The Journal of Microbiology (2011) Vol. 49, No. 6, pp. 935-941 Copyright ⓒ 2011, The Microbiological Society of Korea DOI 10.1007/s12275-011-1163-5 Ｏｊｋｔｚｏｌｏｉｇｚｏｕｔ２＆ Ｕｘｏｍｏｔ２＆ ｇｔｊ＆ Ｋ｜ｕｒ｛ｚｏｕｔ＆ ｕｌ＆ Ｒｋｇｌ＆ Ｔｕｊ｛ｒｇｚｏｔｍ ｭ Ｙ｡ｓｈｏｕｔｚｙ＆ ｕｌ＆ Ｙｋｘｏｉｇｔｚｎｋ＆ ．Ｘ｛ｈｏｇｉｋｇｋ／ ７ ８ ９ ： Ｈｋｔｔ｡＆ Ｒｋｓｇｏｘｋ *２＆ Ｋｒｓｇｘ＆ Ｘｕｈｈｘｋｉｎｚ ２＆ Ｈｘｇｇｓ＆ ｜ｇｔ＆ ］｡ｑ ２＆ Ｙｇｔｊｘｇ＆ ＼ｇｔ＆ Ｕｋ｜ｋｒｋｔ ２ ７ ： ７２；２＜ ８ Ｈｘｋｉｎｚ＆ ＼ｋｘｙｚｘｇｋｚｋ ２＆ Ｋｒｙ＆ Ｖｘｏｔｙｋｔ ２＆ Ｋｘｏｑ＆ Ｙｓｋｚｙ ２＆ ｇｔｊ＆ Ｙｚｋ｜ｋｔ＆ Ｊｋｙｙｋｏｔ 1 Laboratory of Plant Systematics, K.U.Leuven, Kasteelpark Arenberg 31, PO Box 2437, BE-3001 Leuven, Belgium 2 National Botanic Garden of Belgium, Domein van Bouchout, BE-1860 Meise, Belgium 3 H.G.W.J. Schweickerdt Herbarium, University of Pretoria, Pretoria 0002, South Africa 4 Laboratory of Plant Biochemistry and Physiology, University of Antwerp, Universiteitsplein 1, BE-2610 Antwerp, Belgium 5 Netherlands Centre for Biodiversity Naturalis, PO Box 9517, 2300 RA Leiden, the Netherlands 6 National Herbarium of the Netherlands, Leiden University, PO Box 9514, 2300 RA Leiden, the Netherlands (Received March 31, 2011 / Accepted July 20, 2011) Ｈｇｉｚｋｘｏｇｒ＆ ｒｋｇｌ＆ ｙ｡ｓｈｏｕｙｏｙ＆ ｏｙ＆ ｇｔ＆ ｏｔｚｏｓｇｚｋ＆ ｇｙｙｕｉｏｇｚｏｕｔ＆ ｈｋｚ｝ｋｋｔ＆ ｈｇｉｚｋｘｏｇ＆ ｇｔｊ＆ ｖｒｇｔｚｙ＆ ｏｔ＆ ｝ｎｏｉｎ＆ ｋｔｊｕｙ｡ｓｈｏｕｔｚｙ＆ ｇｘｋ＆ ｎｕ｛ｙｋｊ＆ ｝ｏｚｎｏｔ＆ ｒｋｇｌ＆ ｔｕｊ｛ｒｋｙ４＆ Ｚｎｏｙ＆ ｖｎｋｔｕｓｋｔｕｔ＆ ｎｇｙ＆ ｈｋｋｔ＆ ｘｋｖｕｘｚｋｊ＆ ｏｔ＆ ｚｎｘｋｋ＆ ｍｋｔｋｘｇ＆ ｕｌ＆ Ｘ｛ｈｏｇｉｋｇｋ＆ ．Ｖｇ｜ｋｚｚｇ２＆ Ｖｙ｡ｉｎｕｚｘｏｇ２＆ ｇｔｊ＆ Ｙｋｘｏｉｇｔｚｎｋ／２＆ ｈ｛ｚ＆ ｚｎｋ＆ ｈｇｉｚｋｘｏｇｒ＆ ｖｇｘｚｔｋｘ＆ ｎｇｙ＆ ｕｔｒ｡＆ ｈｋｋｔ＆ ｏｊｋｔｚｏｌｏｋｊ＆ ｏｔ＆ Ｖｙ｡ｉｎｕｚｘｏｇ＆ ｇｔｊ＆ Ｖｇ｜ｋｚｚｇ４＆ Ｎｋｘｋ＆ ｝ｋ＆ ｘｋｖｕｘｚ＆ ｚｎｋ＆ ｏｊｋｔｚｏｌｏｉｇｚｏｕｔ＆ ｕｌ＆ ｙ｡ｓｈｏｕｚｏｉ＆ ｈｇｉｚｋｘｏｇ＆ ｏｔ＆ ｚ｝ｕ＆ ｒｋｇｌ＆ ｔｕｊ｛ｒｇｚｏｔｍ＆ Ｙｋｘｏｉｇｔｚｎｋ＆ ｙｖｋｉｏｋｙ４＆ ［ｙｏｔｍ＆ ７＜Ｙ＆ ｘＸＴＧ＆ ｊｇｚｇ＆ ｇｔｊ＆ ｉｕｓｓｕｔ＆ ｎｕ｛ｙｋｑｋｋｖｏｔｍ＆ ｍｋｔｋｚｏｉ＆ ｓｇｘｑｋｘｙ＆ ．ｘｋｉＧ＆ ｇｔｊ＆ ｍ｡ｘＨ／＆ ｝ｋ＆ ｙｚ｛ｊｏｋｊ＆ ｚｎｋ＆ ｖｎ｡ｒｕｍｋｔｋｚｏｉ＆ ｘｋｒｇｚｏｕｔｙｎｏｖｙ＆ ｕｌ＆ ｈｇｉｚｋｘｏｇｒ＆ ｋｔｊｕｙ｡ｓｈｏｕｔｚｙ＆ ｏｔ＆ Ｘ｛ｈｏｇｉｋｇｋ４＆ Ｋｔｊｕｙ｡ｓｈｏｕｔｚｙ＆ ｕｌ＆ ｒｋｇｌ＆ ｔｕｊ｛ｒｇｚｏｔｍ＆ Ｘ｛ｈｏｇｉｋｇｋ＆ ｝ｋｘｋ＆ ｌｕ｛ｔｊ＆ ｚｕ＆ ｈｋ＆ ｉｒｕｙｋｒ｡＆ ｘｋｒｇｚｋｊ＆ ｇｔｊ＆ ｝ｋｘｋ＆ ｖｒｇｉｋｊ＆ ｇｙ＆ ｇ＆ ｓｕｔｕｖｎ｡ｒｋｚｏｉ＆ ｍｘｕ｛ｖ＆ ｝ｏｚｎｏｔ＆ ｚｎｋ＆ ｍｋｔ｛ｙ＆ Ｈ｛ｘｑｎｕｒｊｋｘｏｇ＆ ．β３Ｖｘｕｚｋｕｈｇｉｚｋｘｏｇ／４＆ Ｚｎｋ＆ ｖｎ｡ｒｕｍｋｔｋｚｏｉ＆ ｇｔｇｒ｡ｙｋｙ＆ ｘｋ｜ｋｇｒｋｊ＆ ｇ＆ ｖｇｚｚｋｘｔ＆ ｕｌ＆ ｙｚｘｏｉｚ＆ ｎｕｙｚ＆ ｙｖｋｉｏｌｏｉｏｚ｡＆ ｇｔｊ＆ ｖｒｇｉｋｊ＆ ｚｎｋ＆ ｚ｝ｕ＆ ｏｔ｜ｋｙｚｏｍｇｚｋｊ＆ ｋｔｊｕｙ｡ｓｈｏｕｔｚｙ＆ ｇｚ＆ ｚ｝ｕ＆ ｊｏｙｚｏｔｉｚ＆ ｖｕｙｏｚｏｕｔｙ＆ ｏｔ＆ ｚｎｋ＆ ｚｕｖｕｒｕｍ｡＆ ｕｌ＆ ｚｎｋ＆ ｚｘｋｋ２＆ ｙ｛ｍｍｋｙｚｏｔｍ＆ ｇｚ＆ ｒｋｇｙｚ＆ ｚ｝ｕ＆ ｊｏｌｌｋｘｋｔｚ＆ ｋ｜ｕｒ｛ｚｏｕｔｇｘ｡＆ ｕｘｏｍｏｔｙ４＆ Ｚｎｋ＆ ｊｋｍｘｋｋ＆ ｕｌ＆ ｙｋｗ｛ｋｔｉｋ＆ ｊｏ｜ｋｘｍｋｔｉｋ＆ ｈｋｚ｝ｋｋｔ＆ ｚｎｋ＆ Ｙｋｘｏｉｇｔｚｎｋ＆ ｋｔｊｕｙ｡ｓｈｏｕｔｚｙ＆ ｇｔｊ＆ ｚｎｋｏｘ＆ ｘｋｒｇｚｏ｜ｋｙ＆ ｝ｇｙ＆ ｒｇｘｍｋ＆ ｋｔｕ｛ｍｎ＆ ｚｕ＆ ｖｘｕｖｕｙｋ＆ ｚｎｋ＆ Ｙｋｘｏｉｇｔｚｎｋ＆ ｋｔｊｕｙ｡ｓｈｏｕｔｚｙ＆ ｇｙ＆ ｔｋ｝＆ ｙｖｋｉｏｋｙ＆ ． Ｉｇｔｊｏｊｇｚ｛ｙ＆ Ｈ｛ｘｑｎｕｒｊｋｘｏｇ＆ ｇｔｊｕｔｍｋｔｙｏｙ ＆ ｇｔｊ＆ Ｉｇｔｊｏｊｇｚ｛ｙ＆ Ｈ｛ｘｑｎｕｒｊｋｘｏｇ＆ ｖｋｚｏｚｏｏ ／４＆ Ｏｔ＆ ｇ＆ ｙｋｉｕｔｊ＆ ｖｇｘｚ＆ ｕｌ＆ ｚｎｏｙ＆ ｙｚ｛ｊ｡２＆ ｚｎｋ＆ ｖ｡ｒｕｍｋ３ ｔｋｚｏｉ＆ ｘｋｒｇｚｏｕｔｙｎｏｖｙ＆ ｇｓｕｔｍ＆ ｔｕｊ｛ｒｇｚｏｔｍ＆ ｇｔｊ＆ ｔｕｔ３ｔｕｊ｛ｒｇｚｏｔｍ＆ Ｙｋｘｏｉｇｔｚｎｋ＆ ｙｖｋｉｏｋｙ＆ ｝ｋｘｋ＆ ｏｔ｜ｋｙｚｏｍｇｚｋｊ＆ ｛ｙｏｔｍ＆ ｙｋｗ｛ｋｔｉｋ＆ ｊｇｚｇ＆ ｌｘｕｓ＆ ｙｏ～＆ ｉｎｒｕｘｕｖｒｇｙｚ＆ ｘｋｍｏｕｔｙ＆ ．ｘｖｙ７＜２＆ ｚｘｔＭ２＆ ｚｘｔＲ３ｚｘｔＬ２＆ ｖｋｚＪ２＆ ｖｋｚＧ３ｖｙｈＰ２＆ ｇｔｊ＆ ｇｚｖＯ３ｇｚｖＮ／４＆ Ｕ｜ｋｘｇｒｒ２＆ ｍｋｔｋｚｏｉ＆ ｜ｇｘｏｇｚｏｕｔ＆ ｇｓｕｔｍ＆ ｚｎｋ＆ ｖｒｇｙｚｏｊ＆ ｓｇｘｑｋｘｙ＆ ｝ｇｙ＆ ｏｔｙ｛ｌｌｏｉｏｋｔｚ＆ ｚｕ＆ ｋｔｇｈｒｋ＆ ｖｎ｡ｒｕｍｋｔｋｚｏｉ＆ ｋｙｚｏｓｇｚｏｕｔ４＆ Ｎｕ｝ｋ｜ｋｘ２＆ ｕ｛ｘ＆ ｘｋｙ｛ｒｚｙ＆ ｉｕ｛ｒｊ＆ ｔｕｚ＆ ｘ｛ｒｋ＆ ｕ｛ｚ＆ ｚｎｋ＆ ｖｕｙｙｏｈｏｒｏｚ｡＆ ｚｎｇｚ＆ ｈｇｉｚｋｘｏｇｒ＆ ｒｋｇｌ＆ ｙ｡ｓｈｏｕｙｏｙ＆ ｕｘｏｍｏｔｇｚｋｊ＆ ｕｔｉｋ＆ ｏｔ＆ ｇ＆ ｉｕｓｓｕｔ＆ ｇｔｉｋｙｚｕｘ＆ ｕｌ＆ ｚｎｋ＆ Ｙｋｘｏｉｇｔｚｎｋ＆ ｙｖｋｉｏｋｙ４ Ｑｋ｡｝ｕｘｊｙ＠ Burkholderia, endosymbionts, bacterial leaf nodulation, Sericanthe, Rubiaceae About 500 species of Rubiaceae are known to house bacterial endosymbionts within internal cavities in the leaf lamina, referred to as bacterial leaf nodules or leaf galls (Miller, 1990). Endosymbionts are persistent and obligate associates of the host plants and are required for the successful development and reproduction of their hosts (Gordon, 1963; Miller, 1990). However, knowledge about the exact benefits conferred by these endosymbionts is still incomplete. Many studies have proposed that the endophytes of nodulated species are involved in the production of phytohormones (reviewed in Miller, 1990). From the endosymbiont’s perspective, the colonization of internal plant tissues may provide a stable, uniform, and protective environment. Leaf nodulated plant species are limited to three distantly related genera: Pavetta L., Psychotria L., and Sericanthe Robbr. These three genera of the Rubiaceae family have no close phylogenetic affinity and belong to distant alliances. Psychotria is a member of the subfamily Rubioideae and belongs to the ✽ For correspondence. E-mail: Benny.Lemaire@bio.kuleuven.be; Tel.: 003216-32-86-36; Fax: 0032-16-32-19-55 § Supplemental material for this article may be found at http://www.springer.com/content/120956 tribe Psychotrieae. Pavetta and Sericanthe belong to the subfamily Cinchonoideae and have been placed in the tribes Pavetteae and Coffeeae, respectively (Robbrecht and Manen, 2006). Morphological observations of bacterial endosymbionts have been conducted in all rubiaceous genera (Pavetta and Psychotria in Miller, 1990; Sericanthe in Van Hove, 1972, referred to by the author as ‘Neorosea’). Still, attempts to cultivate and characterize these leaf nodulating bacteria associated have been unsuccessful to date. Molecular sequencing analyses however now enable the identification of uncultivable endosymbionts. Indeed, the taxonomic position of the endosymbionts of Pavetta and Psychotria has been recently clarified (van Oevelen et al., 2001, 2002, 2004; Lemaire et al., 2011). These studies have demonstrated that every nodulating species hosts its own unique Burkholderia endosymbiont. In contrast, the bacterial leaf endosymbionts within the genus Sericanthe remain unknown. The genus Sericanthe is composed mostly by shrubs that occupy rain forests, woodlands, savannas and (sub)montane habitats in Southern and Western Africa. Many Sericanthe species have a very restricted distribution, as reflected by their 936 Lemaire et al. Ｚｇｈｒｋ＆ ７４＆ Taxon accession data with herbarium vouchers, silica-gel collections, (geographical) origins and GenBank accession numbers of leaf nodulated endosymbionts and host plants. Specimens were obtained from the National Botanic Garden of Belgium (BR). Underlined taxa represent accessions that were newly sequenced for this study. Taxon Burkholderia ambifaria Burkholderia caribensis Burkholderia cepacia Burkholderia fungorum Burkholderia gladioli Burkholderia glathei Burkholderia glathei Burkholderia graminis Burkholderia kururiensis Burkholderia multivorans Burkholderia oklahomensis Burkholderia plantarii Burkholderia stabilis Burkholderia tropica Burkholderia tuberum Burkholderia vietnamiensis Candidatus Burkholderia andongensis Candidatus Burkholderia andongensis Candidatus Burkholderia andongensis Candidatus Burkholderia andongensis Candidatus Burkholderia andongensis Candidatus Burkholderia calva Candidatus Burkholderia calva Candidatus Burkholderia hispidae Candidatus Burkholderia hispidae Candidatus Burkholderia kirkii Candidatus Burkholderia kirkii Candidatus Burkholderia nigropunctata Candidatus Burkholderia nigropunctata Candidatus Burkholderia petitii Candidatus Burkholderia petitii Candidatus Burkholderia rigidae Candidatus Burkholderia rigidae Candidatus Burkholderia schumannianae Candidatus Burkholderia schumannianae Ralstonia pickettii Strain/Voucher LMG 19182 LMG 18531 LMG 1223 LMG 16225 LMG 11626 LMG 14190 LMG 14190 LMG 18924 LMG 19447 LMG 13010 LMG 23618 LMG 9035 LMG 14294 LMG 22274 LMG 21444 LMG 10929 BL 259 (BR) BL 271 (BR) BL 286 (BR) BL 293 (BR) SD 1097 (BR) 1962-0512 (BR) 1964-0306 (BR) SD 3176 (BR) OL 732 (BR) 1953-6779 (BR) 2000-1946-61 (BR) PS 13 (BR) SD 1849 (BR) SD 1512 (BR) OL 658 (BR) OL 694 (BR) OL 877 (BR) SD 1099 (BR) 2001-9442-57 (BR) 12J Origin Pea, rhizosphere Vertisol Allium cepa Phanerochaete chrysosporium, fungus Poisoned bongkrek Lateritic soil Lateritic soil Maize senescent root system Aquifer sample Cystic fibrosis patient Soil Oryza sativa, seedling Cystic fibrosis, patient Sugarcane, roots Aspalathus carnosa, root nodule Oryza sativa, rhizosphere soil Sericanthe andongensis (Hiern) Robbr., leaf nodules; South Africa, Louis Trichardt Sericanthe andongensis (Hiern) Robbr., leaf nodules; South Africa, Vhembe Sericanthe andongensis (Hiern) Robbr., leaf nodules; South Africa, Vhembe Sericanthe andongensis (Hiern) Robbr., leaf nodules; South Africa, Tathe Vondo Sericanthe andongensis (Hiern) Robbr., leaf nodules; Zambia Psychotria calva Hiern, leaf nodules; Unknown GenBank accession number 16S rRNA recA gyrB HQ849072 HQ849130 HQ849186 HQ849077 HQ849135 HQ849190 HQ849078 JF295011 HQ849191 HQ849081 HQ849138 HQ849194 HQ849082 HQ849139 HQ849195 U96935 AY619666 EU024198 HQ849084 HQ849141 HQ849197 HQ849086 HQ849143 HQ849199 HQ849088 HQ849145 HQ849201 HQ849090 HQ849203 HQ849092 HQ849148 HQ849205 HQ849098 HQ849153 HQ849210 HQ849103 HQ849159 JF295010 HQ849105 HQ849161 HQ849216 HQ849106 HQ849162 HQ849217 HQ849107 HQ849163 HQ849218 JF916912 JF916907 JF916918 JF916913 JF916908 JF916919 - JF916906 JF916920 JF916914 JF916909 JF916921 JF916915 JF916905 HQ849116 HQ849172 JF295009 Psychotria calva Hiern, leaf nodules; Ivory HQ849117 Coast Pavetta hispida Hiern, leaf nodules; Cameroon, HQ849122 Ebolowa Pavetta hispida Hiern, leaf nodules; Cameroon, HQ849123 Efoulan Psychotria kirkii Hiern, leaf nodules; Unknown HQ849109 HQ849173 HQ849227 HQ849178 HQ849231 HQ849179 HQ849232 HQ849165 HQ849220 Psychotria kirkii Hiern, leaf nodules; D.R. Congo, Kantanga Psychotria nigropunctata Hiern; D.R. Congo, Kisantu Psychotria nigropunctata Hiern, leaf nodules; Gabon, Bemboudié Sericanthe aff. petitii (N.Hallé) Robbr., leaf nodules; Cameroon, Mbikiliki Sericanthe aff. petitii (N.Hallé) Robbr., leaf nodules; Cameroon, Efoulan Pavetta rigida Hiern, leaf nodules; Cameroon, Efoulan Pavetta rigida Hiern, leaf nodules; Cameroon, Nkolakié Pavetta schumanniana F.Hoffm.ex K.Schum., leaf nodules; South Africa Pavetta schumanniana F.Hoffm.ex K.Schum., leaf nodules; D.R. Congo HQ849110 HQ849166 HQ849221 HQ849118 HQ849174 HQ849228 HQ849119 HQ849175 JF295008 JF916923 JF916916 JF916911 JF916922 JF916917 JF916910 HQ849120 HQ849176 HQ849229 HQ849121 HQ849177 HQ849230 HQ849124 HQ849180 HQ849233 HQ849126 HQ849182 HQ849235 NC010678 NC010682 NC010682 Leaf nodulating endosymbionts of Sericanthe 937 Ｚｇｈｒｋ＆ ７４ Continued Taxon Strain/Voucher Origin GenBank accession number trnL-trnF trnG petD petA-psbJ JF916964 JF916953 JF916975 JF916931 atpI-atpH JF916924 Coffea stenophylla G.Don 1937-0053 (BR) rps16 D.R. Congo JF916942 Empogona kirkii Hook.f. Sericanthe andongensis (Hiern) Robbr. Sericanthe andongensis (Hiern) Robbr. Sericanthe auriculata (Keay) Robbr. Sericanthe auriculata (Keay) Robbr. Sericanthe odoratissima (K.Schum.) Robbr. Sericanthe odoratissima (K.Schum.) Robbr. Sericanthe petitii (N.Hallé) Robbr. Sericanthe petitii (N.Hallé) Robbr. Sericanthe spec. nov. 1976-1052 (BR) SD 1097 (BR) Zimbabwe Zambia JF916943 JF916944 JF916965 JF916966 JF916954 JF916955 JF916976 JF916977 JF916932 JF916933 JF916925 - Chapman 6150 (BR) Malawi JF916945 JF916967 JF916956 JF916978 JF916934 - SD 1467 (BR) Cameroon JF916946 JF916968 JF916957 JF916979 JF916935 JF916926 SD 1516 (BR) Cameroon JF916947 JF916969 JF916958 JF916980 JF916936 JF916927 Polhill et al. 5007A (BR) Tanzania JF916948 JF916970 JF916959 - JF916937 - Salubeni 3135 (BR) Malawi JF916949 JF916971 JF916960 JF916981 JF916938 - SD 1512 (BR) Cameroon JF916950 JF916972 JF916961 JF916982 JF916939 JF916928 OL 658 (BR) Cameroon JF916951 JF916973 JF916962 JF916983 JF916940 JF916929 SD 2608 (BR) Cameroon JF916952 JF916974 JF916963 JF916984 JF916941 JF916930 rarity and infrequent collection (Robbrecht, 1978a). For a complete description of the geographical distribution of all Sericanthe species see Robbrecht (1978b). In Sericanthe, leaf nodules have been reported in 11 or 12 out of 17 species (Robbrecht, 1978a). This small nodulating genus contrasts with the more specious genera Pavetta and Psychotria, which contain approximately 350 and 80 nodulating species, respectively. Bacterial leaf galls of nodulated Sericanthe species are always located on the abaxial side of the leaf and are hardly visible from the adaxial side. The shape and distribution of the nodules on the leaves differ substantially among species and range from linear galls along the mid-vein (e.g. Sericanthe andongensis) to dot-shaped or branched nodules scattered in the leaf blade (e.g. Sericanthe petitii) (Robbrecht, 1978a, 1981). A similar variation in nodule localization and morphology has been reported in Pavetta and Psychotria (Bremekamp, 1933). In the present study, we focus on the endosymbiont identification and evolutionary history of bacterial leaf symbiosis in the genus Sericanthe. We propose the hypotheses that (1) leaf nodulated Sericanthe species accommodate their own specific endosymbionts and that (2) the bacteria-plant interaction is the result of an ancient and single infection event within an ancestral leaf nodulated Sericanthe host. Ｓｇｚｋｘｏｇｒ＆ ｇｔｊ＆ Ｓｋｚｎｕｊｙ Ｚｇ～ｕｔ＆ ｙｇｓｖｒｏｔｍ Silica-dried material from S. andongensis and S. petitii were collected during botanical field expeditions in South Africa, Cameroon and Zambia. Five accessions of S. andongensis and two accessions of S. petitii were sampled from different regions in the field and were used to identify the bacterial endosymbionts. A detailed list of sampled species, voucher information and localities is given in Table 1. To determine the phylogenetic position of the endosymbionts of Sericanthe, we included related bacterial sequences of Burkholderia obtained from GenBank (Table 1). Three additional Sericanthe species (i.e. S. auriculata, S. odoratissima, S. spec. nov.), the latter two of which were collected in the field and the first of which was obtained from an herbarium sample at the National Botanic Garden of Belgium (BR), were included in this study to construct the host phylogeny (Table 1). ＪＴＧ＆ ｋ～ｚｘｇｉｚｏｕｔ２＆ ｇｓｖｒｏｌｏｉｇｚｏｕｔ２＆ ｉｒｕｔｏｔｍ２＆ ｇｔｊ＆ ｙｋｗ｛ｋｔｉｏｔｍ Before extraction of the bacterial DNA the silica-dried leaves were rinsed with 70% ethanol to avoid bacterial contamination. Total DNA was extracted from silica-dried collections and herbarium specimens (BR) using the modified CTAB protocol of Tel-Zur et al. (1999). Bacterial DNA (16S rRNA, recA and gyrB) and host chloroplast DNA regions (rps16, trnL-trnF, trnG, petD, petA-psbJ, and atpI-atpH) were amplified with the primers listed in Supplementary data Table 1. All amplification reactions were performed using a GeneAmp PCR System 9700 (Applied Biosystems, USA). Each amplification reaction was performed in 25 μl reaction mix containing 1 μl total DNA, 16 μl H20, 2.5 μl 10× PCR buffer, 0.75 μl 25 μM MgCl2, 1 μl of 20 μM forward and reverse primers, 2.5 μl 2 μM dNTP, and 0.2 μl Taq DNA polymerase. PCR amplification of endosymbiont DNA regions was performed with PCR parameters as described previously (Lemaire et al., 2011). Amplification of rps16, trnL-trnF, trnG, petA-psbJ, and atpI-atpH was carried out under the following conditions: 94°C for 3 min; 30 cycles at 94°C for 60 sec, 52°C for 60 sec, 72°C for 90 sec; final extension at 72°C for 5 min. The amplification parameters for petD were 94°C for 3 min; 30 cycles at 94°C for 60 sec, 55°C for 60 sec, 72°C for 90 sec; final extension at 72°C for 5 min. Amplified products were purified using a modification of the Exo/ SAP enzyme cleaning protocol (Werle et al., 1994). Amplified 16S rRNA products were cloned into a pGEM-T vector (Promega), according to the manufacturer’s instructions, and transformed into JM109 E. coli by heat shock. Plasmid purification was TM obtained with a PureYield Plasmid Miniprep System (Promega). 938 Lemaire et al. Purified plasmids and PCR products were sent to Macrogen for sequencing (Macrogen Inc., Korea). Ｖｎ｡ｒｕｍｋｔｋｚｏｉ＆ ｇｔｇｒ｡ｙｋｙ Sequences were assembled and edited using the program Geneious 5.0.3 (http://www.geneious.com). A preliminary sequence alignment was created with Muscle (Edgar, 2004) followed by manual adjustments with MacClade 4.04 (Maddison and Maddison, 2001). Molecular data were analyzed using Maximum Likelhood (ML) and Bayesian Inference (BI) criteria, both of which were implemented in the CIPRES web portal (http:// www.phylo.org). ML analyses were performed with RAxML-VI-HPC v2.0 using GTR-GAMMA as the nucleotide substitution model (Stamatakis, 2006). We performed 100 RAxML runs and selected the best ML tree by comparing the likelihood scores. The robustness of the ML tree was calculated with multi-parametric bootstrap resampling and 1000 pseudo-replicates. Model selection for the Bayesian analysis was conducted with MrModeltest v. 3.06 (Posada and Crandall, 1998) under the Akaike information criterion. For the different datasets, Modeltest selected the following models of evolution: petD – GTR; rps16 – GTR+I; trnG – F81; trnL-trnF – HKY; petA-psbJ – HKY; atpI-atpH – GTR; 16S rRNA – GTR+I+G; recA – GTR+I+G; gyrB – GTR+I+G. Ｌｏｍ４＆ ７４＆ Phylogenetic tree of bacterial endosymbionts based on 16S rRNA, recA and gyrB data. Support values for the Bayesian and Maximum Likelihood analyses are given at the nodes (Bayesian posterior probabilities-bootstrap values from the Maximum Likelihood analysis). Branches of leaf nodulating endosymbionts are shown in bold. Names of newly proposed bacterial taxa are shown in bold. Leaf nodulating endosymbionts of Sericanthe In the combined BI analyses, the concatenated datasets (petD + rps16 + trnG + trnL-trnF + petA-psbJ + atpI-atpH and 16S rRNA + recA + gyrB) were partitioned and the same models were assigned to the separate partitions as selected for the single analyses. Gaps in the chloroplast data were coded according to the simple indel coding method described by Simmons and Ochoterena (2000). Bayesian analyses were conducted with MrBayes 3.1 (Huelsenbeck and Ronquist, 2001; Ronquist and Huelsenbeck, 2003), and four Markov chains were ran simultaneously for five million generations and sampling trees every 100 generations. The 25% initial trees were discarded as conservative “burnin”. Convergence of the chains was checked using Tracer v.1.4 (Rambaut and Drummond, 2007). Ｓｕｘｖｎｕｒｕｍｏｉｇｒ＆ ｕｈｙｋｘ｜ｇｚｏｕｔｙ Morphological observations of leaf material of S. andongensis (accession: Lemaire et al. 293) and S. petitii (accession: Lachenaud et al. 658) were conducted to illustrate the bacterial endosymbionts. Sections through leaf nodules were made with a razor blade and the dissected material was washed repeatedly in 70% ethanol and dehydrated in a 1:1 mixture of ethanol and dimethoxymethan (DMM) for 20 min and in pure DMM for 20 min. After critical-point drying (CPD 030, BAL-TEC AG, Liechtenstein), dried samples were mounted onto aluminium stubs, coated with gold (SPI Module Sputter Coater, Spi Supplies, USA) and observed under a scanning electron microscope (JEOL JSM-6360; Jeol Ltd, Japan). Ｘｋｙ｛ｒｚｙ＆ ｇｔｊ＆ Ｊｏｙｉ｛ｙｙｏｕｔ Ｖｎ｡ｒｕｍｋｔｋｚｏｉ＆ ｇｔｇｒ｡ｙｋｙ＆ ｕｌ＆ ｚｎｋ＆ ｋｔｊｕｙ｡ｓｈｏｕｔｚ＆ ｊｇｚｇ The use of 16S rRNA, recA and gyrB data to infer the phylogenetic relationships in the genus Burkholderia is quite common and offers high resolution at both high and low taxonomic levels (Payne et al., 2005; Tabacchioni et al., 2008). These genes have been shown to provide a robust framework to determine the phylogenetic placement of the symbiotic bacteria of leaf nodulating Rubiaceae, as previously described in Lemaire et al. (2011). In the present study, a similar approach was used to identify the endosymbionts of nodulating Sericanthe species: molecular identification of endosymbionts was first performed by 16S rRNA sequencing and 16S rRNA BLAST searches, and recA and gyrB genes were used to increase the relative discriminatory power. Direct sequencing of full-length 16S rRNA from 10 clones per plant species produced consistent results and assigned the endosymbionts of the leaf nodulating Sericanthe species (S. andongensis and S. petitii) in the Burkholderia genus. This β-Proteobacteria genus also includes the endosymbionts of the two other nodulating genera of the family, Psychotria and Pavetta. Amplified recA and gyrB data with Burkholderia specific primers (Supplementary data Table 1) were analyzed in combination with the 16S rRNA data, including 19 nodulated endosymbionts and 14 related Burkholderia strains. The phylogenetic analyses of the separate datasets showed similar topologies, except for few terminal branches. The phylogenies produced separately by the three datasets (16S rRNA, recA and gyrB) are shown in supplementary data Fig. 1. Both the BI and ML analyses produced similar tree topologies and support values (Fig. 1). A well-supported clade (100% Bayesian posterior probability, BPP / 99% bootstrap support, BS) with endosymbionts of leaf nodulating Psychotria, 939 Pavetta, and Sericanthe plants was recovered as sister to Burkholderia glathei. All S. andongensis endosymbionts were positioned in a clade with maximum support that was sister to the endosymbionts of Pavetta rigida and Pavetta hispida. The intersequence similarities between the lineages from both clades ranged from 96% (Candidatus Burkholderia andongensis vs. Candidatus Burkholderia hispidae) to 96.5% (Candidatus Burkholderia andongensis vs. Candidatus Burkholderia rigidae). The endosymbionts of S. petitii (100% BPP / 100% BS) were related to the ‘Candidatus Burkholderia kirkii – Candidatus Burkholderia schumannianae’ clade. The sequence divergence between both nodulating clades ranged from 94.6% (Candidatus Burkholderia petitii vs Candidatus Burkholderia kirkii) to 95.0% (Candidatus Burkholderia petitii vs Candidatus Burkholderia schumannianae). These phylogenetic patterns indicate that endosymbiosis occurred multiple times in Rubiaceae, thus rejecting the hypothesis of a single infection event within the ancestor of extant leaf nodulated Pavetta, Psychotria, and Sericanthe species. Five different samples of S. andongensis and two accessions of Sericanthe petitii from different geographical locations were investigated (Table 1). Intraspecific sequence variability among the endosymbiont strains of both species was low (average sequence identity between S. andongensis accessions: 16S rRNA - 100%; recA - 100%; gyrB - 99.9% and S. petitii accessions: 16S rRNA - 100%; recA - 99.8%; gyrB - 100%), suggesting a stable interaction and high specificity between host and endosymbiont. A similar pattern of host specificity has been documented in Psychotria and Pavetta (van Oevelen et al., 2001, 2002, 2004; Lemaire et al., 2011). The phylogenetic analyses presented in this study show that the evolutionary distances between the Sericanthe endosymbionts and their closest relatives were significant compared to the observed intraspecific polymorphism to recognize these endosymbionts as novel Burkholderia species. As long as the cultivation of Sericanthe endosymbionts is not possible (E. Prinsen 2011, pers. comm.), we propose to record these endosymbionts under a Candidatus designation, according to Murray and Stockebrandt (1995). The endosymbionts of S. andongensis and S. petitii can be described using the specific epithets of their host species as specific epithets for these candidate Burkholderia species: ‘Candidatus Burkholderia andongensis’ (andongensis, from the specific epithet of the host plant) (β-proteobacteria, genus Burkholderia); NC; G-; R; NAS (GenBank nos. JF916921, JF916915, JF916905), oligonucleotide sequence complementary to unique region of 16S rRNA gene 5′-ACTTCGTCCCTAATA ATGGATGGAG-3′, oligonucleotide sequence complementary to unique region of recA 5′-CGCGTTCATCGATGCCGAAC ACGCGCTC-3′, oligonucleotide sequence complementary to unique region of gyrB gene 5′-TCGCACGGCGTCGTGCAG AACCGTGAAGT-3′; S (S. andongensis, leaf galls). Lemaire et al. this study. ‘Candidatus Burkholderia petitii’ (petitii, from the specific epithet of the host plant) (β-proteobacteria, genus Burkholderia); NC; G-; R; NAS (GenBank nos. JF916923, JF916916, JF916911), oligonucleotide sequence complementary to unique region of 16S rRNA gene 5′-GCTTCGGGGTTAATACCCCT GGGG-3′, oligonucleotide sequence complementary to unique region of recA 5′-ACGTGCAATACGCCTCGAAGCTTGGC GTGAACGTGCCGGAT-3′, oligonucleotide sequence com- 940 Lemaire et al. and Pavetta [see previous observations in the study of van Oevelen et al. (2004) and Lemaire et al. (2011)]. Ｖｎ｡ｒｕｍｋｔｋｚｏｉ＆ ｇｔｇｒ｡ｙｋｙ＆ ｕｌ＆ ｎｕｙｚｙ Ｌｏｍ４＆ ８４＆ SEM photographs of leaf nodulating endosymbionts of (A) S. petitii and (B) S. andongensis. Non-flagellated rod-shaped bacteria with a mean length of 2 μM are visible. plementary to unique region of gyrB gene 5′-ATGGAGTTC GCGCGTGGAGTCGTGCAGAACCGC-3′; S (S. petitii, leaf galls). Lemaire et al. this study. Ｓｕｘｖｎｕｒｕｍｏｉｇｒ＆ ｕｈｙｋｘ｜ｇｚｏｕｔｙ＆ ｕｌ＆ ｒｋｇｌ＆ ｔｕｊ｛ｒｇｚｏｔｍ＆ ｋｔｊｕ３ ｙ｡ｓｈｏｕｔｚｙ The phylogentic analyses showed that leaf nodulated Sericanthe species accommodate a single species-specific endosymbiont. As a result, we were able to use the non-specific scanning electron microscopy to illustrate the endosymbionts in leaf nodule structures. The bacterial endosymbionts within leaf nodules of S. andongensis and S. petitii are shown in Figs. 2A and B. Cross SEM sections of leaves were made to illustrate the bacterial morphology and the localization of the endosymbionts within nodules. The endosymbionts were restricted to the leaf gall structures and were clearly visible as rod shaped bacteria with an average length of 2 μm. No flagella were observed. The endosymbionts of Sericanthe were similar in size (1-2 μm) and shape (bacterial rods) compared to the symbionts of Psychotria To reconstruct the phylogenetic relationships between nodulated and non-nodulated Sericanthe species, 66 sequences were generated including six chloroplast regions (Table 1). Genetic variation among all chloroplast DNA regions was extremely low, ranging from 0.8% to 3.5% of variable sites (Supplementary data Table 2). In contrast, the alignment of the 16S rRNA, recA and gyrB sequences revealed higher levels of genetic variability. This difference in sequence variability between plants and bacteria is probably linked to different rates of molecular evolution associated with differences in body size, metabolic rate, DNA repair and generation time (Bromham, 2009). The phylogenetic relationships obtained from the six individual plastid markers were analyzed separately, and the resulting tree topologies were phylogenetically consistent. Consequently, the datasets were combined in subsequent analyses to increase phylogenetic resolution. Indels were binary coded and added to data matrices to increase support values. The Bayesian majority rule consensus tree and the Maximum Likelihood tree were congruent and are shown in Fig. 3. Overall, most phylogenetic relationships were resolved with high support values. However, the phylogenetic relationships between the nodulating Sericanthe species (showed in bold) and non-nodulating species were not completely resolved, showing a polytomy with members of S. andongensis, S. odoratissima, S. petitii, and S. auriculata. All nucleotide positions within the alignment were examined by eye and no single character was informative to resolve this node. Nevertheless, the observed phylogenetic relationships in this study do not rule out the possibility that bacterial endosymbiosis evolved in a parsimonious way, as demonstrated for other Ｌｏｍ４＆ ９４＆ Phylogenetic tree of hosts based on chloroplast data (rps16, trnG, trnL-trnF, petD, petA-psbJ, and atpI-atpH). Support values for the Bayesian and Maximum Likelihood analyses are given at the nodes (Bayesian posterior probabilities - bootstrap values from the Maximum Likelihood analysis). Branches of leaf nodulated representatives are shown in bold. Leaves with leaf nodules are redrawn from Robbrecht (1978a). Top: leaf galls located along the midvein (S. andongensis var. andongensis). Bottom: leaf galls dispersed over the leaf blade (S. petitii). Leaf nodulating endosymbionts of Sericanthe nodulated genera, i.e. Psychotria (Andersson, 2002) and Pavetta (de Block et al. unpublished). Ｉｕｔｉｒ｛ｙｏｕｔｙ The three nodulating genera have no close affinity and have been placed within different tribes and Rubiaceae subfamilies, which could lead to the conclusion that bacterial leaf nodule symbiosis originated independently in these three genera. Surprisingly, our results demonstrate that all endosymbionts of leaf nodulating Rubiaceae are closely related, but that neither the endosymbionts of Sericanthe nor the endosymbionts of Pavetta or Psychotria are monophyletic. These findings contrast with previous results showing that these three nodulating taxa are monophyletic (Andersson, 2002; Davis et al., 2007; Tosh et al., 2009; de Block et al., unpublished). The present results suggest thereby that the history of bacterial leaf symbiosis is characterized by horizontal symbiont transfers and reject the hypothesis of strict co-speciation between plant and bacteria at generic level. Ｇｉｑｔｕ｝ｒｋｊｍｋｓｋｔｚｙ The authors thank Elsa van Wyk and Magda Nel for their excellent support at the H.G.W.J. Schweickerdt Herbarium, Department of Plant Science, University of Pretoria (South Africa). We are also grateful to Norbert Hahn, who accompanied us during the expedition in South Africa. We thank the King Léopold III Fund and the Fund for Scientific Research – Flanders (FWO) which provided financial support for fieldwork in South Africa. This work was supported by the Institute for the Promotion of Innovation by Science and Technology in Flanders (IWT Vlaanderen, no. 71488). General financial support was provided by a grant of the Research Program of the Fund for Scientific Research – Flanders (Belgium) (FWO – Vlaanderen, G.0343.09N) and the K.U. Leuven (OT/05/35). Ｘｋｌｋｘｋｔｉｋｙ Andersson, L. 2002. Relationships and generic circumscriptions in the Psychotria complex (Rubiaceae, Psychotrieae). Syst. Geogr. Pl. 72, 167-202. Bremekamp, C.E.B. 1933. The bacteriophilous species of Psychotria. Lond. J. Bot. 71, 271-281. Bromham, L. 2009. Why do species vary in their rate of molecular evolution? Biol. Lett. 5, 401-404. Davis, A.P., M. Chester, O. Maurin, and M. Fay. 2007. Searching for the relatives of Coffea (Rubiaceae, Ixoroideae): the circumscription and phylogeny of Coffeeae based on plastid sequence data and morphology. Am. J. Bot. 94, 313-329. Edgar, R.C. 2004. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res. 32, 1792-1797. Gordon, J.F. 1963. The nature and distribution within the plant of the bacteria associated with certain leaf-nodulated species of the families Myrsinaceae and Rubiaceae, p. 370. Thesis. Univ. London. Huelsenbeck, J.P. and F. Ronquist. 2001. MRBAYES: Bayesian inference of phylogeny. Bioinformatics 17, 754-755. Lemaire B., P. de Block, S. van Oevelen, B. Verstraete, E. Smets, E. Prinsen, and S. Dessein. 2011. Identification of the bacterial endosymbionts in leaf nodulating Pavetta species (Rubiaceae). Int. J. Syst. Evol. Microbiol. (in press). 941 Maddison, D.R. and W.P. Maddison. 2001. MacClade 4: Analysis of phylogeny and character evolution, version 4.01. Sinauer Associates, Sunderland, Massachusetts, USA. Miller, I.M. 1990. Bacterial leaf nodule symbiosis. Adv. Bot. Res. 17, 163-243. Murray, R.G.E. and E. Stackebrandt. 1995. Taxonomic note: implementation of the provisional status Candidatus for incompletely described prokaryotes. Int. J. Syst. Bacteriol. 45, 186-187. Payne, G.W., P. Vandamme, S.H. Morgan, J.J. LiPuma, T. Coeyne, A.J. Weightman, T.H. Jones, and E. Mahenthiralingam. 2005. Development of a recA gene-based identification approach for the entire Burkholderia genus. Appl. Environ. Microbiol. 71, 39173927. Posada, D. and K.A. Crandall. 1998. Modeltest: testing the model of DNA substitution. Bioinformatics 14, 817-818. Rambaut, A. and A.J. Drummond. 2007. Tracer v1http://evolve.zoo. ox.ac.uk/software.html. Robbrecht, E. 1978a. Sericanthe, a new African genus of Rubiaceae (Coffeeae). Bull. Nat. Plantentuin Belg. 48, 3-78. Robbrecht, E. 1978b. Rubiaceae. Distr. Pl. Afr. 13: map 406-429. Robbrecht, E. 1981. Studies in tropical African Rubiaceae (I). Bull. Nat. Plantentuin Belg. 51, 165-189. Robbrecht, E. and J.F. Manen. 2006. The major evolutionary lineages of the coffee family (Rubiaceae, angiosperms). Combined analysis (nDNA and cpDNA) to infer the position of Coptosapelta and Luculia, and supertree construction based on rbcL, rps16, trnLtrnF and atpB-rbcL data. A new classification in two subfamilies, Cinchonoideae and Rubioideae. Syst. Geogr. Pl. 76, 85-146. Ronquist, F. and J.P. Huelsenbeck. 2003. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19, 15721574. Simmons, M.P. and H. Ochoterena. 2000. Gaps as characters in sequence-based phylogenetic analyses. Syst. Biol. 49, 369-381. Stamatakis, A. 2006. RAxML-VI-HPC: maximum likelihood-based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics 22, 2688-2690. Tabacchioni, S., L. Ferri, G. Manno, M. Mentasti, P. Cocchi, S. Campana, N. Ravenni, and et al. 2008. Use of the gyrB gene to discriminate among species of the Burkholderia cepacia complex. FEMS Microbiol. Lett. 281, 175-182. Tel-Zur, N., S. Abbo, D. Mylaboski, and Y. Mizrahi. 1999. Modified CTAB procedure for DNA isolation from epiphytic cacti of the genus Hylocereus and Selenicerus (Cactaceae). Plant Mol. Biol. Rep. 17, 249-254. Tosh, J., A.P. Davis, S. Dessein, P. de Block, S. Huysmans, M.F. Fay, E. Smets, and E. Robbrecht. 2009. Phylogeny of Tricalysia (Rubiaceae) and its relationships with allied genera based on plastid DNA data: resurrection of the genus Empogona. Ann. Mossouri. Bot. Gard. 96, 194-213. van Hove, C. 1972. Structure and initiation of nodules in the leaves of Neorosea andongensis (Hiern) N. Hallé. Ann. Bot. 36, 259-262. van Oevelen, S., R. de Wachter, P. Vandamme, E. Robbrecht, and E. Prinsen. 2002. Identification of the bacterial endosymbionts in leaf galls of Psychotria (Rubiaceae, angiosperms) and proposal of ‘Candidatus Burkholderia kirkii’ sp. nov. Int. J. Syst. Evol. Microbiol. 52, 2023-2027. van Oevelen, S., R. De Wachter, P. Vandamme, E. Robbrecht, and E. Prinsen. 2004. ‘Candidatus Burkholderia calva’ and ‘Candidatus Burkholderia nigropunctata’ as leaf gall endosymbionts of African Psychotria. Int. J. Syst. Evol. Microbiol. 54, 2237-2239. van Oevelen, S., E. Prinsen, R. de Wachter, and E. Robbrecht. 2001. The taxonomic value of bacterial symbiont identification in African Psychotria (Rubiaceae). Syst. Geogr. Pl. 71, 557-563. Werle, E., C. Schneider, M. Renner, M. Volker, and W. Fiehn. 1994. Convenient single step, one tube purification of PCR products for direct sequencing. Nucleic Acids Res. 22, 4354-4355.