enzymes, the NBS profiles contained 50–90% fragments that were significantly similar to known R-gene and RGA sequences. The protocol was similarly successful in other crops (including tomato, barley, and lettuce) without modifications. NBS profiling can thus be used to produce markers tightly linked to R genes and R-gene clusters for genomic mapping and positional cloning and to mine for new alleles and new sources of disease resistance in available germplasm.Communicated by H.F. Linskens     

Molecular characterization of NBS-LRR-RGAs in the rose genome

To isolate resistance gene analogues (RGAs) from roses we used various degenerate oligonucleotide primers targeting conserved motifs within the NBS region of nucleotide binding site (NBS)-leucine-rich repeat (LRR) resistance genes. A large RGA sublibrary consisting of 7000 clones was constructed. This sublibrary contains at least 40 unique RGA families of the TIR (toll-/interleukin-1 receptor) and the LZ (leucine zipper) type, which were further analysed. Phylogenetic studies revealed close relationships of some rose RGAs to R genes and RGAs from other plants and gave rise to the assumption that rose R genes evolved from different starting points, prior to and subsequent to speciation. Southern blot analyses showed that the RGAs were organized as single, low and multicopy loci in the rose genome. None of the analysed sequences detected any hybridization signal in Prunus cérasus indicating that the analysed RGAs are not conserved across genera. The efficiency and selectivity of the different degenerate primers used for the RGA isolation is discussed in detail.     

Characterization of disease resistance gene candidates of the nucleotide binding site (NBS) type from banana and correlation of a transcriptional polymorphism with resistance to <Emphasis Type="Italic">Fusarium oxysporum</Emphasis> f.sp. <Emphasis Type="Italic">cubense</Emphasis> race 4

Most plant disease resistance (R) genes encode proteins with a nucleotide binding site and leucine-rich repeat structure (NBS-LRR). In this study, degenerate primers were used to amplify genomic NBS-type sequences from wild banana (Musa acuminata ssp. malaccensis) plants resistant to the fungal pathogen Fusarium oxysporum formae specialis (f. sp.) cubense (FOC) race 4. Five different classes of NBS-type sequences were identified and designated as resistance gene candidates (RGCs). The deduced amino acid sequences of the RGCs revealed the presence of motifs characteristic of the majority of known plant NBS-LRR resistance genes. Structural and phylogenetic analyses grouped the banana RGCs within the non-TIR (homology to Toll/interleukin-1 receptors) subclass of NBS sequences. Southern hybridization showed that each banana RGC is present in low copy number. The expression of the RGCs was assessed by RT-PCR in leaf and root tissues of plants resistant or susceptible to FOC race 4. RGC1, 3 and 5 showed a constitutive expression profile in both resistant and susceptible plants whereas no expression was detected for RGC4. Interestingly, RGC2 expression was found to be associated only to FOC race 4 resistant lines. This finding could assist in the identification of a FOC race 4 resistance gene.     

Isolation and diversity analysis of resistance gene analogues (RGAs) from cultivated and wild strawberries

Degenerate oligonucleotide primers, designed based on conserved regions of Nucleotide Binding Site (NBS) domains from previously cloned plant resistance genes, were used to isolate Resistance Gene Analogues (RGAs) from wild and cultivated strawberries. Seven distinct families of RGAs of the NBS-LRR type were identified from two related wild species, Fragaria vesca and F. chiloensis, and six different Fragaria × ananassa cultivars. With one exception (GAV-3), the deduced amino acid sequences of strawberry RGAs showed strong similarity to TIR (Toll Interleukin I Receptor)-type R genes from Arabidopsis, tobacco and flax, suggesting the existence of common ancestors. GAV-3 seemed to be more closely related to the non-TIR type. Further studies showed that the recombination level and the ratio of non-synonymous to synonymous substitutions within families were low. These data suggest that NBS-encoding sequences of RGAs in strawberry are subject to a gradual accumulation of mutations leading to purifying selection, rather than to a diversifying process. The present paper is the first report on RGAs in strawberry.Communicated by M.-A. Grandbastien     

Isolation, characterization and evolutionary analysis of resistance gene analogs in annual ryegrass, perennial ryegrass and their hybrid

Recently, a number of disease-resistance genes related to a diverse range of pathogens were isolated from a wide variety of plant species. The majority of plant disease-resistance genes encoded a nucleotide-binding site (NBS) domain. According to the comparisons of the NBS domain of cloned R -genes, it has shown highly conserved amino acid motifs in this structure, which made it possible to isolate resistance gene analogs (RGAs) by PCR using degenerate primers. We have designed three pairs of degenerate primers based on two conserved motifs in the NBS domain of resistance proteins encoded by R -genes to amplify genomic sequences from ryegrass ( Lolium sp.). Sixteen NBS-like RGAs were isolated from turf and forage type grasses. The sequence analysis of these RGAs revealed that there existed a high similarity (up to 85%) between RGA sequences among ryegrass species and other plants. The alignment of the predicted amino acid sequences of RGAs showed that ryegrass RGAs contained four conserved motifs (P-Loop, kinase-2, kinase-3a, GLPL) present in other known plant NBS-leucine rich repeat resistance genes. These ryegrass RGAs all belonged to non-toll and interleukin-1 receptor subclass. Phylogenetic analysis of ryegrass RGAs and other cloned R -genes indicated that gene mutation was the predominant source of gene variations, and the sequence polymorphism was due to purifying selection rather than diversifying selection. We further analyzed the source of gene variation in other monocots, rice, barley, wheat, and maize based on the data published before. Our analysis indicated that the source of RGA diversity in these monocots was the same as in ryegrass. Thus, monocots were probably the same as dicots in the source of RGA diversity. Ryegrass RGAs in the present paper represented a large group of resistance gene homologs in monocots. We discussed the origin and the evolution of R -genes in grass species.     

Resistance and virulence interactions between two white pine species and blister rust in a 30-year field trial

We monitored infection and mortality of full-sib families of sugar pine (SP) and western white pine (WWP) selected for different mechanisms of resistance to white pine blister rust for more than 30 years in a field test in northern California. Natural infection was enhanced by interplanting alternate host Ribes spp. among test seedlings. Parents of the families were from three geographic provinces, representing the northern Rocky Mountains of Idaho (WWP), the Cascade Range of Oregon and Washington (WWP), and the Siskiyou Mountains of California and Oregon (SP). Several families of each pine species had major gene resistance (MGR), or genes for partial resistance (PR), or both types. Under the chronic epidemic conditions that prevailed, susceptible controls of both species became infected and died within a few years. Families with MGR segregated in expected Mendelian ratios (susceptibility/resistance), but each R gene—Cr1 in sugar pine and Cr2 in western white pine—was eventually defeated by biotypes of the rust with specific virulence to it (vcr1 and vcr2, respectively). WWP families from Idaho, which lacked Cr2, performed the best overall, ranging from 44 to 64% rust-free; a few WWP families from the Cascades, both with and without Cr2, performed comparably. Families with PR were not specifically vulnerable to vcr1 or vcr2. PR was less frequent in SP than WWP, and SP families had steeper infection rate curves than WWP. Combining PR in pedigrees with MGR may prolong protection of MGR by inhibiting selection and epidemic increase of vcr genes in the rust.     

Microdissection and molecular manipulation of single chromosomes in woody fruit trees with small chromosomes using pomelo (<Emphasis Type="Italic">Citrus grandis</Emphasis>) as a model. II. Cloning of resistance gene analogs from single chromosomes

Amplification of resistance gene analogs (RGAs) is both a useful method for acquiring DNA markers closely linked to disease resistance (R) genes and a potential approach for the rapid cloning of R genes in plants. However, the screening of target sequences from among the numerous amplified RGAs can be very laborious. The amplification of RGAs from specific chromosomes could greatly reduce the number of RGAs to be screened and, consequently, speed up the identification of target RGAs. We have developed two methods for amplifying RGAs from single chromosomes. Method 1 uses products of Sau3A linker adaptor-mediated PCR (LAM-PCR) from a single chromosome as the templates for RGA amplification, while Method 2 directly uses a single chromosomal DNA molecule as the template. Using a pair of degenerate primers designed on the basis of the conserved nucleotide-binding-site motifs in many R genes, RGAs were successfully amplified from single chromosomes of pomelo using both these methods. Sequencing and cluster analysis of RGA clones obtained from single chromosomes revealed the number, type and organization of R-gene clusters on the chromosomes. We suggest that Method 1 is suitable for analyzing chromosomes that are unidentifiable under a microscope, while Method 2 is more appropriate when chromosomes can be clearly identified.Communicated by P. Langridge     

Diversity and evolutionary relationship of nucleotide binding site-encoding disease-resistance gene analogues in sweet potato (<Emphasis Type="Italic">Ipomoea batatas</Emphasis> Lam.)

Most plant disease-resistance genes (R-genes) isolated so far encode proteins with a nucleotide binding site (NBS) domain and belong to a superfamily. NBS domains related to R-genes show a highly conserved backbone of an amino acid motif, which makes it possible to isolate resistance gene analogues (RGAs) by degenerate primers. Degenerate primers based on the conserved motif (P-loop and GLPL) of the NBS domain from R -genes were used to isolate RGAs from the genomic DNA of sweet potato cultivar Qingnong no.2. Five distinct clusters of RGAs (22 sequences) with the characteristic NBS representing a highly diverse sample were identified in sweet potato genomic DNA. Sequence identity among the 22 RGA nucleotide sequences ranged from 41.2% to 99.4%, while the deduced amino acid sequence identity from the 22 RGAs ranged from 20.6%to 100%. The analysis of sweet potato RGA sequences suggested mutation as the primary source of diversity. The phylogenetic analyses for RGA nucleotide sequences and deduced amino acids showed that RGAs from sweet potato were classified into two distinct groups--toll and interleukin receptor-1 (TIR)-NBS-LRR and non-TIR-NBS-LRR. The high degree of similarity between sweet potato RGAs and NBS sequences derived from R-genes cloned from tomato, tobacco, flax and potato suggest an ancestral relationship. Further studies showed that the ratio of non-synonymous to synonymous substitution within families was low. These data obtained from sweet potato suggest that the evolution of NBS-encoding sequences in sweet potato occur by the gradual accumulation of mutations leading to purifying selection and slow rates of divergence within distinct R-gene families.     

Characterisation and genetic mapping of resistance and defence gene analogs in cocoa (Theobroma cacao L.)

Disease resistance and defence gene analog (RGA/DGA) sequences were isolated in cocoa using a PCR approach with degenerate primers designed from conserved domains of plant resistance and defence genes: the NBS (nucleotide binding site) motif present in a number of resistance genes such as the tobacco N, sub-domains of plant serine/threonine kinases such as the Pto tomato gene, and conserved domains of two defence gene families: pathogenesis-related proteins (PR) of classes 2 and 5. Nucleotide identity between thirty six sequences isolated from cocoa and known resistance or defence genes varied from 58 to 80%. Amino acid sequences translated from corresponding coding sequences produced sequences without stop codons, except for one NBS –like sequence. Most of the RGAs could be mapped on the cocoa genome and three clusters of genes could be observed : NBS-like sequences clustered in two regions located on chromosomes 7 and 10, Pto-like sequences mapped in five genome regions of which one, located on chromosome 4, corresponded to a cluster of five different sequences. PR2-like sequences mapped in two regions located on chromosome 5 and 9 respectively. An enrichment of the genetic map with microsatellite markers allowed us to identify several co-localisations of RGAs, DGAs and QTL for resistance to Phytophthora detected in several progenies, particularly on chromosome 4 where a cluster of Pto-like sequences and 4 QTL for resistance to Phytophthora were observed. Many other serious diseases affect cocoa and the candidate genes, isolated in this study, could be of broader interest in cocoa disease management.     

Barley disease resistance gene analogs of the NBS-LRR class: identification and mapping

The majority of verified plant disease resistance genes isolated to date are of the NBS-LRR class, encoding proteins with a predicted nucleotide binding site (NBS) and a leucine-rich repeat (LRR) region. We took advantage of the sequence conservation in the NBS motif to clone, by PCR, gene fragments from barley representing putative disease resistance genes of this class. Over 30 different resistance gene analogs (RGAs) were isolated from the barley cultivar Regatta. These were grouped into 13 classes based on DNA sequence similarity. Actively transcribed genes were identified from all classes but one, and cDNA clones were isolated to derive the complete NBS-LRR protein sequences. Some of the NBS-LRR genes exhibited variation with respect to whether and where particular introns were spliced, as well as frequent premature polyadenylation. DNA sequences related to the majority of the barley RGAs were identified in the recently expanded public rice genomic sequence database, indicating that the rice sequence can be used to extract a large proportion of the RGAs from barley and other cereals. Using a combination of RFLP and PCR marker techniques, representatives of all barley RGA gene classes were mapped in the barley genome, to all chromosomes except 4H. A number of the RGA loci map in the vicinity of known disease resistance loci, and the association between RGA S-120 and the nematode resistance locus Ha2 on chromosome 2H was further tested by co-segregation analysis. Most of the RGA sequences reported here have not been described previously, and represent a useful resource as candidates or molecular markers for disease resistance genes in barley and other cereals.     

Origin, diversity and evolution of NBS-type disease-resistance gene homologues in coffee trees (Coffea L.)

The majority of plant disease-resistance genes (R-genes) isolated so far encode a predicted nucleotide-binding site (NBS) domain. NBS domains related to R-genes show a highly conserved backbone of amino acid motifs, which makes it possible to isolate resistance gene analogues (RGAs) by PCR with degenerate primers. Multiple combinations of primers with low degeneracy, designed from two conserved motifs in the NBS regions of R-genes of various plants, were used on genomic DNA from coffee trees, an important perennial tropical crop. Nine distinct classes of RGAs of the NBS-like type, representing a highly diverse sample, were isolated from Coffea arabica and C. canephora species. The analysis of one coffee RGA family suggested point mutations as the primary source of diversity. With one exception, coffee RGA families appeared to be closely related in sequence to at least one cloned R-gene. In addition, deduced amino acid sequences of coffee RGAs were identified that showed strong sequence similarity to almost all known non-TIR (Toll/Interleukin 1 Receptor)-type R-genes. The high degree of similarity between particular coffee RGAs and R-genes isolated from other angiosperm species, such as Arabidopsis, tomato and rice, indicates an ancestral relationship and the existence of common ancestors. The data obtained from coffee species suggests that the evolution of NBS-encoding sequences involves the gradual accumulation of mutations and slow rates of divergence within distinct R-gene families, rather than being a rapid process. Functional inferences drawn from the suggested pattern of evolution of NBS-type R-genes is also discussed.     

Isolation and linkage mapping of NBS-LRR resistance gene analogs in red raspberry (<Emphasis Type="Italic">Rubus idaeus</Emphasis> L.) and classification among 270 Rosaceae NBS-LRR genes

Plant R genes confer resistance to pathogens in a gene-for-gene mode. Seventy-five putative resistance gene analogs (RGAs) containing conserved domains were cloned from Rubus idaeus L. cv. ‘Latham’ using degenerate primers based on RGAs identified in Rosaceae species. The sequences were compared to 195 RGA sequences identified from five Rosaceae family genera. Multiple sequence alignments showed high similarity at multiple nucleotide-binding site (NBS) motifs with homology to Drosophila Toll and mammalian interleukin-1 receptor (TIR) and non-TIR RNBSA-A motifs. The TIR sequences clustered separately from the non-TIR sequences with a bootstrap value of 76%. There were 11 clusters each of TIR and non-TIR type sequences of multiple genera with bootstrap values of more than 50%, including nine with values of more than 75% and seven of more than 90%. Polymorphic sequence characterized amplified region and cleaved amplified polymorphic sequence markers were developed for nine Rubus RGA sequences with eight placed on a red raspberry genetic linkage map. Phylogenetic analysis indicated four of the mapped sequences share sequence similarity to groupTIR I, while three others were spread in non-TIR groups. Of the 75 Rubus RGA sequences analyzed, members were placed in five TIR groups and six non-TIR groups. These group classifications closely matched those in 12 of 13 studies from which these sequences were derived. The analysis of related DNA sequences within plant families elucidates the evolutionary relationship and process involved in pest resistance development in plants. This information will aid in the understanding of R genes and their proliferation within plant genomes. Electronic supplementary material  The online version of this article (doi:) contains supplementary material, which is available to authorized users.