Genome

Barcoding of ( and Chalcidoidea) associated with wild and cultivated in the Western Cape of South Africa

Journal: Genome

Manuscript ID gen-2018-0068.R3

Manuscript Type: Article

Date Submitted by the 28-Sep-2018 Author:

Complete List of Authors: Powell, Chanté; Stellenbosch University Faculty of AgriSciences, Caleca, Virgilio; University of Palermo , Department of Agricultural and Forestry SciencesDraft Sinno, Martina; University of Naples Federico II, Department of Agricultural Sciences van Staden, Michaela; Stellenbosch University Faculty of AgriSciences, Genetics van Noort, Simon; Iziko South African Museum, Department of Natural History Rhode, Clint; Stellenbosch University Faculty of Agrisciences, Genetics Department Allsopp, Elleunorah; Agricultural Research Council, Infruitec- Nietvoorbij van Asch, Barbara; Stellenbosch University Faculty of AgriSciences, Genetics;

Keyword: Braconidae, Chalcidoidea, DNA barcoding, olives, species identification

Is the invited manuscript for consideration in a Special 7th International Barcode of Life Issue? :

https://mc06.manuscriptcentral.com/genome-pubs Page 1 of 48 Genome

1 TITLE

2 Barcoding of parasitoid wasps (Braconidae and Chalcidoidea) associated with wild and cultivated olives

3 in the Western Cape of South Africa

4 RUNNING HEAD

5 DNA barcoding of braconid and chalcid wasps

6 AUTHORS

7 Chante Powell1, Virgilio Caleca2, Martina Sinno3, Michaela van Staden1, Simon van Noort4, Clint

8 Rhode1, Elleunorah Allsopp5, Barbara van Asch1 9 Draft 10 AFFILIATIONS

11 1Department of Genetics, Faculty of Agrisciences, University of Stellenbosch, Stellenbosch, South

12 Africa

13 2Department of Agricultural and Forestry Sciences, University of Palermo, Italy.

14 3Department of Agricultural Sciences, University of Naples "Federico II", Naples, Italy.

15 4Division of Entomology, Department of Natural History, Iziko South African Museum, Cape Town, South

16 Africa.

17 5Agricultural Research Council, Infruitec-Nietvoorbij, Stellenbosch, South Africa.

18

19 CORRESPONDING AUTHOR

20 Barbara van Asch - [email protected]

1 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 2 of 48

21 ABSTRACT

22 Wild and cultivated olives harbor and share a diversity of , some of which are considered

23 agricultural pests, such as the olive fruit . The assemblage of olive-associated and seed

24 wasps is rich and specialized in sub-Saharan Africa, with native species possibly coevolving with their

25 hosts. Although historical entomological surveys reported on the diversity of olive species in the

26 Western Cape Province of South Africa, no comprehensive study has been performed in the region in the

27 molecular era. In this study, a dual approach combining morphological and DNA-based methods was

28 used for the identification of adult specimens reared from olive fruits. Four species of Braconidae and six

29 species of Chalcidoidea were identified, and DNA barcoding methodologies were used to investigate

30 conspecificity among individuals, based on randomly selected representative specimens. Morphological 31 identifications were congruent with DNA data,Draft as NJ and ML trees correctly placed the sequences for 32 each species either at the genus or species level, depending on the available taxa coverage, and genetic

33 distances strongly supported conspecificity. No clear evidence of cryptic diversity was found. Overall

34 seed infestation and rates were higher in wild olives compared to cultivated olives, and

35 highest for Eupelmus spermophilus and Utetes africanus. These results can be used for early DNA-based

36 detection of wasp larvae in olives, and further investigating the and ecology of these species.

37

38 KEYWORDS

39 Braconidae, Chalcidoidea, DNA barcoding, olives, species identification

40

2 https://mc06.manuscriptcentral.com/genome-pubs Page 3 of 48 Genome

41 INTRODUCTION

42 The wild olive tree ( europaea L. subsp. cuspidata) is closely related to the cultivated olive tree (Olea

43 europaea L. subsp. europaea var. europaea) (Green 2002). The subspecies is widely distributed on the

44 African continent from the Southern tip of Africa to South Egypt (Rubio De Casas et al. 2006) where it

45 occurs mainly in Afro-montane forests, and often near water sources; it is also present in small areas of

46 the Asian continent (Green 2002). It is known to host a wide variety of leaf, sap and fruit-feeding insects,

47 and their associated parasitoids (Silvestri 1915; Copeland et al. 2004; Mkize et al. 2008). Wild olive trees

48 are often found in close proximity to non-native cultivated olive trees in the Western Cape Province of

49 South Africa, the main commercial producer of olives, due to its typically Mediterranean with

50 warm, dry summers and mild, moist winters. The region comprising the Western and the Eastern Cape

51 provinces has been identified as home to a high diversity of wasp species described as natural enemies

52 of olive fruit and phytophagous olive seedDraft wasps (Silvestri 1913; Silvestri 1915; Neuenschwander

53 1982).

54 Two olive fruit flies, oleae (Rossi) and Bactrocera biguttula (Bezzi), have also been identified

55 and are presently associated with olives in Africa. Bactrocera oleae, a major pest of cultivated and wild

56 olives, is believed to have originated and disseminated in Africa, and to have accompanied the

57 geographic expansion and domestication of olive trees in the Mediterranean Basin (Zohary 1994; Nardi

58 et al. 2005; Daane & Johnson 2010; Nardi et al. 2010). Bactrocera biguttula is a closely related species

59 endemic to the continent, probably also matching the natural range of the geographic distribution of

60 wild olive trees in sub-Saharan Africa (Munro 1925; Munro 1984; Mkize et al. 2008). Infestation of

61 cultivated olives by B. biguttula has never been reported. The infestation rates of cultivated olives by B.

62 oleae in South Africa are lower than under similar conditions in the Mediterranean Basin, and the

63 limiting factors have been attributed to the action of indigenous parasitoid wasps (Neuenschwander

64 1982; Costa 1998; Hoelmer et al. 2011) and, more recently, to the specific climatic patterns of the region

65 (Giacalone 2011; Caleca et al. 2015; Caleca et al. 2017).

3 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 4 of 48

66 The potential utility of parasitoid wasps for the biological control of B. oleae in the Mediterranean basin,

67 where it causes significant damage (Daane & Johnson 2010), and regions with similar such as

68 the Middle East and California, where invasion has occurred more recently (Rice et al. 2003; Ramezani et

69 al. 2015), has sparked interest in assembling detailed catalogues of Southern and

70 Eastern African wasp species since the early 20th century (Silvestri 1915). Surveys conducted in sub-

71 Saharan Africa have reported the presence of a distinct and broad complex of wasps, including species

72 endemic to the region (Hoelmer et al. 2011). Southern European surveys have shown wasp assemblages

73 less diverse and comprised of a smaller number of specialized species. Five parasitoid species are

74 commonly found in Southern Europe, of which four are chalcids [(Eupelmus urozonus Dalman,

75 mediterraneus Ferrière et Delucchi, martellii Domenichini, and Cyrtoptyx latipes (Rondani), and

76 only one is a braconid [(Psyttalia concolor (Szépligeti)]. Psyttalia concolor is native to North Africa and

77 some southern Italian regions or sub-regionsDraft (Sicily, southern and southern Calabria) (Silvestri

78 1939; Caleca et al. 2017), and was purposefully imported into most of Southern Europe for the control of

79 B. oleae (Hoelmer et al. 2011; Borowiec et al. 2012).

80 Earlier studies provided the first descriptions of wasps associated with wild and cultivated olives in South

81 Africa (the Western Cape, and the Transvaal, a former province that now comprises Gauteng, Limpopo,

82 Mpumalanga and part of the North-West province), and Eritrea (Silvestri 1913; Silvestri 1915;

83 Neuenschwander 1982). A recent survey in Kenya reported wasps associated with B. oleae in wild olives,

84 but only three braconids were identified (Copeland et al. 2004). A more recent study on wild olives in the

85 Eastern Cape Province of South Africa reported the occurrence of both parasitoid and seed wasps, and

86 additionally provided estimates of relative infestation rates in wild olives. Four braconids and seven

87 chalcids were found, although some groups were only identified to genus level (Mkize et al. 2008).

88 Similar results were obtained in the Western Cape (Giacalone 2011; Caleca et al. 2017). None of these

89 works included molecular analyses, and reference DNA barcoding sequences for the majority of the

90 species reported remained unavailable.

4 https://mc06.manuscriptcentral.com/genome-pubs Page 5 of 48 Genome

91 The morphological identification of small hymenopterans requires the expertise of well-trained

92 taxonomists, and is difficult to perform on immature life stages. Additional challenges result from sexual

93 dimorphisms, natural intraspecific variation, and the potential presence of cryptic species (Rowley et al.

94 2007; Al Khatib et al. 2014). DNA barcoding provides a methodological framework for identifying

95 organisms by comparing their degree of nucleotide sequence similarity (expressed as genetic distance)

96 to sets of reference taxa (Hebert et al. 2003). Sequence similarities can then be interpreted using

97 numerical methods such as hierarchical clustering of genetic distances (e.g. the Neighbour-joining

98 algorithm) and statistical evaluation of thresholds of genetic distances. The underlying assumption is that

99 interspecific genetic variation exceeds intraspecific variation.

100 DNA barcoding in relies on nucleotide sequence similarity at a standard region (~650 bp) of the

101 5’-end of the mitochondrial cytochrome oxidase I gene (COI). In recent years, researchers worldwide

102 have been depositing high-quality referenceDraft sequences in public databases (e.g. Barcode of Life Data

103 System, BOLD, www.boldsystems.org) (Ratnasingham & Hebert 2007) that will increasingly allow for the

104 assignment of unknown specimens to morphologically determined taxa, the discrimination of cryptic

105 species, and the elucidation of synonymies (Hebert & Gregory 2005). Although the potential applications

106 of DNA barcoding are indisputable, methodological limitations and the nature of mitochondrial evolution

107 may restrict its applicability in particular taxa. The use of a single marker also confines the amount of

108 genetic variation, thus limiting the ability to understand the patterns of species boundaries (Dupuis et al.

109 2012). Another potential limitation of DNA barcoding based on COI sequences is the possibility that

110 ancestral polymorphic haplotypes have not sorted according to independent speciation events

111 (incomplete lineage sorting) (Ball et al. 2005). Therefore, it is advisable to combine morphological and

112 DNA-based methods for species identification, as sole reliance on either approach has limitations.

113 The aim of this study was to assess the congruence between morphological identification of braconid

114 and chalcid wasps and patterns of genetic clustering and genetic distances within and amongst groups,

115 using novel and publicly available COI sequences. The objectives included a sampling strategy that aimed

5 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 6 of 48

116 at capturing the total assemblage of wasp species associated with wild and cultivated olives, and the

117 assessment of the novel sequences as representative of the species within the context of each particular

118 genus. Additionally, this work also represented an opportunity to report estimates of braconid and

119 chalcid infestation rates across the distribution range of wild and cultivated olive trees in the Western

120 Cape of South Africa, a region known to harbor a rich diversity of these parasitoid and phytophagous

121 wasps.

122

123 MATERIAL & METHODS

124 Sample collection

125 Wild and cultivated olive fruits were collected haphazardly from 16 different areas across the Western

126 Cape Province of South Africa and one area in the Eastern Cape, between March and October 2016 (Fig.

127 1). As the objective of this study was to rear,Draft identify and barcode as many parasitoids and seed wasp

128 species as possible, and fly and wasp infestation is known to be higher in wild olives, the sampling effort

129 focused particularly on wild olives. Sampling of cultivated olives included unsprayed fruit collected on

130 commercial farms, as well as in urban areas. Wild olives were collected according to accessibility in

131 diverse contexts, including the vicinity of cultivated olives, wilderness areas, and ornamental trees in

132 urban settings. Olive fruits were stored in ventilated boxes until the emergence of adults. Adult wasps

133 were euthanized by freezing, and stored individually in absolute ethanol at -20°C until DNA extraction.

134 Morphological identification of all specimens was performed on ethanol-preserved adults.

135 Morphological identification and photographic imaging

136 Braconidae were identified to the genus level using the key available in Parasitoids of Fruit infesting

137 (PAROFFIT) database (http://paroffit.org) (Wharton & Yoder). The genera were identified to

138 the species level following currently available descriptions (Silvestri 1913), and by comparison to

139 photographic images available on PAROFFIT. Chalcidoidea groups/species were identified as follows:

140 according to Gates & Delvare (2008) and Lotfalizadeh et al. (2007); Eupelmus Dahlman

6 https://mc06.manuscriptcentral.com/genome-pubs Page 7 of 48 Genome

141 according to Al Khatib et al. (2014) and Gibson & Fusu (2016), while the only way to identify specimens

142 at species level for these two families and was to refer to species descriptions provided by

143 Silvestri (1915). Identification of Westwood specimens was performed following Bouček et al.

144 (1981) and Nieves-Aldrey et al. (2007).

145 Voucher male and female representatives of each species were randomly selected for photographic

146 imaging. Specimens were washed thrice in absolute ethanol with one-hour intervals followed by an

147 additional overnight wash step. Prior to imaging, specimens were processed in a Leica EM CPD300

148 Critical Point Dryer (Leica Microsystems, Wetzlar, Germany) to maintain the integrity of morphological

149 structures. Specimens were mounted on felt tips and photographed using a Microscope EntoVision

150 Mobile Imaging System, consisting of a Leica Z16 APO zoom lens attached to a digital camera and

151 computer workstation running on the Leica Application Suite v.4.7.1 (Leica Microsystems). The images

152 were deposited onto BOLD Systems and WaspWebDraft (www.waspweb.org), an online bioinformatics

153 resource of wasps, , and documented from the Afrotropical biogeographical region. Female and

154 male specimens were deposited in the entomology collection at the Iziko South African Museum in Cape

155 Town for future reference (Table S1). DNA sequences were not generated from the deposited specimens

156 but from other specimens equally representative of each species, according to morphological

157 identifications.

158 DNA extraction, PCR amplification and sequencing

159 Individual specimens were randomly selected from the total sample of morphologically identified wasps

160 for total DNA extraction and barcoding, and were destroyed in the process. A standard phenol-

161 chloroform protocol (Sambrook et al. 1989) was used for total DNA extraction. The standard COI

162 barcoding region (~710 bp) was amplified using the universal invertebrate barcoding primers (LCO1490

163 and HCO2198) (Folmer et al. 1994) for six species (Bracon celer, formosus, Psyttalia

164 humilis, Psyttalia lounsburyi, aethiopica, and Utetes africanus). Species-specific primers were

7 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 8 of 48

165 designed for Eupelmus spermophilus (Eupel-COI-F and Eupel-COI-R), and genus-specific primers were

166 designed for Eurytoma (Euryt-COI-F2 and Euryt-COI-R2) (Table S2).

167 PCR amplifications were performed in 5 μL reactions containing 1x Kapa HiFi HotStart Ready Mix Kit

168 (KAPA Biosystems), 10 μM of each primer, and 1 μL template DNA. Thermocycling conditions were as

169 follows: initial denaturation at 95°C for 3 min; 5 cycles at 98°C for 20 s, 41°C for 15 s, and 72°C for 1min,

170 followed by 35 cycles of 98°C for 20 s, 56°C for 15 s, 72°C for 1 min; and a final extension at 72°C for 10

171 min. Amplification of the expected fragments was confirmed on a 1.5% agarose gel electrophoresis. PCR

172 products that presented non-specific bands were separated on a 0.8% agarose gel. The correct fragment

173 was then excised from the gel and purified using Zymoclean™ Gel DNA Recovery Kit (Zymo Research).

174 Sequencing was performed using the BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems)

175 at the Central Analytical Facilities of Stellenbosch University, South Africa. Sequences were manually

176 edited, and homology with known taxa wasDraft verified by BLASTn search (www.ncbi.nlm.nih.gov). All

177 sequences were translated into amino acids for the detection of premature stop codons and/or

178 frameshift mutations indicative of pseudogenes with Geneious R11 (www.geneious.com; Kearse et al.

179 2012), using the invertebrate mitochondrial genetic code.

180 Genetic clustering and estimates of sequence divergence

181 All publicly available COI sequences for the Braconidae and Chalcidoidea genera represented in this

182 study were downloaded from Genbank for estimating intra- and interspecific genetic distances, and

183 illustrating sequence clustering based on Neighbour-joining (NJ) and Maximum Likelihood (ML) methods

184 (Table S3). Sequences shorter than 500 bp, containing nucleotide ambiguities, and non-overlapping with

185 the COI region under study were excluded from downstream analyses. Only sequences identified to the

186 species level were included in the analyses, except in the case of the genus Sycophila for which only

187 sequences identified as Sycophila sp. were publicly available. To avoid excessively dense trees in the

188 genetic clustering analyses, duplicate haplotypes in the public dataset were identified and deleted using

189 Geneious R11, and a maximum of six sequences were randomly selected when a large number of

8 https://mc06.manuscriptcentral.com/genome-pubs Page 9 of 48 Genome

190 representatives was available for a single species. For estimates of genetic distances, this last step was

191 not performed (i.e. duplicate sequences were not removed).

192 Nucleotide sequences were aligned with the MAFFT algorithm implemented in Geneious R11. NJ

193 clustering analyses were performed for each genus in MEGA7 (Kumar et al. 2016) using the Kimura-2-

194 parameter (K2P) model (Kimura 1980), with pairwise deletion of the only gap (a 6 bp difference between

195 braconids and chalcids representing two consecutive amino acids). ML trees were reconstructed based

196 on the same alignments with RAxML-HPC Black Box v8.2.10 (Stamatakis, 2014) using the GTRCAT

197 evolutionary model of substitution rate heterogeneity and rapid bootstrapping included in the method

198 (Stamatakis, 2006), and ran on the CIPRES Science Gateway Portal (www.phylo.org; Miller et al. 2010).

199 Clade support in NJ trees was assessed by 1,000 bootstrap replications. Psyttalia humilis and B. celer

200 were used as outgroups for Chalcidoidea, and E. spermophilus and E. varicolor as outgroups for

201 Braconidae. Estimates of intra- and interspecificDraft sequence divergence, and relative standard errors were

202 estimated in MEGA7 using the K2P model.

203 Infestation rates

204 Apparent parasitism rates (APR) for Braconidae species and Neochrysocharis formosus were estimated as

205 follows: APR = total number of adult specimens of the particular species / (total number of tephritid flies

206 + total number of adult braconid and N. formosus specimens). For Chalcidoidea species, apparent seed

207 infestation rates (AIR) were estimated as: AIR = total number of adult specimens of the particular species

208 / total number of olives. For the sake of simplicity, apparent parasitism and apparent seed infestation

209 rates will be generically designated as infestation rates (IFs) in this report.

210

211 RESULTS

212 A total of 83,381 olive fruits (wild = 76,960 and cultivated = 6,421) were haphazardly collected in 16

213 areas in the Western Cape and one area in the Eastern Cape Provinces of South Africa between March

214 and October 2016 (Table S4). No adult wasp specimens were reared from wild or cultivated olive fruits 9 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 10 of 48

215 collected in two areas (Tulbagh and Worcester), or from cultivated olives collected in five areas

216 (Malmesbury, McGregor, Porterville, Riebeek Kasteel and Somerset West). A total of 843 adult wasp

217 specimens was reared from wild (n = 836, 99.2%) and cultivated olives (n = 7, 0.8%). Specimens were

218 distributed among six species of Chalcidoidea and four species of Braconidae: Eupelmus spermophilus, n

219 = 321 (38.1%); Eurytoma oleae, n = 58 (6.9%); Eurytoma varicolor, n = 136 (16.1%); Sycophila aethiopica,

220 n = 50 (5.9%); Neochrysocharis formosus, n = 23 (2.7%); Ormyrus sp., n = 47 (5.6%), Psyttalia humilis, n =

221 28 (3.3%); Psyttalia lounsburyi, n = 22 (2.6%); Utetes africanus, n = 130 (15.4%); and Bracon celer, n = 28

222 (3.3%). Photographic images of one male and one female specimen representative of each species are

223 presented in Fig. 2-5. Overall, Chalcidoidea (n = 635, 75.3%) were more abundant than Braconidae (n =

224 208, 24.7%). The most abundantly reared chalcid was E. spermophilus, and the most abundant braconid

225 was U. africanus. Only three species were recovered from a total of 2,583 cultivated olives in three areas

226 (Franschhoek, Paarl and Stellenbosch): E. spermophilusDraft (n = 4), B. celer (n = 2), and P. lounsburyi (n = 1)

227 (Fig. 6).

228 PCR amplification using the universal invertebrate barcoding primers LCO1490 and HCO2198 only

229 generated the expected product in DNA samples from P. humilis, P. lounsburyi, B. celer, and S.

230 aethiopica. Eupelmus spermophilus and Eurytoma species did not consistently amplify with these

231 primers; therefore, new primers were designed for the amplification of a shorter amplicon (650 bp)

232 within the standard COI barcoding region (Table S2). The new Euryt-COI-F2 and Euryt-COI-R2 primers

233 generated non-specific products in E. varicolor; therefore, purification of the specific band from a 0.8%

234 agarose gel was performed prior to sequencing reactions. Utetes africanus also presented non-specific

235 amplifications with the universal primers, and purification of the specific band from a 0.8% agarose gel

236 was necessary.

237 One Wolbachia sequence, identified by BLASTn search during the sequence quality control procedure,

238 was unintentionally obtained from an E. spermophilus DNA sample using the universal primers. These

239 primers did not consistently generate PCR products in E. spermophilus, and were subsequently replaced

10 https://mc06.manuscriptcentral.com/genome-pubs Page 11 of 48 Genome

240 by newly-designed species-specific primers (Eupel-COI-F and Eupel-COI-R, Table S2). The species-specific

241 primers were then used for generating the DNA data presented in this study, and did not amplify

242 Wolbachia sequences. Three putative pseudogene fragments, amplified from E. varicolor using the Euryt-

243 COI-F2/Euryt-COI-R2 primer pair, were also detected during the sequence quality control procedure.

244 These sequences were similar to the functional COI region, but amino acid translation showed several

245 stop codons; therefore, they were excluded from downstream analyses. Novel reference barcoding

246 sequences were generated for six of the nine species identified in this study: Bracon celer (n = 1), Utetes

247 africanus (n = 10), Eupelmus spermophilus (n = 10), Eurytoma oleae (n = 9), Eurytoma varicolor (n = 6),

248 and Sycophila aethiopica (n = 1). Additional sequences were generated for Psyttalia lounsburyi (n = 4)

249 and Psyttalia humilis (n = 3), and Neochrysocharis formosus (n = 2) (Table S5). All sequences with the

250 corresponding trace files, specimen images, GPS coordinates and biological data were deposited on

251 BOLD (projects UTET, SYCPH, PSYT, NCHRY,Draft EURYT, EUPEL and BRCN), and made publicly available. All

252 sequences were also deposited in Genbank (Table S3).

253 For an overview of the current taxonomic coverage of Braconidae and Chalcidoidea, two separate family

254 trees were constructed, with posterior condensation of species clusters into single branches (Fig. 7 and

255 8). Genus-specific trees were also constructed to provide a detailed illustration of the relationships

256 between the individual sequences generated in this study (Fig. S1-S7). Species clusters were strongly

257 supported by the NJ distance-based method, and were in agreement with the ML analysis; therefore,

258 only NJ-based trees are shown, with reference to the relevant topological differences in the text.

259 BRACONIDAE

260 Bracon celer Szépligeti (Fig. 2 A-B) represented 13.5% of the total braconids, similarly to P. humilis and P.

261 lounsburyi. Bracon celer was reared almost exclusively from wild olives, with a single specimen found in

262 cultivated olives from Paarl (Table S4). The genetic clustering analyses for the genus Bracon included 22

263 sequences distributed among nine species, with four species represented by a single sequence. NJ and

264 ML trees recovered identical topology, and showed B. celer (n = 1, this study) nested as an internal

11 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 12 of 48

265 branch. The genetic clustering was consistent with species designations except for the non-monophyly

266 with strong nodal support for B. asphondyliae (Fig. S1).

267 Psyttalia humilis (Silvestri) (Fig. 2 C-D) was reared exclusively from wild olives in five areas, and

268 represented 13.5% of all braconids, while Psyttalia lounsburyi (Silvestri) (Fig. 3 A-B) was reared from

269 both cultivated (a single specimen in Paarl) and wild olives in five areas, and represented 10.6% of all

270 braconids (Table S4). The NJ and ML analyses for the genus Psyttalia included 54 sequences distributed

271 among 11 species, with only three species represented by a single sequence. The topology of the trees

272 showed monophyletic clustering of sequences in congruence with species designations (Fig. S2). The P.

273 humilis (n = 3) and P. lounsburyi (n = 4) specimens identified and sequenced in this survey grouped with

274 the publicly available sequences in their respective monophyletic clusters. The same pairs of sister

275 species were recovered in NJ and ML (e.g. P. lounsburyi/P. phaeostigma; P. humilis/P. concolor),

276 although the topology of the deeper branchesDraft differed, albeit with low statistical support. Interspecific

277 sequence divergence ranged between 8.3% for the species pair P. carinata/P. ponephoraga and 16.3%

278 for P. carinata/P. fletcheri. Intraspecific sequence divergence was estimated for six species, and

279 maximum distances were lower than 1.7% in all cases (Table S6). Intraspecific genetic distances were

280 also estimated separating the new P. lounsburyi and P. humilis sequences from the conspecific

281 sequences available on Genbank. No differences (i.e. high genetic distances) were found; therefore, the

282 public dataset did not seem to include sequences incorrectly assigned to species.

283 Utetes africanus (Szépligeti) (Fig. 3 C-D) was the most abundantly reared braconid (62.5%), and was

284 exclusively found in wild olives in nine areas (Table S4). The NJ and ML trees for the genus Utetes

285 included 52 sequences distributed among eight species, with four species represented by a single

286 sequence. U. canaliculatus (n = 34) represented 65.4% of the total sequence dataset for the genus (Fig.

287 S3). The NJ and ML trees showed the same topology, except for the position of U. magnus. A

288 polyphyletic pattern was recovered for U. canaliculatus, with a highly diverged monophyletic group

289 (cluster 3), a polyphyletic group including U. frequens (cluster 2), and a monophyletic (in NJ) or

12 https://mc06.manuscriptcentral.com/genome-pubs Page 13 of 48 Genome

290 polyphyletic (in ML, where it included U. magnus) cluster 1 (Fig. S3). The maximum intraspecific genetic

291 divergence considering U. canaliculatus as a single group was 11.0%, whereas for each of the separate

292 clusters it was lower than 0.7%. The divergence was highest between cluster 3 and the other two

293 clusters, and the lowest between clusters 1 and 2 (Table S7), as suggested by the topology of the tree.

294 Utetes africanus (n = 10, this study) formed a monophyletic cluster, and the sequences had high

295 similarity (maximum intraspecific distance = 0.4%).

296 CHALCIDOIDEA

297 Eupelmus spermophilus Silvestri (Fig. 4 A-B) represented 50.6% of the total chalcids, and was recovered

298 from the three areas where wasps were reared from cultivated olives, and in nine areas from wild olives

299 (Table S4). The NJ and ML trees of the genus Eupelmus included 99 sequences distributed among 37

300 species, with 18 species represented by a single sequence (Fig. S4). The general topology of the trees

301 recovered monophyletic clustering for all EupelmusDraft species, including E. spermophilus (n = 10, this

302 study), albeit with different topology and low statistical support of deeper nodes in ML and NJ.

303 Intraspecific sequence divergence was estimated for nine species (Table S8). Maximum intraspecific

304 distances ranged between 2.3% for E. spermophilus and 8.7% for E. annulatus. Interspecific sequence

305 divergence ranged between 16.9% for the species pair E. spermophilus/E. azureus, and lowest for E.

306 minozonus/E. urozonus (7.8%).

307 Eurytoma oleae Silvestri (Fig. 4 C-D) represented 9.1% of the total chalcids, and was reared exclusively

308 from wild olives in three areas (Table S4). The highest IF was found in Riversdale (1.89%), whereas the

309 average was 0.10% in the other areas. Eurytoma varicolor Silvestri (Fig. 4 E-F) was reared exclusively

310 from wild olives in five areas, and represented 21.4% of the total chalcids (Table S4). The NJ and ML trees

311 for the genus Eurytoma comprised 59 sequences distributed among 16 species, with only four species

312 represented by a single sequence. Although deeper nodes had different topologies with low statistical

313 support in NJ and ML, all species formed monophyletic clusters, including E. oleae and E. varicolor (n = 9

314 and n = 6, respectively, this study) (Fig. S5). All maximum intraspecific distances were lower than 2.7%,

13 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 14 of 48

315 and interspecific distances were higher than 7.7% (E. morio/E. striolata), with the highest between E.

316 oleae and E. varicolor (18.1%) (Table S9).

317 Sycophila aethiopica (Silvestri) (Fig. 5 C-D) was reared exclusively from wild olives in four areas, and

318 represented 7.9% of the total chalcids (Table S4). The NJ and ML trees of the genus Sycophila included 71

319 sequences, with only one sequence identified to the species level (S. aethiopica, this study). Identical

320 sequence clusters were recovered in NJ and ML, albeit with different topology and low statistical support

321 of deeper nodes. Sycophila aethiopica nested in the interior branches of the trees (Fig. S6).

322 Neochrysocharis formosus (Westwood) (Fig. 5 A-B) was reared exclusively from wild olives in Paarl (n =

323 23), and represented the lowest proportion (3.6%) of the total chalcids (Table S4). A previous phylogeny

324 of Eulophidae based on morphological and molecular markers (COI and 28S rRNA D2-D5) showed that N.

325 formosus and N. clinias were a paraphyletic group with respect to sp., although the study

326 included a single sequence for each NeochrysocharisDraft species from Italy (Burks et al. 2011). Due to the

327 poor sequence coverage of the genus Neochrysocharis, public Asecodes sequences with high quality (A.

328 lucens, n = 6) were included in the NJ and ML analyses, along with the available sequences for

329 Neochrysocharis identified to the species level (N. formosus, n = 3; and N. clinias, n = 1). The NJ and ML

330 trees recovered N. formosus (n = 2, this study) and A. lucens as sister species with high statistical

331 support. However, N. formosus HM365028 (Genbank) did not cluster with the new N. formosus

332 sequences (Fig. S7). Genetic distance estimates showed that the maximum divergence between the

333 three N. formosus sequences was 12.9% (Table S10). A closer inspection revealed that N. formosus

334 HM365028 was highly polymorphic relative to the two new N. formosus sequences, which diverged

335 between them by only 1.1%.

336 Ormyrus sp. (Fig. 5 E-F) was reared exclusively from wild olives in five areas, and represented 7.4% of the

337 total chalcids (Table S4). Identification to the genus level (Ormyrus Westwood) was performed using

338 solely morphological characters, as molecular analyses were not successful. Although PCR amplification

14 https://mc06.manuscriptcentral.com/genome-pubs Page 15 of 48 Genome

339 products were generated and sequenced, BLASTn search resulted in no matches with known COI

340 sequences, or any other sequence.

341 Overall, apparent parasitism rate was the highest for U. africanus (11.80%) (Fig. S8). Psyttalia, B. celer

342 and N. formosus had approximately five-fold lower IFs (2.54% for P. humilis, 2.00% for P. lounsburyi,

343 2.54% for B. celer, and 2.09% for N. formosus). Apparent seed infestation rate was the highest for

344 Eupelmus spermophilus (0.38%). Eurytoma oleae, S. aethiopica and Ormyrus sp. had similar IFs (average

345 0.06%), and E. varicolor had an intermediate IF of 0.16%. A richer wasp assemblage was reared from wild

346 olives compared to cultivated olives, from which only two braconids (B. celer and P. lounsburyi) and one

347 chalcid (E. spermophilus) were reared at low IFs in three areas (Franschhoek, Paarl and Stellenbosch)

348 (Fig. S9 – S10).

349

350 DISCUSSION Draft

351 In this study, 10 wasp species (four braconids and six chalcids) were reared from wild and cultivated

352 olives, and identified based on morphological characters following the currently available keys. All

353 groups reported in a previous survey of wild olives performed in the Eastern Cape (Mkize et al. 2008)

354 were observed, except for N. formosus, which was only recovered in the present survey. Resolution to

355 the species level was improved for two of the genera reported in the Eastern Cape: Eurytoma, with the

356 identification of E. oleae and E. varicolor, and Sycophila, with the identification of S. aethiopica.

357 Eupelmus afer was reportedly reared in the Eastern Cape (Mkize et al. 2008), but was not identified

358 among the specimens reared in the present study.

359 The sampling strategy was haphazard and opportunistic, and not designed to consistently survey and

360 compare the assemblage of species in wild and cultivated olives in the Western Cape, but to potentiate

361 the rearing of the widest possible range of wasps for morphological species identification and DNA

362 barcoding. Some areas were visited only once, while other areas were sampled multiple times (e.g. Paarl

363 and Stellenbosch). Additionally, the number of cultivated olives collected was much lower than the

15 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 16 of 48

364 number of wild olives, for which more areas were also sampled. This sampling bias was deliberate

365 because wild olives are known to harbor more olive flies, braconids and chalcids than cultivated olives.

366 The low presence of braconids in cultivated olives is most probably due to three main reasons. First,

367 infestation is relatively low in the Western Cape, thus precluding high levels of parasitoid

368 wasp populations. Second, braconids have difficulty reaching the third instar fly larvae, the larval stage

369 most attacked by parasitoids, as third instar larvae feed close to the olive kernel, and the thick pulp layer

370 of cultivated olives limits the action of parasitoids. This is especially relevant in the case of Utetes

371 africanus, a species that has a very short . As for chalcids, seed wasps attack olives when the

372 fruit is small and the kernel is still soft, but infestation is also limited because of the thicker pulp layer of

373 cultivated olives compared to wild olives. The atypical climatic conditions of extremely low rainfall in the

374 Western and the Eastern Cape Provinces since 2015 may have also contributed to the absence of wasps

375 in olives collected in five areas, and to non-representativeDraft infestation rates. Therefore, the estimates for

376 apparent parasitism and seed infestation rates here presented should be interpreted with caution.

377 However, it is relevant to note that Utetes africanus was the most abundantly reared braconid, and

378 Eupelmus spermophilus was the most abundant chalcid, with the latter the most abundant wasp overall

379 (38%). It is also relevant to note that, despite the haphazard olive sampling across the Western Cape, a

380 higher diversity of species and higher IFs were found in wild olives than in cultivated olives, as expected.

381 At least one specimen per species was sequenced for the standard barcoding COI region for all species,

382 except for Ormyrus sp. These nucleotide sequences represent the first DNA barcoding references for all

383 species except N. formosus and the two Psyttalia species, for which at least one sequence was publicly

384 available. The consistency between sequence similarity and morphological identification was then

385 investigated using K2P distances, and NJ and ML trees. Phylogenetic reconstruction and estimates of

386 genetic distances offer useful insights into evolutionary relationships among taxa, thus assisting species

387 identification, provided that the specific taxonomic group is well represented in the reference dataset

388 (Hebert et al. 2003; Ross et al. 2003; Hebert et al. 2004; Ross et al. 2008; Collins et al. 2012). This was not

16 https://mc06.manuscriptcentral.com/genome-pubs Page 17 of 48 Genome

389 the case for all species identified in this study. For example, publicly available sequences for the genus

390 Sycophila, although abundant (n = 70), were not identified to the species level. The genus Bracon

391 followed a pattern of incomplete identification: 75% of the 447 public COI sequences were identified as

392 Bracon sp., and 68% of the sequences identified to the species level were duplicates (i.e. sequences with

393 identical residues), resulting in a final dataset of 22 overlapping COI sequences. Neochrysocharis was

394 similarly covered, as 34% of the public sequences were only identified to the genus level. Additionally,

395 after the removal of duplicates, 92% of the remaining 24 Neochrysocharis public sequences identified to

396 the species level had a short overlap with the standard COI barcoding region. Therefore, the final dataset

397 for the genus Neochrysocharis included only the four sequences used in the NJ and ML analyses. These

398 difficulties highlight the importance of good taxonomic coverage for the generation of reliable species

399 reference sequences. In the context of the purpose of this study, which aimed at associating

400 morphologically identified specimens with DraftDNA barcodes, phylogenetic reconstruction using the ML

401 methodology did not show improved resolution or reliability over the distance-based NJ method, as NJ

402 and ML recovered the same monophyletic species clusters with high statistical support. Deeper

403 branches, on the other hand, were as poorly supported in NJ as in ML, as expected when using relatively

404 short sequences (~500 bp) of closely related species (Min & Hickey 2007).

405 Overall, we found complete concordance between morphological identification of specimens, sequence

406 clusters on NJ and ML trees, and genetic distances for six species (Eupelmus spermophilus, Eurytoma

407 oleae, Eurytoma varicolor, Psyttalia humilis, Psyttalia lounsburyi and Utetes africanus) of the ten species

408 reared from olives. No clear evidence for cryptic diversity was found, as these species formed

409 monophyletic clusters with high statistical support, and maximum intraspecific genetic distances were

410 within the range of the commonly used barcoding thresholds of 2-3%, and lower than 1.3% in all cases,

411 except for Eurytoma oleae (2.7%). Interestingly, the maximum intraspecific genetic distance in Eupelmus

412 spermophilus was 2.3%, the lowest in the genus Eupelmus, for which high intraspecific divergence was

413 found, the most striking case being E. annulatus (8.7%).

17 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 18 of 48

414 BRACONIDAE

415 Bracon celer is an idiobiont ectoparasitoid of third (last) instar olive fruit fly , and the only Bracon

416 species known to be an olive fruit fly parasitoid (Silvestri 1913). In sub-Saharan Africa, B. celer has been

417 reported in Kenya, Ethiopia, Namibia, and South Africa (Silvestri 1913; Silvestri 1915; Neuenschwander

418 1982; Mkize et al. 2008; Daane et al. 2011). The genus was previously found to be monophyletic with

419 high statistical support, using 658 bp COI sequences (Matsuo et al. 2016). Our NJ and ML analyses

420 recovered B. asphondilae and B. tamabae as non-monophyletic with low statistical support, probably

421 due to the shorter COI region utilized (547 bp). Only one B. celer specimen was sequenced in this study,

422 therefore precluding estimates of intraspecific variation. However, B. celer nested as an interior tree

423 branch, suggesting that it can be used as a reference for the species.

424 Psyttalia lounsburyi and Psyttalia humilis are endoparasitoids of tephritids endemic to sub-Saharan

425 Africa. The two species have been found inDraft South Africa, Namibia and Kenya (Copeland et al. 2004; Mkize

426 et al. 2008; Rugman-Jones et al. 2009; Daane et al. 2011), and both have been tested as exotic biocontrol

427 agents of B. oleae in Europe and California, albeit with limited success (Daane et al. 2008; Borowiec et al.

428 2012). Previous studies of the genus Psyttalia based on COI sequences showed monophyly of Psyttalia

429 species, including P. lounsburyi and P. humilis (Cheyppe-Buchmann et al. 2011; Borowiec et al. 2012;

430 Schuler et al. 2016), and phylogenetic reconstruction based on COI and 28sD2 sequences provided

431 further support (Rugman-Jones et al. 2009). Our NJ and ML analyses and the estimates of intra- and

432 interspecific genetic distances support the morphological identification of the specimens analyzed in this

433 study, and the utility of standard DNA barcoding for the molecular identification of species belonging to

434 the genus Psyttalia, at least for those with good intraspecific coverage.

435 Utetes africanus is a parasitoid reported in South Africa, Namibia and Kenya (Silvestri 1913; Copeland et

436 al. 2004; Mkize et al. 2008; Daane et al. 2011). It has been reported as being more abundant in wild

437 olives than in cultivated olives (Neuenschwander 1982; Mkize et al. 2008; Giacalone 2011; Caleca et al.

438 2017), most likely because its short ovipositor is unable to reach fly larvae buried deep inside the pulp of

18 https://mc06.manuscriptcentral.com/genome-pubs Page 19 of 48 Genome

439 the large fruit of cultivated olives. The genus Utetes was shown to be polyphyletic (Hamerlinck et al.

440 2016), and three main clusters were recovered for U. canaliculatus with an exact correspondence

441 between microsatellite genetic distances and a COI maximum parsimony tree (Hood et al. 2015). Our NJ

442 and ML trees also recovered non-monophyly for the genus, the same three U. canaliculatus clusters, and

443 inconsistency between species designations and sequence clustering (e.g. U. tabellariae was positioned

444 within U. canaliculatus in cluster 2). Comparison of genetic distances suggested that U. canaliculatus

445 cluster 3 represents an evolutionary unit highly diverged from U. canaliculatus clusters 1 and 2. In

446 agreement with the morphological identification, U. africanus was a monophyletic cluster with low

447 intraspecific divergence (0.4%), thus supporting the use of these sequences as references for the species.

448 CHALCIDOIDEA

449 Eupelmus spermophilus was found “emerging from the seeds of wild olive fruits” (Silvestri 1915). This

450 species was previously reported in Eritrea, Draftand in the Western and Eastern Cape Provinces of South

451 Africa (Silvestri 1915; Mkize et al. 2008). In agreement with previous phylogenetic analyses focusing on

452 the Eupelmus urozonus species complex (Al Khatib et al. 2014; Al Khatib et al. 2016), our NJ and ML trees

453 showed concordance between monophyletic clustering and morphological identification for the genus

454 Eupelmus, including E. spermophilus. Interspecific genetic distances were generally high, and supported

455 the species designations. Interspecific divergence was exceptionally low for E. urozonus/E. minozonus

456 (7.8%), suggesting a more recent divergence for this pair, represented as sister clades in the NJ tree.

457 Maximum intraspecific genetic distances within this genus were exceptionally high for some species,

458 suggesting the presence of cryptic diversity. For example, E. annulatus had a maximum intraspecific

459 genetic distance of 8.7%), and the sequences were distributed between two well-supported clades

460 composed by E. annulatus 1, 2 and 3, and E. annulatus 5, 6 and 7 (Fig. S4). The mean genetic distance

461 between the two clades was 7.2%, and the maximum within-clade distance was lower than 3.1%, thus

462 suggesting that not all sequences designated as E. annulatus are conspecific. Although this was not the

463 case for E. spermophilus (2.3%, this study), a pattern of high maximum intraspecific distances (4.0 - 8.7%)

19 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 20 of 48

464 was found for the all Eupelmus species reported in a previous assessment of this genus, except for E.

465 minozonus and E. gemellus (2.7%) (Al Khatib et al. 2014).

466 Eurytoma oleae and Eurytoma varicolor were reported to develop on the seeds of olives, and the

467 species may be phytophagous seed wasps or parasitoids of seed wasps (Silvestri 1915; Neuenschwander

468 1982). Both species were previously found in Eritrea and South Africa (Western Cape) (Silvestri 1915).

469 Eurytoma oleae was also identified in a previous study in the Eastern Cape, as well as a Eurytoma sp. that

470 most likely represented E. varicolor (Mkize et al. 2008). The geographic range of Eurytoma species

471 associated with olive trees probably extends to Kenya, where unidentified Eurytomidae were reportedly

472 reared from wild olives (Copeland et al. 2004). Our NJ and ML analyses recovered monophyletic clusters

473 in accordance with species designations, including E. oleae and E. varicolor. As maximum intraspecific

474 genetic distances in the genus Eurytoma ranged between 0.4 and 1.5%, future investigation of potential

475 cryptic diversity in E. oleae using additionalDraft genetic markers may be warranted. The range of interspecific

476 genetic distances (>10.2%) support the utilization of the standard barcoding COI region for species

477 identification within this genus.

478 Sycophila aethiopica is possibly a parasitoid of seed wasps (Silvestri 1915). The species was previously

479 reported in Eritrea and South Africa (Western Cape) (Silvestri 1915; Neuenschwander 1982), and most

480 probably reported in the Eastern Cape as Sycophila sp. (Mkize et al. 2008). Sycophila aethiopica was

481 represented by a single sequence generated in this study, and none of the publicly available sequences

482 were identified to the species level, therefore hampering estimation of intra- and interspecific

483 divergences, and specific NJ clustering. However, the single Sycophila aethiopica sequence nested within

484 the interior branches of the NJ and ML trees, thus supporting its utility as reference for the species. The

485 genus Sycophila was shown to be monophyletic using nuclear markers (28S and 18S rRNA), and non-

486 monophyletic using mitochondrial markers (16S and COI) (Chen et al. 2004). The low statistical support

487 for the deeper-level divergences in the NJ and ML trees also suggest that future phylogenetic

20 https://mc06.manuscriptcentral.com/genome-pubs Page 21 of 48 Genome

488 reconstructions may have to include a combination of nuclear and mitochondrial sequences for the

489 recovery of reliable branching patterns within this genus.

490 Neochrysocharis formosus (formerly N. formosa) is a non-specialized endoparasitoid with worldwide

491 distribution except for Australia (Chien & Ku 2001). This species was also previously found in the Western

492 Cape in several areas (including Paarl) (Neuenschwander 1982), but it was not reported in the Eastern

493 Cape (Mkize et al. 2008). The classification of genera and species in the tribe Entedonini is controversial,

494 particularly in the case of small-bodied species, such as Neochrysocharis. Neochrysocharis (N. formosus

495 HM365028 and N. clinias HM365038) was previously shown to be, with respect to Asecodes,

496 paraphyletic in molecular analyses, paraphyletic in combined (molecular and morphological) parsimony

497 analysis, and monophyletic in combined Bayesian analysis (Burks et al. 2011). The inclusion of the N.

498 formosus Nf21 and N. formosus Nf18 sequences recovered the of the genus Neochrysocharis

499 with regards to Asecodes. Additionally, bothDraft the trees and the estimates of genetic divergence indicated

500 that the Neochrysocharis COI sequence dataset is composed of two different species, with one species

501 represented by N. formosus HM365028, and the other represented by N. formosus Nf21 and N. formosus

502 Nf18 identified in this study. Deeper molecular coverage of species will be necessary to resolve

503 taxonomic classifications within this group.

504 Ormyrus and similar species are considered to be parasitoids attacking seed wasps both in Eritrea and in

505 the Western Cape (Silvestri 1915), and no ormyrids are known to parasitize B. oleae. Ormyrus sp. was

506 reportedly reared from wild and cultivated olives in the Western and the Eastern Cape (Neuenschwander

507 1982; Mkize et al. 2008; Giacalone 2011). Several attempts were made to obtain PCR products from

508 Ormyrus sp. without success, suggesting that specific primers may have to be designed for future

509 analysis.

510 Wolbachia and pseudogenes

511 Unintended amplification and sequencing of two types of fragments non-representative of the barcoding

512 COI region occurred in some DNA extracts. One Wolbachia sequence, the most common bacterial

21 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 22 of 48

513 endosymbionts in , was sequenced from E. spermophilus. This is known to occur when

514 attempting PCR amplification of COI with universal primers from total genomic DNA (Smith et al.

515 2012). Procedures for the quality control of the data (e.g. BLASTn searches) are mandatory to prevent

516 false results in downstream assessments of genetic variation and phylogenetic reconstructions. Putative

517 pseudogene fragments, possibly nuclear pseudogenes of mitochondrial origin (NUMTs), were also

518 obtained with the genus-specific primers (Euryt-COI-F2/Euryt-COI-R2) in three E. varicolor samples. This

519 could be explained by the non-specificity of the primers for E. varicolor, as these were designed based on

520 E. oleae sequences. PCR amplification of NUMTs is known to occur frequently in DNA barcoding of

521 insects, and quality control for their identification (e.g. amino acid translation) can greatly contribute to

522 detect and purge these sequences from COI datasets (Leite 2012).

523 The present assessment of wasp species associated with wild and cultivated olives represents a

524 comprehensive coverage of the rich endemicDraft parasitoid and seed wasp diversity in South African olives.

525 Sub-Saharan African Braconidae are particularly interesting, due to their potential use as exotic

526 biocontrol agents for controlling olive fly populations in regions where this pest lacks specialized natural

527 enemies (e.g. California). The assemblage of Chalcidoidea associated with olives remains poorly studied,

528 and details of the specific biology remain unknown for several species. For example, Eupelmus

529 spermophilus, Eurytoma oleae, Eurytoma varicolor, Sycophila aethiopica and Ormyrus sp. have been

530 variously reported as possible seed wasps, parasitoids of olive fruit flies, or hyperparasitoids. DNA

531 analyses can be applied to the identification of immature insect life stages such as eggs, larvae, nymphs

532 or pupae, otherwise often impossible to identify morphologically. This is particularly pertinent for the

533 early detection of invasions, disseminations and infestation outbreaks of agricultural pests. In the

534 particular case of olives, methodologies inspired by DNA barcoding for species identification could also

535 be used in the analyses of insect material collected from the interior of the fruits to elucidate the elusive

536 lifestyle of the wasps and other insect groups associated with wild and cultivated olive trees.

537

22 https://mc06.manuscriptcentral.com/genome-pubs Page 23 of 48 Genome

538 ACKNOWLEDGEMENTS

539 The authors are grateful to numerous olive growers and landowners for allowing the collection of

540 samples on their properties. Emma Cook is acknowledged for assistance in laboratory work. This study

541 was funded by the National Research Foundation (FBIP - Small Grants and Biodiversity Surveys,

542 FBIS150603118644, Grant 98112), the National Research and Technology Fund of South Africa

543 (Reference RTF150428117617; Grant 98590), and Stellenbosch University (Staff Grant, Research Sub-

544 committee B).

545

546 REFERENCES

547 Ball, S.L. Hebert, P.D.N., Burian, S.K., and Webb, J.M. 2005. Biological identifications of mayflies

548 (Ephemeroptera) using DNA barcodes. J. N. Am. Benthol. Soc. 24(3): 508. doi:10.1899/04-142.1.

549 Draft

550 Borowiec N., Groussier-Bout G., Vercken E., Thaon M., Auguste-Maros A., Warot-Fricaux S., et al. 2012.

551 Diversity and geographic distribution of the indigenous and exotic parasitoids of the olive fruit fly,

552 Bactrocera oleae (Diptera: Tephritidae), in Southern France. IOBC-WPRS Bulletin. 79(i): 71–78.

553

554 Bouček, Z., Watsham, A., and Wiebes, J.T. 1981. The fauna of the receptacles of Ficus

555 thonningii (, Chalcidoidea). Tijdschr. Entomol. 124(5): 149-233.

556

557 Burks, R.A., Heraty, J.M., Gebiola, M., and Hansson, C. 2011. Combined molecular and morphological

558 phylogeny of Eulophidae (Hymenoptera: Chalcidoidea), with focus on the subfamily .

559 Cladistics, 27(6): 581–605. doi: 10.1111/j.1096-0031.2011.00358.x.

560

23 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 24 of 48

561 Caleca, V., Giacalone, C., Maltese, M., and Tortorici, F. 2017. Contenimento naturale di Bactrocera oleae

562 (Rossi): Clima o parassitoidi? Confronto tra Western Cape (Sud Africa) e Sicilia. Atti Acc. Naz. Ital.

563 Entomol. 64: 99–105.

564

565 Caleca, V., Giacalone, C. and Allsopp, E., 2015. Olive fruit fly: a threat to the South African olive industry?

566 Technology, South African Olive Industry Association. August/September, 71–72.

567

568 Chen, Y., Xiao, H., Fu, J., and Huang, D-W. 2004. A molecular phylogeny of eurytomid wasps inferred

569 from DNA sequence data of 28S, 18S, 16S, and COI genes. Mol. Phyl. Evol. 31(1): 300–307.

570 doi:10.1016/S1055-7903(03)00282-3. PMID: 15019626.

571

572 Cheyppe-Buchmann, S., Bon, M.C., Warot, S.,Draft Jones, W., Malausa, T., Fauverge, X., et al. 2011. Molecular

573 characterization of Psyttalia lounsburyi, a candidate biocontrol agent of the olive fruit fly, and its

574 Wolbachia symbionts as a pre-requisite for future intraspecific hybridization. BioControl, 56(5): 713–724.

575 doi:10.1007/s10526-011-9346-x.

576

577 Chien, C.C. and Ku, S.C. 2001. Appearance and life history of Neochrysocharis formosus (Hymenoptera:

578 Eulophidae). Formosan Entomol. 21(4): 383–393.

579

580 Collins, R.A., Boykin, L., Cruickshank, R.H., and Armstrong, K.F. 2012. Barcoding’s next top model: an

581 evaluation of nucleotide substitution models for specimen identification. Methods Ecol. Evol. 3(3): 457–

582 465. doi: 10.1111/j.2041-210X.2011.00176.x.

583

584 Copeland, R., White, I.M., Okumu, M., Machera, P., and Wharton, R.A. 2004. Insects associated with

585 fruits of the Oleaceae (Asteridae, Lamiales) in Kenya, with special reference to the Tephritidae (Diptera).

24 https://mc06.manuscriptcentral.com/genome-pubs Page 25 of 48 Genome

586 In Evenhuis, N.L., and Kaneshiro, K.Y. (eds.), D. Elmo Hardy Memorial Volume. Contribution to the

587 Systematics and Evolution of Diptera. Bishop Mus. Bull. Entomol.. 12: 135–164.

588

589 Costa, C., 1998. Olive production in South Africa: a handbook for olive growers. ARC - Institute for

590 Tropical and Subtropical Crops, Nelspruit, South Africa. ISBN-10: 1868491005.

591

592 Daane, K.M., Johnson, M.W., Pickett, C.H., Simen, K., Wang, X., Nadel, H., et al. 2011. Biological controls

593 investigated to aid management of olive fruit fly in California. Calif. Agric. 65(1): 21–28.

594 doi:10.3733/ca.v065n01p21.

595

596 Daane, K.M., Sime, K.R., Johnson, M.W., Walton, V.M., et al. 2008. Psyttalia lounsburyi (Hymenoptera:

597 Braconidae), potential biological control agentDraft for the olive fruit fly in California. Biol. Control. 44(1): 79–

598 89. doi:10.1016/j.biocontrol.2007.08.010.

599

600 Daane, K.M. and Johnson, M.W. 2010. Olive fruit fly: managing an ancient pest in modern times. Annu.

601 Rev. Entomol. 55: 151–169. doi:10.1146/annurev.ento.54.110807.090553. PMID: 19961328.

602

603 Dupuis, J.R., Roe, A.D. and Sperling, F.A. 2012. Multi-locus species delimitation in closely related animals

604 and fungi: one marker is not enough. Mol. Ecol. 21:4422–4436. doi:10.1111/j.1365-294X.2012.05642.x.

605 PMID: 22891635.

606

607 Folmer, O., Black, M., Hoeh, W., Lutz, R. and Vrijenhoek, R. 1994. DNA primers for amplification of

608 mitochondrial cytochrome c oxidase subunit I from diverse metazoan invertebrates. Mol. Mar. Biol.

609 Biotechnol. 3(5): 294–299. PMID: 7881515.

610

25 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 26 of 48

611 Gates, M., and Delvare, G. 2008. A new species of Eurytoma (Hymenoptera: Eurytomidae) attacking

612 spp. (Hymenoptera: Eulophidae) galling Erythrina spp. (Fabaceae), with a summary of

613 African Eurytoma biology and species checklist. Zootaxa, 1751: 1–24.

614

615 Giacalone, C., 2011. Il controllo di Bactrocera oleae (Rossi) e di altri carpofagi negli oliveti biologici in

616 Sicilia e Sud Africa. PhD thesis, Entomology Department, Palermo University, Palermo, Italy. Available

617 from: https://iris.unipa.it/retrieve/handle/10447/95135/121949/merged_document_3.pdf

618

619 Green, P.S. 2002. A revision of Olea L. (Oleaceae). Kew Bulletin, 57: 91-140. doi:10.2307/4110824

620

621 Hamerlinck, G., Hulbert, D., Hood, G.R., Smith, J.J. and Forbes, A.A. 2016. Histories of host shifts and co-

622 speciation among free-living parasitoids of DraftRhagoletis flies. J. Evol. Bio. 29(9): 1766–1779.

623 doi:10.1111/jeb.12909. PMID: 27234648.

624

625 Hebert, P.D., Cywinska, A., Ball, S.L. and deWaard, J.R. 2003. Biological identifications through DNA

626 barcodes. Proc. Royal Soc. B: Bio. Sci. 270(1512): 313–321. doi:10.1098/rspb.2002.2218. PMID:

627 12614582.

628

629 Hebert, P.D.N., Penton, E.H., Burns, J.M., Janzen, D.H., and Hallwachs, W. 2004. Ten species in one: DNA

630 barcoding reveals cryptic species in the Neotropical skipper butterfly Astraptes fulgerator. Proc. Natl.

631 Acad. Sci. USA. 101(41): 14812–14817. doi:10.1073pnas.0406166101. PMID: 15465915.

632

633 Hebert, P.D. and Gregory, T.R. 2005. The promise of DNA barcoding for . Syst. Biol. 54(5): 852–

634 859. doi:10.1080/10635150500354886. PMID: 16243770.

635

26 https://mc06.manuscriptcentral.com/genome-pubs Page 27 of 48 Genome

636 Hoelmer, K.A., Kirk, A.A, Pickett, C.H., Daane, K.M., and Johnson, M.W. 2011. Prospects for improving

637 biological control of olive fruit fly, Bactrocera oleae (Diptera: Tephritidae), with introduced parasitoids

638 (Hymenoptera). Biocontrol Sci. Technol. 21(9): 1005–1025. doi:10.1080/09583157.2011.594951.

639

640 Hood, G.R., Forbes, A.A., Powell, T.H., Egan, S.P., Hamerlinck, G. and Smith, J.J., et al. 2015. Sequential

641 divergence and the multiplicative origin of community diversity. Proc. Natl. Acad. Sci. USA. 112(44):

642 E5980–E5989. doi: 10.1073/pnas.1424717112. PMID: 26499247.

643

644 Kearse, M., Moir, R., Wilson, A., Stones-Havas, S., Cheung, M., Sturrock, S., et al. 2012. Geneious Basic:

645 an integrated and extendable desktop software platform for the organization and analysis of sequence

646 data. Bioinformatics, 28(12): 1647-1649. doi: 10.1093/bioinformatics/bts199 PMID: 22543367.

647 Draft

648 Al Khatib, F., Fusu, L., Cruaud, A., Gibson, G., Borowiec, N., Rasplus, J.-Y., et al. 2014. An integrative

649 approach to species discrimination in the Eupelmus urozonus complex (Hymenoptera, ), with

650 the description of 11 new species from the Western Palaearctic. Syst. Entomol. 39(4): 806-862.

651 doi.org/10.1111/syen.12089.

652

653 Al Khatib, F., Cruaud, A., Fusu, L., Genson, G., Rasplus, J.-Y., Ris, N., et al. 2016. Multilocus phylogeny and

654 ecological differentiation of the “Eupelmus urozonus species group” (Hymenoptera, Eupelmidae) in the

655 West-Palaearctic. BMC Evol. Biol. 16(1): 13. doi: 10.1186/s12862-015-0571-2. PMID:26781031.

656

657 Kimura, M. 1980. A simple method for estimating evolutionary rate of base substitutions through

658 comparative studies of nucleotide sequences. J. Mol. Evol. 16(2):111–120. PMID:7463489.

659

27 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 28 of 48

660 Kumar, S., Stecher, G. and Tamura, K. 2016. MEGA7: Molecular Evolutionary Genetics Analysis version

661 7.0 for bigger datasets. Mol. Bio. Evol. 33(7): 1870–1874. doi: 10.1093/molbev/msw054.

662 PMID:27004904.

663

664 Lotfalizadeh, H., Delvare, G., and Rasplus, J.-Y. 2007. Phylogenetic analysis of Eurytominae (Chalcidoidea:

665 Eurytomidae) based on morphological characters. Zool. J. Linnean Soc. 151: 441–510.

666 doi:10.1111/j.1096-3642.2007.00308.x.

667

668 Leite, L.A.R. 2012. Mitochondrial pseudogenes in insect DNA barcoding: differing points of view on the

669 same issue. Biota Neotrop. 12(3): 301–308. doi:10.1590/S1676-06032012000300029.

670

671 Matsuo, K., Uechi, N., Tokuda, M., Maeto, K.,Draft and Yukawa, J. 2016. Host range of braconid species

672 (Hymenoptera: Braconidae) that attack Asphondyliini (Diptera: ) in Japan. Entomol. Sci.

673 19(1): 3–8. doi: 10.1111/ens.12167.

674

675 Miller, M.A., Pfeiffer, W. and Schwartz, T. 2010. Creating the CIPRES Science Gateway for inference of

676 large phylogenetic trees. In Proceedings of the Gateway Computing Environments Workshop, 14 Nov.

677 2010, New Orleans, LA. pp.1 - 8. doi:10.1109/GCE.2010.5676129.

678

679 Min, X.J. and Hickey D.A. 2007. Assessing the effect of varying sequence length on DNA barcoding of

680 fungi. Mol. Ecol. Notes. 7(3): 365-373. doi:10.1111/j.1471-8286.2007.01698.x. PMID: 18784789.

681

682 Mkize, N., Hoelmer, K.A. and Villet, M.H. 2008. A survey of fruit-feeding insects and their parasitoids

683 occurring on wild olives, Olea europaea ssp. cuspidata, in the Eastern Cape of South Africa. Biocontrol

684 Sci. Technol. 18(10): 991–1004. doi:10.1080/09583150802450154.

28 https://mc06.manuscriptcentral.com/genome-pubs Page 29 of 48 Genome

685

686 Munro, H.K., 1926. Biological notes on South-African Trypaneidae (Trypaneidae: fruit flies). II. Entomol.

687 Mem. S. Afr. Dep. Agric. 5: 17–40.

688

689 Munro, H.K. 1984. A taxonomic treatise on the Dacidae (Tephritidae, Diptera) of Africa. Entomol. Mem.

690 S. Afr. Dep. Agric. 61: I-IX, 1-313.

691

692 Nardi, F., Carapelli, A., Boore, J.L., Roderick, G.K., Dallai, R. and Frati, F. 2010. Domestication of olive fly

693 through a multi-regional host shift to cultivated olives: comparative dating using complete mitochondrial

694 genomes. Mol. Phyl. Evol. 57(2): 678–686. doi:10.1016/j.ympev.2010.08.008. PMID: 20723608.

695

696 Nardi, F., Carapelli, Dallai, R., Roderick, G.K.Draft and Frati, F. 2005. Population structure and colonization

697 history of the olive fly, Bactrocera oleae (Diptera, Tephritidae). Mol. Ecol. 14(9): 2729–2738. doi:

698 10.1111/j.1365-294X.2005.02610.x. PMID: 16029474.

699

700 Neuenschwander, P., 1982. Searching parasitoids of Dacus oleae (Gmel.) (Dipt., Tephritidae) in South

701 Africa. Zeit. Ang. Entomol. 94(1): 509–522. doi:10.1111/j.1439-0418.1982.tb02598.x.

702

703 Nieves-Aldrey, J.L., Hernández Nieves, M., and Gómez, J.F. 2007. A new afrotropical species

704 of Ormyrus Westwood, 1832 (Hymenoptera, Chalcidoidea, ). Graellsia 63(1): 53-60. doi:

705 10.3989/graellsia.2007.v63.i1.80.

706

707 Ramezani, S., Blibech, I., Trindade-Rei, F., van Asch, B. and Teixeira da Costa, L. 2015. Bactrocera oleae

708 (Diptera: Tephritidae) in Iran: An invasion from the Middle West. European J. Entomol. 112(4): 713–721.

709 doi:10.14411/eje.2015.097.

29 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 30 of 48

710

711 Ratnasingham, S. and Hebert, P.D. 2007. Bold: The Barcode of Life Data System

712 (http://www.barcodinglife.org). Mol. Ecol. Notes. 7(3): 355-364. doi:10.1111/j.1471-8286.2007.01678.x.

713 PMID: 18784790.

714

715 Rice, R.E., Philips, P.A., Stewart-Leslie, J., Sibbett, G.S. 2003. Olive fruit fly populations measured in

716 Central and Southern California. Calif. Agric. 57(4): 122–127.

717

718 Ross, H.A., Lento, G.M., Dalebout, M.L., Goode, M., Ewing, G., McLaren, P., et al. 2003. DNA Surveillance:

719 Web-based molecular identification of whales, dolphins, and porpoises. J. Heredity, 94(2): 111–114.

720 doi:10.1093/jhered/esg027. PMID: 12721222.

721 Draft

722 Ross, H.A., Murugan, S. and Li, W.L.S. 2008. Testing the reliability of genetic methods of species

723 identification via simulation. Syst. Biol. 57(2): 216–230. doi:10.1080/10635150802032990. PMID:

724 18398767.

725

726 Rowley, D.L., Coddington, J.A., Gates, M.W., Norrbom, A.L., Ochoa, R.A., Vandenberg, N.J., et al. 2007.

727 Vouchering DNA-barcoded specimens: test of a nondestructive extraction protocol for terrestrial

728 arthropods. Mol. Ecol. Notes. 7(6): 915–924. doi:10.1111/j.1471-8286.2007.01905.x.

729

730 Rubio De Casas, R., Besnard, G., Schönswetter, P., Balaguer, L. and Vargas, P. 2006. Extensive gene flow

731 blurs phylogeographic but not phylogenetic signal in Olea europaea L. Theo. App. Genet. 113(4): 575–

732 583. doi:10.1007/s00122-006-0306-2. PMID: 16835765.

733

30 https://mc06.manuscriptcentral.com/genome-pubs Page 31 of 48 Genome

734 Rugman-Jones, P.F., Wharton, R., van Noort, T. and Stouthamer, R. 2009. Molecular differentiation of

735 the Psyttalia concolor (Szépligeti) species complex (Hymenoptera: Braconidae) associated with olive fly,

736 Bactrocera oleae (Rossi) (Diptera: Tephritidae), in Africa. Biol. Control 49(1): 17–26. doi:

737 10.1016/j.biocontrol.2008.12.005.

738

739 Sambrook, J., Fritsch, E. and Maniatis, T. 1989. Molecular cloning: a laboratory manual. Cold Spring

740 Harbour, N.Y.: Cold Spring Harbour Laboratory Press.

741

742 Schuler, H., Kern, P., Arthofer, W., Vogt, H., Fischer, M., Stauffer, C., et al. 2016. Wolbachia in parasitoids

743 attacking native European and introduced Eastern cherry fruit flies in Europe. Environ. Entomol. 45(6):

744 1424–1431. doi:10.1093/ee/nvw137. PMID: 28028089.

745 Draft

746 Silvestri, F. 1913. Viaggio in Africa per cercare parassiti di mosche dei frutti. Boll. Lab. Zoo. Gen. Agr. R.

747 Scu. Sup. d’Agric. Portici, 8: 1-164.

748

749 Silvestri, F. 1915. Contributo alla conoscenza degli insetti dell’ olivo dell’ Eritrea e dell’ Africa meridionale.

750 Boll. Lab. Zoo. Gen. Agr. Portici, 9: 240–334.

751

752 Silvestri F. 1939. La lotta biologica contro le mosche dei frutti della famiglia Trypetidae. In Proceedings of

753 VII International Congress of Entomology, 15-20 August, Berlin, Germany. 4: 2396-2418.

754

755 Smith, M.A., Bertrand, C., Crosby, K., Eveleigh, E.S., Fernandez-Triana, J., Fisher, B.L., et al. 2012.

756 Wolbachia and DNA barcoding insects: patterns, potential, and problems. PLoS ONE. 7(5). doi:

757 10.1371/journal.pone.0036514. PMID: 22567162.

758

31 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 32 of 48

759 Stamatakis, A. 2006. Phylogenetic models of rate heterogeneity: a high performance computing

760 perspective. In Proceedings of the 20th IEEE International Parallel & Distributed Processing Symposium

761 (IPDPS2006). Washington: IEEE Computer Society Press. pp. 278–286. doi:10.1109/IPDPS.2006.1639535.

762

763 Stamatakis, A. 2014. RAxML version 8: a tool for phylogenetic analysis and post-analysis of large

764 phylogenies. Bioinformatics, 30(9): 1312-3. doi:10.1093/bioinformatics/btu033. PMID: 24451623.

765

766 Thompson, J.D., Higgins, D.G. and Gibson, T.J. 1994. CLUSTAL W: improving the sensitivity of progressive

767 multiple sequence alignment through sequence weighting, position-specific gap penalties and weight

768 matrix choice. Nucleic Acids Res. 22(22): 4673–4680. PMID: 7984417.

769

770 Wharton, R. and Yoder, M. Parasitoids of Fruit-InfestingDraft Tephritidae [online]. Available from

771 http://paroffit.org [accessed 10 March 2017].

772

773 Zohary, D. 1994. The wild genetic resources of the cultivated olive. Acta Hortic. 356: 62-65. doi:

774 10.17660/ActaHortic.1994.356.12.

775

776

32 https://mc06.manuscriptcentral.com/genome-pubs Page 33 of 48 Genome

777 MAIN FIGURE LEGENDS

778

779 Figure 1. Areas of collection of wild and cultivated olives in the Western Cape (16) and in the Eastern

780 Cape (1) Provinces of South Africa. Pie charts represent the relative proportion of Braconidae and

781 Chalcidoidea species reared from olives collected in each area. The size of the circles is proportional to

782 the total number of adult wasp specimens recovered from each area. Black dots represent collection

783 areas from which no specimens were reared.

784

785 Figure 2. A) Bracon celer female; B) Bracon celer male; C) Psyttalia humilis female; D) Psyttalia humilis

786 male.

787

788 Figure 3. A) Psyttalia lounsburyi female; B)Draft Psyttalia lounsburyi male. C) Utetes africanus, female; D)

789 Utetes africanus male.

790

791 Figure 4. A) Eupelmus spermophilus female; B) Eupelmus spermophilus male; C) Eurytoma oleae female;

792 D) Eurytoma oleae male; E) Eurytoma varicolor female; F) Eurytoma varicolor male.

793

794 Figure 5. A) Neochrysocharis formosus female; B) Neochrysocharis formosus male; C) Sycophila

795 aethiopica female; D) Sycophila aethiopica male; E) Ormyrus sp. female; and F) Ormyrus sp. male.

796

797 Figure 6. Relative proportions of adult braconid and chalcid wasps reared from wild and cultivated olives

798 collected in 16 areas in the Western Cape Province and one area in the Eastern Cape Province of South

799 Africa.

800

33 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 34 of 48

801 Figure 7. Neighbour-joining (K2P) tree of the family Braconidae, based on a 485 bp alignment of 128 COI

802 sequences and two outgroups, with pairwise deletion of sites. Values indicate nodal bootstrap support

803 (1,000 replicates). The scale bar represents the percentage of sequence divergence. Species surveyed in

804 this study are shown in bold underlined. Triangles represent condensed species clades.

805

806 Figure 8. Neighbour-joining (K2P) tree of the superfamily Chalcidoidea, based on a 460 bp alignment of

807 225 COI sequences and two outgroups, with pairwise deletion of sites. Values indicate nodal bootstrap

808 support (1,000 replicates). The scale bar represents the percentage of sequence divergence. Species

809 surveyed in this study are shown in bold underlined. Triangles represent condensed species clades.

810

811 Draft

34 https://mc06.manuscriptcentral.com/genome-pubs Page 35 of 48 Genome

812 SUPPLEMENTARY TABLE AND FIGURE LEGENDS

813

814 Table S1. List of specimens of Braconidae and Chalcidoidea wasps associated with wild and cultivated

815 olives in the Western Cape Province of South Africa, as deposited in the entomology collection at the

816 Iziko South African Museum in Cape Town.

817

818 Table S2. List of primers used for the PCR amplification and bidirectional sequencing of the barcoding

819 region of the COI gene in Braconidae and Chalcidoidea wasps reared from wild and cultivated olives in

820 the Western Cape and Eastern Cape Provinces of South Africa. *Previously published (Folmer et al,

821 1994); **Developed in this study.

822

823 Table S3. List of publicly available and newDraft COI sequences of Braconidae and Chalcidoidea used in the NJ

824 and ML analyses of wasps reared from wild and cultivated olives in the Western and the Eastern Cape

825 Provinces of South Africa, with Genbank accession numbers and BOLD identification numbers.

826 Sequences generated in this study are shown in bold underlined.

827

828 Table S4. Number of tephritid flies, and braconid and chalcid adult specimens reared from wild and

829 cultivated olives, and apparent parasitism and seed wasp estimates of infestation rates in 16 areas in the

830 Western Cape and one area in the Eastern Cape (Grahamstown) Provinces of South Africa. AIR %* -

831 Apparent seed infestation rate, calculated as the total number of adult specimens of the particular

832 species / total number of olives at each collection site and date. APR %* - Apparent parasitism rate,

833 calculated as the total number of adult braconid and Neochrysocharis formosus specimens / (total

834 number of tephritid flies + total number of braconids + N. formosus) at each collection site and date. Nd

835 – No data.

836

35 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 36 of 48

837 Table S5. Specimen data of Braconidae and Chalcidoidea adult specimens reared from wild and

838 cultivated olives collected in 17 areas in the Western Cape and Eastern Cape Provinces of South Africa,

839 and used for DNA barcoding and phylogenetic analyses. Geographic coordinates indicate the site of olive

840 collection. Primer pair codes can be found in Table S1. Nd – No data.

841

842 Table S6. a), b) and c) Minimum, maximum and mean intraspecific p-distances (K2P) for Psyttalia species

843 based on a 609 bp alignment of 63 COI sequences, with pairwise deletion of sites. n - Number of

844 sequences used in estimates for each species; d), e) and f) Mean interspecific and intergroup p-distances

845 (K2P) for Psyttalia species based on a 609 bp alignment of 63 COI sequences, with pairwise deletion of

846 sites. Standard errors are shown above the diagonal.

847

848 Table S7. a) Minimum, maximum and meanDraft intraspecific p-distances (K2P) for two Utetes species based

849 on a 515 bp alignment of COI sequences, with pairwise deletion of sites. n - Number of sequences used

850 in estimates for each species; b) Mean interspecific p-distances (K2P) for U. africanus and three U.

851 canaliculatus NJ clusters (see Fig. S3) based on a 515 bp alignment of COI sequences, with pairwise

852 deletion of sites. Standard errors are shown above the diagonal.

853

854 Table S8. a) Minimum, maximum and mean intraspecific p-distances (K2P) for nine Eupelmus species

855 based on a 532 bp alignment of 150 COI sequences, with pairwise deletion of sites. n - Number of

856 sequences used in estimates for each species; b) Mean interspecific p-distances (K2P) for nine Eupelmus

857 species based on a 532 bp alignment of 150 COI sequences, with pairwise deletion of sites. Standard

858 errors are shown above the diagonal.

859

860 Table S9. a) Minimum, maximum and mean intraspecific p-distances (K2P) for seven Eurytoma species

861 based on a 530 bp alignment of 74 COI sequences, with pairwise deletion of sites. n - Number of

36 https://mc06.manuscriptcentral.com/genome-pubs Page 37 of 48 Genome

862 sequences used in estimates for each species; b) Mean interspecific p-distances (K2P) for seven

863 Eurytoma species, based on a 532 bp alignment of 74 COI sequences, with pairwise deletion of sites.

864 Standard errors are shown above the diagonal.

865

866 Table S10. Minimum, maximum and mean intraspecific p-distances (K2P) for Neochrysocharis based on

867 633 bp alignments of COI sequences, with pairwise deletion of sites. n - Number of sequences used in

868 estimates for each group. na – not applicable.

869

870 Figure S1. Neighbour-joining (K2P) tree of the Bracon genus, based on a 524 bp alignment of 22 COI

871 sequences and two outgroups, with pairwise deletion of sites. Values indicate nodal bootstrap support

872 (1,000 replicates). The scale bar represents the percentage of sequence divergence. Sequences

873 generated in this study are shown in bold underlinedDraft.

874

875 Figure S2. Neighbour-joining (K2P) tree of the Psyttalia genus, based on a 524 bp alignment of 54 COI

876 sequences and two outgroups, with pairwise deletion of sites. Values indicate nodal bootstrap support

877 (1,000 replicates). The scale bar represents the percentage of sequence divergence. Sequences

878 generated in this study are shown in bold underlined.

879

880 Figure S3. Neighbour-joining (K2P) tree of the Utetes genus, based on a 485 bp alignment of 52 COI

881 sequences and two outgroups, with pairwise deletion of sites. Values indicate nodal bootstrap support

882 (1,000 replicates). The scale bar represents the percentage of sequence divergence. Sequences

883 generated in this study are shown in bold underlined.

884

885 Figure S4. Neighbour-joining (K2P) tree of the Eupelmus genus, based on a 545 bp alignment of 99 COI

886 sequences and two outgroups, with pairwise deletion of sites. Values indicate nodal bootstrap support

37 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 38 of 48

887 (1,000 replicates). The scale bar represents the percentage of sequence divergence. Sequences

888 generated in this study are shown in bold underlined.

889

890 Figure S5. Neighbour-joining (K2P) tree of the Eurytoma genus, based on a 536 bp alignment of 51 COI

891 sequences and two outgroups, with pairwise deletion of sites. Values indicate nodal bootstrap support

892 (1,000 replicates). The scale bar represents the percentage of sequence divergence. Sequences

893 generated in this study are shown in bold underlined.

894

895 Figure S6. Neighbour-joining (K2P) tree of the Sycophila genus, based on a 549 bp alignment of 71 COI

896 sequences and two outgroups, with pairwise deletion of sites. Values indicate nodal bootstrap support

897 (1,000 replicates). The scale bar represents the percentage of sequence divergence. Sequences

898 generated in this study are shown in bold underlinedDraft.

899

900 Figure S7. Neighbour-joining (K2P) tree of the Neochrysocharis genus and the sister group Asecodes

901 lucens, based on a 561 bp alignment of 10 COI sequences and two outgroups, with pairwise deletion of

902 sites. Values indicate nodal bootstrap support (1,000 replicates). The scale bar represents the percentage

903 of sequence divergence. Sequences generated in this study are shown in bold underlined.

904

905 Figure S8a. Overall apparent parasitism rate (%) in wild and cultivated olives collected across all 17 areas

906 in the Western Cape and the Eastern Cape Provinces of South Africa. Overall apparent parasitism rate =

907 total number of adult braconid and Neochrysocharis formosus* specimens / (total number of tephritid

908 flies + the total number of adult braconid and N. formosus* specimens). *Neochrysocharis formosus is a

909 with a parasitoid lifestyle.

910

38 https://mc06.manuscriptcentral.com/genome-pubs Page 39 of 48 Genome

911 Figure S8b. Overall apparent seed infestation rate (%) in wild and cultivated olives collected across all 17

912 areas in the Western Cape and the Eastern Cape Provinces of South Africa. Overall apparent seed

913 infestation rate = total number of adult chalcid specimens / total number of wild and cultivated olives.

914

915 Figure S9. Apparent parasitism and apparent seed infestation rates (%) in cultivated olives collected from

916 three areas in the Western Cape Province of South Africa. Apparent parasitism rate = number of adult

917 braconid specimens / (number of tephritid flies + number of adult braconid specimens). Apparent seed

918 infestation rate = number of adult chalcid specimens / number of cultivated olives. Bracon celer (BR) and

919 Psyttalia lounsburyi (PL) are braconid wasps. Eupelmus spermophilus (ES) is a chalcid wasp.

920

921 Figure S10. Apparent parasitism and seed infestation rates (%) in wild olives collected from 11 areas in

922 the Western Cape Province and one area inDraft the Eastern Cape Province of South Africa. Apparent

923 parasitism rate = number of adult braconid and Neochrysocharis formosus* specimens / (number of

924 tephritid flies + number of adult braconid and N. formosus* specimens). Apparent seed infestation rate =

925 number of adult chalcid specimens / number of wild olives. Bracon celer (BR), Psyttalia humilis (PS),

926 Psyttalia lounsburyi (PL) and Utetes africanus (UA) are braconid wasps. Eupelmus spermophilus (ES),

927 Eurytoma oleae (EYO), Eurytoma varicolor (EYV), Sycophila aethiopica (SY) and Ormyrus sp. (OR) are

928 chalcid wasps. *Neochrysocharis formosus (NF) is a chalcid wasp with a parasitoid lifestyle.

929

930

931

932

933

934

935 39 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 40 of 48

936

Draft

40 https://mc06.manuscriptcentral.com/genome-pubs Page 41 of 48 Genome

Eupelmus spermophilus Eurytoma oleae Eurytoma varicolor Sycophila aethiopica SOUTH AFRICA NORTHERN CAPE Psyttalia humilis SOUTH AFRICA Psyttalia lounsburyi Bracon celer Neochrysocharis formosus Utetes africanus Ormyrus striatus

Riebeek Kasteel Draft

Porterville Malmesbury Tulbagh WESTERN CAPE EASTERN CAPE Paarl Grahamstown Worcester

Franschhoek Brackenfell Bonnievale Stellenbosch Riversdale Cape Town Swellendam McGregor

Stanford INDIAN OCEAN https://mc06.manuscriptcentral.com/genome-pubs Somerset West 100 km Genome Page 42 of 48

Draft

Figure 2. A) Bracon celer female; B) Bracon celer male; C) Psyttalia humilis female; D) Psyttalia humilis male.

220x165mm (300 x 300 DPI)

https://mc06.manuscriptcentral.com/genome-pubs Page 43 of 48 Genome

Draft

Figure 3. A) Psyttalia lounsburyi female; B) Psyttalia lounsburyi male. C) Utetes africanus, female; D) Utetes africanus male.

220x165mm (300 x 300 DPI)

https://mc06.manuscriptcentral.com/genome-pubs Genome Page 44 of 48

Draft

Figure 4. A) Eupelmus spermophilus female; B) Eupelmus spermophilus male; C) Eurytoma oleae female; D) Eurytoma oleae male; E) Eurytoma varicolor female; F) Eurytoma varicolor male.

220x248mm (300 x 300 DPI)

https://mc06.manuscriptcentral.com/genome-pubs Page 45 of 48 Genome

Draft

Figure 5. A) Neochrysocharis formosus female; B) Neochrysocharis formosus male; C) Sycophila aethiopica female; D) Sycophila aethiopica male; E) Ormyrus sp. female; and F) Ormyrus sp. male.

220x248mm (300 x 300 DPI)

https://mc06.manuscriptcentral.com/genome-pubs Genome Page 46 of 48

Draft

https://mc06.manuscriptcentral.com/genome-pubs Page 47 of 48 99 Genome Utetes canaliculatus

Utetes magnus

Utetes frequens Utetes canaliculatus 10 99 Utetes canaliculatus 95 Utetes juniperi 100 Utetes rosicola Utetes anastrephae

Utetes africanus 100 Utetes canaliculatus 100 Bracon greeni 100 Bracon atrator

Bracon intercessor 100 Bracon crassicornis 80 100 Bracon kopelkei Draft95 Bracon picticornis Bracon celer Br02 87 Bracon asphondylae 9 100 99 Bracon tamabae Bracon asphondylae 100 100 98 Psyttalia carinata Psyttalia ponerophaga 100 Psyttalia phaeostigma 86 100 99 Psyttalia fletcheri Psyttalia incisi Psyttalia cf. perproxima PFRJ-2008 100 Psyttalia cosyrae 100 Psyttalia lounsburyi 100 Psyttalia perproxima 100 Psyttalia concolor

95 Psyttalia humilis 100 Eurytoma spermophilus ES99 100 Eurytoma oleae EY14

https://mc06.manuscriptcentral.com/genome-pubs 0.02 Genome Page 48 of 48

Sycophila sp.

Sycophila aethiopica SY50 Sycophila sp. 83 Sycophila sp.

Sycophila sp. 99 Eurytoma striolata Eurytoma morio 99 Eurytoma oleae 99 Eurytoma maura 99 Eurytoma afra 99 Eurytoma gigantea Eurytoma juniperina Eurytoma discordans Eurytoma aciculata 99 Eurytoma laricis 99 Eurytoma arctica Eurytoma longavena 99 Eupelmus varicolor 99 99 Eupelmus spermophilus Eupelmus matranus Eupelmus testaceiventris Eupelmus linearis 99 Eupelmus cushmani Eupelmus falcatus Eupelmus phragmitis Draft Eupelmus fuscipennis Eupelmus cicadae Eupelmus juniperinus thuriferae Eupelmus atropurpureus Eupelmus seculatus Eupelmus microzonus 99 Eupelmus tibicinis Eupelmus muellneri Eupelmus simizonus Eupelmus vesicularis 9 Eupelmus pini 99 99 82 Eupelmus gemellus Eupelmus martellii Eupelmus acinellus 99 Eupelmus annulatus 99 99 Eupelmus vindex Eupelmus azureus 99 Eupelmus cerris 99 Eupelmus confusus Eupelmus pistaciae 99 Eupelmus fulgicalvus 99 Eupelmus fulvipes 99 99 Eupelmus kiefferi Eupelmus tremulae 99 93 Eupelmus opacus 91 Eupelmus priotoni 97 Eupelmus purpuricollis 99 Eupelmus janstai 99 Eupelmus minozonus 98 Eupelmus urozonus 99 Neochrysocharis formosus HM365028 Neochrysocharis clinias HM365038 Neochrysocharis formosus 99 Psyttalia humilis Ps24 99 https://mc06.manuscriptcentral.com/genome-pubsBracon celer Br02

0.05