Development and characterization of microsatellite markers in the African timber tree species Cylicodiscus gabunensis (Fabaceae)

Cylicodiscus gabunensis (Fabaceae) or ‘Okan’ is a Central African multipurpose timber species that is used for heavy construction and traditional medicine. Despite being currently heavily exploited, the species shows a low population density and a natural regeneration deficit in dense forest. We aimed to characterize polymorphic microsatellite markers that can be used to study patterns of genetic structure and gene flow (mating system, pollen and seed dispersal and inbreeding depression) and ultimately, help to develop sustainable forest management practices. We developed 24 polymorphic markers that can be amplified in three PCR multiplexes that were tested in 647 individuals of C. gabunensis from three populations. The number of alleles ranged from 4 to 27 and the average observed and expected heterozygosity across loci and populations were Ho = 0.585 (0.081–0.936) and He = 0.510 (0.076–0.914), respectively. This set of markers is a useful tool for exploring intra-specific diversity, genetic structure and gene flow of C. gabunensis.


Introduction
Cylicodiscus gabunensis HARMS (Fabaceae), locally known as 'Okan', is a very large tree (up to 60 m height) endemic to the Congo Basin and the only species of the genus Cylicodiscus. This non-gregarious, light-demanding species is distributed from Ivory Coast to the Republic of the Congo where it grows in semi-deciduous and dense humid evergreen forests [1]. It bears white-greenish spike-like inflorescences that are hermaphroditic and most-likely pollinated by insects [2]. Fruits are large dehiscent pods containing numerous winged seeds that are wind-dispersed [1,3]. Like several other lightdemanding species, population structure and population density of C. gabunensis are variable according to the forest type and are scarce in dense humid evergreen forests [2,4]. The species is of high socio-economical value. Its dense and hard wood is harvested for heavy constructions including bridges, gates, marine constructions and heavy floor [1]. The tree is also sacred for some communities and its bark and leaves are used for multiple purposes including traditional medicine [2,3].
Cylicodiscus gabunensis was still considered as a secondary species not highly desired a few years ago but there has been a rapid growth of interest the last two decades and it is now the sixth most exploited species in the Congo Basin [5,6]. Despite the high socio-economical value of C. gabunensis, very little is known about its ecological and genetical characteristics [2]. Several reproductive, ecological, and genetic processes including seed and pollen dispersal can affect gene flow and spatial genetic structure (SGS), i.e. the non-random spatial distribution of genotypes [7]. By decreasing population density, selective logging is expected to increase SGS because of higher local genetic drift [8]. Intra-specific genetic diversity and gene flow also influence the reproductive success and the capacity of tree populations to adapt to habitat fragmentation. Understanding the processes involved in the reproduction of C. gabunensis is therefore crucial to develop sustainable forest management practices, including the definition of a minimal diameter of 1 3 exploitation and a minimal post-exploitation density allowing to maintain a sufficient potential for regeneration and species adaptation [4,6,9,10].
Nuclear simple sequence repeat markers (SSR), or microsatellites, have gained considerable importance in plant genetics because of their high level of transferability to related species, co-dominant inheritance, important genome coverage, high polymorphism and amenability to high-throughput genotyping [11]. They are used to study the mating system, gene flow, inbreeding depression and the determinants of reproductive success through parentage analyses, for example in African trees [12][13][14]. They are also used to study the genetic structure of populations, phylogeographic patterns and other studies of genetic variation [8,14]. The natural regeneration deficit and low observed densities of C. gabunensis in dense forest suggest that special precautions should be taken to achieve sustainable exploitation of this species [2,5]. Therefore, we aimed to develop polymorphic microsatellites primers for C. gabunensis using high-throughput sequencing. These nuclear markers will be characterized in three populations to study patterns of genetic structure and gene flow (mating system, pollen and seed dispersal and inbreeding depression).

Microsatellite development
DNA was isolated from 15 to 25 mg of silica dried leaf samples of two Cylicodiscus gabunensis individuals collected in the Republic of Congo (OH5903, 3.500°S, 12.305°E) and in Cameroon (JFG1811, 3.855°N, 14.768°E), using the Nucleospin 96 Plant II kit (Macherey-Nagel, Duren, Germany). We sheared these DNA extracts using a Biorup-tor® Pico (Diagenode SA.) in volume 100 μL of TE 1X pH 8 (Life Technologies AM9858), 6 cycles of 12 s ON and 90 s OFF to obtain fragments of c. 500 bp. We further used the AMPure XP protocol (Agencourt) too remove too small (< 400 bp) and too large (> 800 bp) DNA fragments. We then prepared non-enriched genomic libraries using the NEBNext Ultra II DNA Library Prep Kit for Illumina (New England Biolabs E7645) and the protocol of Mariac et al. [15]. These libraries were sequenced on an Illumina MiSeq platform at GIGA (Liège, Belgium). The sequencing run generated 1.8 to 2.0 million paired-end 250 bp reads, which were merged with FLASH [16] resulting in 1.5 to 1.6 million sequences with a modal length of 360 to 390 bp (up to 492 bp). We used the QDD pipeline [17] to identify microsatellite markers, for which suitable primers were proposed using the parameters described in Malausa et al. [18]. More than 30,000 microsatellite loci were detected. After selecting pure microsatellites with at least 12 repetitions of di-, tri-or tetra-nucleotide repeats, with primers situated at least 16 bp from the microsatellite sequence and a primer design 'A' (i.e. best quality according to QDD output), we ordered primers for 45 loci. On each forward primer we added one of four universal sequences (Q1-Q4, Table 1) associated to fluorescent dyes (6-FAM, NED, VIC and PET, respectively), allowing low-cost M13-like fluorescent labelling [19].

Results and Discussion
Microsatellite sequences were deposited on GenBank (Table 1). One or two alleles per locus were observed, indicating that C. gabunensis is diploid. The 24 markers genotyped in the three populations showed 4 to 27 alleles with an average of 9.96 alleles per locus. PIC values ranged from 0.108 and 0.934 with an average of 0.560 ( Table 2). The average expected and observed heterozygosity were respectively 0.585 (0.081-0.936) and 0.510 (0.076-0.914). Of the 24 loci studied, nine loci significantly deviated from HWE in at least one of the populations, five of them usually showing heterozygosity deficiency while four of them showed slight heterozygote excess in population Alpicam (Table 2). Heterozygosity deficiency was associated with higher frequencies of null alleles suggesting that it is probably not a consequence of inbreeding, as supported by the absence of inbreeding in each population according to INEst. Of the 24 loci characterized in this study, eight showed the presence of null alleles in at least one population. However, the frequencies of null alleles were always below 0.20 except for three markers in at least one population (Cyl07, Cyl14 and Cyl36; Table 2).
In conclusion, we developed 24 microsatellite markers that can be used to study intra-specific diversity, genetic structure and gene flow of C. gabunensis. They are also expected to be polymorphic enough to conduct paternity or parentage analyses. The biological knowledge that can be acquired with these markers will help determine which practices allow a sustainable exploitation of this timber species. The number of microsatellite loci developed is expected to be sufficient to obtain accurate measures of genetic variation and structure.

Conflict of interest
The authors have no relevant financial or non-financial interests to disclose.
Ethical approval This article does not contain any studies with human participants or animals performed by any of the authors.