Cross-amplification and characterization of microsatellite markers in species of Manihot Mill. (Euphorbiaceae) endemic to the Brazilian Cerrado

The genus Manihot Mill. contains about 120 species of which about 104 occur in Brazil. We tested the cross-amplification of ten microsatellite markers developed for Manihot esculenta in 15 species of Manihot endemic to the Brazilian Cerrado. We also evaluated the genetic diversity of Manihot irwinii , M. orbicularis , and M. purpureocostata . Ten pairs of primers were amplified among 14 species of Manihot . The percentage of polymorphic loci per species varied from 70 to 100 %. Nine markers showed amplification and polymorphism when evaluated on polyacrylamide gel. The markers were combined to form three sets for multiplex genotyping for genetic diversity analysis, and showed 51, 75, and 75 alleles in M. irwinii , M. orbicularis , and M. purpureocostata , respectively. The levels of genetic diversity for the transferred markers were high for the three species and proved to be useful for population genetics studies of species of Manihot endemic to the Cerrado. The results of this study will help to better understand the genetic diversity, taxonomy and relationships among species Manihot , and to develop conservation programs for the genus.


Introduction
The genus Manihot has a Mesoamerican origin with its center of diversity in Brazil (Silva 2014). The Cerrado of Central Brazil presents about 104 documented species, so the main center of diversity for the genus within the country, (Silva & Amaral 2020).
Floristic and taxonomic studies have shown that species of this genus have problems of taxonomic delimitation due to the lack of taxonomic studies in Brazil. Nonetheless, several new species have been described in recent years Silva 2016;Silva et al. 2017;Mendoza et al. 2018), showing significant advances in knowledge of the diversity of the genus. However, little is known about the genetic diversity of these species, especially in the Central-West Region of Brazil.
The development of microsatellites or simple sequence repeats (SSRs) provides an ideal tool for investigating patterns of genetic variation due to their codominant inheritance and multiallelic and highly polymorphic properties, as well as being abundant and well distributed throughout the genome (Li et al. 2002;Ellegren 2004). However, given the time-consuming and expensive process of isolating SSRs, it is advantageous to test available microsatellite markers for phylogenetically close species by cross-amplification before investing in the development of species-specific markers.
The success of cross-species amplification depends on the conservation of primer sites within flanking sequences and on the maintenance of sequences that promote polymorphisms (FitzSimmons et al. 1995). Several studies have demonstrated the utility of using primer pairs designed from one species for others of the same genus (Bernardes et al. 2014;Buzatti et al. 2016) and even for species of other genera (Barbará et al. 2007;Santos et al. 2015;Fagundes et al. 2016;Miranda et al. 2016).
Considering the insufficient knowledge available regarding the diversity of wild species of the genus Manihot, along with the taxonomic and phylogenetic complexity of the genus and the lack of SSR and molecular information, we tested microsatellite markers developed for M. esculenta (cultivated cassava) (Chavarriaga-Aguirre et al. 1998;Mba et al. 2001) by cross-amplification in 15 congeners endemic to the Cerrado. Moreover, we evaluated the genetic diversity of the markers in three of these species (M. irwinii, M. orbicularis, and M. purpureocostata), to provide tools for population genetics studies on species of Manihot endemic to the Cerrado.

Plant material and cross-species amplification
Amplification tests used leaves sampled from three individuals from each of the know 15 species of Manihot of the Cerrado, Goiás State, Central-West Brazil (Tab. 1). Marker polymorphism evaluation used eight individuals per species. A standard protocol of 2 % CTAB was used for DNA extraction (Doyle & Doyle 1987

Voucher/ Herbarium Universidade Federal de Goiás
The annealing temperature of the primers was adjusted for each marker until an acceptable amplification pattern was found on 3 % agarose gel (Tab. 2). Polymorphism of the markers in the 15 species of Manihot was determined through standard 6 % acrylamide gel electrophoresis visualized by silver staining procedures (Creste et al. 2001). Allele size was determined by reference to 10 bp and 50 bp DNA standards (Invitrogen ™).

Genetic variability of polymorphic loci
The markers that showed the best polyacrylamide gel amplification profiles, associated with verified polymorphism in the 15 species, were selected for characterizing the genetic diversity of three Manihot species: M. orbicularis Pohl, M. purpureocostata Pohl, and M. irwinii DJ Rogers & Appan. A sample of 24 individuals, representing known occurrence, was used for each species, for a total of 72 individuals. DNA extraction and amplification followed the same protocols as described above.
Forward sequences of selected primer pairs were labeled with one of four fluorescent dyes: VIC, NED, 6-FAM, or PET. The sizes of amplification products were determined using a GeneScan 600 LIZ internal marker (Applied Biosystems) in an ABI PRISM® 3500 DNA Genetic Analyzer (Applied Biosystems). Microsatellite loci with greater clarity in their amplification detected by capillary electrophoresis were arranged in multiplex panels for analysis of the three Manihot species.
Allele calling was performed using GeneMapper 5.0 software (Applied Biosystems). Genotypes were confirmed using an allelic ladder constructed with all alleles found for each locus in this study. Micro-Checker software (Oosterhout et al. 2004) was then used to detect errors due to stuttering, allele dropout, and null alleles.
The power of individual discrimination with the total loci set and with each locus was evaluated by estimates of probability of genetic identity (I) (Paetkau et al. 1995) and the probability of paternity exclusion (Q) (Weir 1996), using Identity 1.0 software (Wagner & Sefc 1999). Genetic variability analysis, including allelic richness, observed heterozygosity (H o ), and expected heterozygosity under the Hardy-Weinberg equilibrium (H e ), were estimated using Genetic Data Analysis 1.0 software (GDA) (Lewis & Zaykin 2001). Linkage disequilibrium was evaluated using Bonferroni correction in FSTAT 2.9.3.2 software (Goudet 2002).

Results and discussion
The markers amplified in all 15 species of Manihot were dinucleotides, most with the GA motif. The least conserved marker was GAGG5 (tetranucleotide), which did not amplify for one of the 15 species (M. irwinii).
All ten primer pairs tested in the wild Manihot species were polymorphic (100 %) in nine of them (M.   (Roa et al. 2000). Thus, in the present study, the percentage of polymorphic loci obtained from crossamplification among species within the genus Manihot tends to maintain close to 100 %, reflecting high conservation of genomic regions among species. This conservation may be evidence of recent diversification of the genus Manihot, which would explain the low taxonomic resolution for some sets of Manihot species (Duputié et al. 2011 (Roa et al. 2000), showing that, in general, microsatellite primers work throughout the genus. However, as phylogenetic distance increases, successful amplification of loci tends to decrease. This relationship with phylogenetic distance was also observed by Bressan et al. (2012) when reporting cross-amplification with the species Jatropha curca, which belongs to the same family as Manihot (Euphorbiaceae). They found amplification of alleles in species of the same genus, but not in other genera of Euphorbiaceae, such as Hevea brasiliensis, Manihot esculenta, and Ricinus communis. Successful transferability of microsatellite markers among closely related species has also been verified for other Cerrado species, such as Anacardium humile (Soares et al. 2013), Byrsonima cydoniifolia (Bernardes et al. 2014), and Campomanesia adamantium and C. pubescens (Miranda et al. 2016). This success can be explained by the conservation of microsatellite flanking regions in closely related species (FitzSimmons et al. 1995;Barbará et al. 2007).
The markers were polymorphic for most species (eight individuals), varying among 12 to 15 species. The most polymorphic markers were GA 136, GA 134, GA131, and SSRY12, which were polymorphic in all 15 species (Tab. 2). Such high levels of polymorphism were expected according to data in the literature for wild species of the genus Manihot (Roa et al. 2000;Raji et al. 2009).
Out of the ten markers, nine exhibited better amplification and polymorphism patterns and so were combined into three sets for multiplex genotyping and characterization of the loci (Tab. 3). Out of the nine markers with clear amplicons tested on the three wild species of Manihot, four loci had null alleles, the locus GA136 for M. orbicularis and M. irwinii, GA126 for M. orbicularis and M. purpureocostata, GA16 for M. irwinii, and SSRY12 for M. orbicularis. Null alleles have been commonly found in studies of transferability of microsatellite markers in wild species of Manihot (Roa et al. 2000;Raji et al. 2009). However, no null alleles have been observed in M. esculenta (Roa et al. 2000). In this sense, the the occurrence de null alleles may be an artifact in the sequences flanking the microsatellite, caused by transferability (Dabrowski et al. 2015).
No significant changes in linkage disequilibrium (P> 0.05) were found for any pair of loci in any species. Deviations from HWE (P <0.05) were observed at locus GA21 for all species, GA136 for M. irwinii and M. orbicularis, GA16 and SSRY82 for M. irwinii, GA126 for M. orbicularis and M. purpureocostata, and GA131 for M. orbicularis. Among the loci that deviated from HWE, some showed null alleles, which may explain the deviation. Analysis with more populations and a larger number of individuals per population may confirm this result.
The present study detected 51 alleles in M. irwinii, 75 alleles in M. orbicularis and 75 alleles in M. purpureocostata, which demonstrates a high level of polymorphism. These results are similar to those reported in the literature for wild species of the genus Manihot, which found 79 alleles (Roa et al. 2000) and 50 alleles (Silva et al. 2017). Studies of M. esculenta have found 46 alleles (Siqueira et al. 2009), 45 alleles (Roa et al. 2000), and 47 alleles (Aragon et al. 2012). These results follow Roa et al. (2000), who state that the wild species of the genus Manihot have a larger pool of SSR alleles than M. esculenta.
The number of alleles per locus found for the three species of Manihot varied from three to 16, with averages Thus, the high average number of alleles per locus found in the present study suggests that the set of markers used substantially represents the polymorphism of the loci. This can be confirmed by the high polymorphism along with the low probability of identity (1.672 x 10 -7 , 9.13 x 10 -10 , and 2.61 x 10 -9 ) and high power of paternity exclusion (0.999, 0.999, and 0.998), observed for the species M. irwinii, M. orbicularis, and M. purpureocostata, respectively. These results show that the nine markers are suitable for discriminating individuals at each locus under analysis (Paetkau et al. 1995) and demonstrated a high power of paternity exclusion (Weir & Evett 1998), allowing an efficient characterization of the genetic variability existing in populations of wild Manihot species (Tab. 4). The average genetic diversity (H e ) of the markers was high for M. irwinii (74 %), M. orbicularis (82 %), and M. purpureocosta (78 %). These results are equivalent to the maximum genetic diversity expected (0.833, 0.879, and 0.879), according to Hennink & Zeven (1991), considering the number of alleles found per locus.
The average observed heterozygosity (H o = 0.517, 0.615, and 0.669) was lower than the expected (H e =0.618, 0.724, and 0.689; Tab. 4) for M. irwinii, M. orbicularis, and M. purpureocosta, respectively. This result suggests that the heterozygote deficiency may be due to several factors, such as inbreeding (Halsey et al. 2008), limited sample size, and presence of null alleles, with the latter being a common factor in transferability studies with wild species of Manihot (Roa et al. 2000;Raji et al. 2009), and with other species of the Cerrado (Ciampi et al. 2008;Feres et al. 2009;Soares et al. 2013;Fagundes et al. 2016;Miranda et al. 2016).
The present study documented ten polymorphic microsatellite markers for the 15 studied species of Manihot. Nine of these markers were indicated as having high potential to detect genetic variation in the three analyzed wild species of Manihot. Thus, the results of this study are promising and valuable for developing further studies of genetic variability of these species and for studies aiming to properly understand the relationships among the species of the genus Manihot, as well as their genetic diversity, taxonomy, and conservation.