Variability of and seeds of crassifolia () genotypes cultivated in northern Mato Grosso State, Brazil.

J.F.L. Santos1, A.A.B. Rossi1, G.F. Pena1, A.V. Tiago1, K.E.M. Zortéa1, E.S. Cardoso1, E.C.M. Pedri1, I.C.B. Santos1, P.H.A.D. Santos2, D.B. Santos3 and I.R.B. Santos4

1 Laboratório de Genética Vegetal e Biologia Molecular, Universidade do Estado de Mato Grosso, Alta Floresta, MT, Brasil 2 Centro de Ciências e Tecnologias Agropecuárias, Universidade Estadual do Norte Fluminense Darcy Ribeiro, Rio de Janeiro, RJ, Brasil 3 Departamento Educacional, Escola Estadual Ludovico da Riva Neto, Alta Floresta, MT, Brasil 4 Instituto de Ciências e Saúde, Universidade Federal de Mato Grosso, Sinop, MT, Brasil

Corresponding author: A.A.B. Rossi E-mail: [email protected]

Genet. Mol. Res. 19 (2): gmr18620 Received April 16, 2020 Accepted March 31, 2020 Published May 30, 2020 DOI http://dx.doi.org/10.4238/gmr18620

ABSTRACT. Due to the increasing popularity of consuming only products, including vegetarianism and more recently veganism, vegetables and fruits have gradually become an increasingly important component of the human diet, worldwide. The large diversity of Brazilian tropical fruits, such as murici, , (Malpighiaceae), associated with a dearth of information concerning this and other exotic , has stimulated research on techniques to obtain increases in productivity, inserting these species into regional and national agribusiness. We evaluated the variability of B. crassifolia genotypes and quantified the relative contribution of 12 morphological traits of its fruits and seeds. Ripe fruits were collected from 20 different genotypes (trees. The fruits

Genetics and Molecular Research 19 (2): gmr18620 ©FUNPEC-RP www.funpecrp.com.br

J.F.L. Santos et al. 2

were collected during the harvest season from December 2017 to January 2018. Twenty fruits of each genotype were selected; that were visually appealing, whole and without deformation, totaling 400 samples. We measured the length, width, thickness, weight, and volume of the fruits, the length, width, thickness and weight of the seed and the thickness and weight of the pulp. The data were obtained with a digital caliper and the degrees brix measured with a refractometer. Multivariate analyses were made with UPGMA hierarchical methods and Tocher optimization, using the standardized average Euclidean distance as a measure of dissimilarity, and principal components analysis (PCA). Four distinct groups were formed based on the data. The first three principal components explained 87.51% of the total variation. The brix measurement was the trait with the greatest contribution to the differentiation of the genotypes.

Key words: Amazon fruits; Variability; Morphological traits; UPGMA; PCA

INTRODUCTION

Information about the fruits of murici, Byrsonima crassifolia, are scarce; however, the development of research aimed at generating knowledge that allows better use and commercialization may contribute to insert this species into regional agribusiness, for exploitation of the fruits in this biome, standing out as important alternative for income generation in several Brazilian regions (Pereira, 2001; Alberto et al., 2011). Murici is a fruit species that has gradually aroused interest in its functional properties, nutritional value and sensory attractions (Oliveira et al., 2008). Currently, fruit growing is one of the most important agribusiness segments, since it has been expanding from more traditional agricultural regions in recent years to the north, northeast and midwest regions, where climatic conditions are more favorable than in other parts of Brazil (Costa et al., 2013). In general, murici fruits vary a lot, as they are present in all regions of the legal amazon, cerrado, fields and coastal forests, and can be found in countries near Brazilian, such as Central America and the (Brasil, 2015). In Brazilian, the family Malpighiaceae has 38 genera and about 300 species recorded (Souza and Lorenzi, 2005). Muricis are distinguished mainly by the color of their fruits and growing region, and go by the following names: murici-amarelo, murici-branco, murici-vermelho, murici-de-flor-branca, murici-de-flor-vermelha, murici-de-chapada, murici-da-mata, murici-da-serra, murici-das-capoeiras, murici-do-campo, murici-do-brejo, murici-da-praia, among others (Costa et al., 2013). The flowering period of this species starts in late August, while fruiting starts in late September and ends in mid-January but may persist until March/April. Fruits can be consumed in their natural state or as pulp, sweets, ice creams, liqueurs, jam, popsicles, alcoholic beverages, and cereal bars (Guimarães and Silva, 2008). Byrsonima crassifolia fruits are a source of C with significant amounts of and fibers. One hundred grams of dry fruit weight can contain nutritional concentrations of 73 kcal, 1.2g total , 3mg sodium, 244g potassium, 17g carbohydrates,

Genetics and Molecular Research 19 (2): gmr18620 ©FUNPEC-RP www.funpecrp.com.br

Variation of fruits and seeds of Byrsonima crassifolia 3

8g , 8g , 0.7g protein, 74 IU , 92.5mg , 46mg calcium, 0.4mg iron, and 20mg magnesium (GreenMel, 2017). In general, vitamin sources contribute to proper vascular function, reducing cholesterol and consequent heart problems. They also help in the proper functioning of the brain, boost the body immune system, reduce the probability of appearance of cancer, lower insulin peaks, prevent anemia, among other benefits (GreenMel, 2017). The food market, especially that of organic products and foods consumed in their natural state, has gained prominence in recent years. This agribusiness component presents great potential for success, as it involves a production process that is free of agrochemicals, in addition to the practical aspects these products offer in the preparation of juices, , and pulp (Gadelha et al., 2009). The biometric characterization of fruits is an important strategy for detecting genetic variability within populations of the same species and also for drawing inferences about the relationships between this variability and environmental factors (Macedo et al., 2009), which are important aspects for the implementation of a breeding program for an edible fruit-producing species. We evaluated the genetic variability of genotypes of B. crassifolia, quantifying the relative contribution of 12 morphological characteristics of fruits and seeds.

MATERIAL AND METHODS

Fruits were harvested from 20 genotypes of B. crassifolia, cultivated in urban yards, in the municipality of Alta Floresta-MT, Brazil. being the points: (Figure 1).

Figure 1. Byrsonima crassifolia fruit collection sites of the 20 genotypes sampled in the urban perimeter of Alta Floresta – MT, Brazil. Genotypes 1 to 4 – point 1, genotypes 5 to 19 – point 2 and genotype 20 point 3.

Genetics and Molecular Research 19 (2): gmr18620 ©FUNPEC-RP www.funpecrp.com.br

J.F.L. Santos et al. 4

According to (Inmet, 2017), the climate of the region is an AWI type with a marked dry season and temperatures varying between 29 and 35ºC, with an annual mean of 32ºC. The region is situated at 292 m above sea level, and annual precipitation ranges from 1,600 to 1,800 mm. For the study, ripe fallen fruits were harvested in the crown projection area of the 20 sampled trees. Collection took place during the fruiting period, from December 2017 to January 2018. Random trees were chosen in the collection area of the urban perimeter from which 20 visually healthy, intact, and non-deformed fruits of each genotype (tree) were selected. Fruits were identified and packed in plastic bags during collection, and 400 fruits (20 fruits per genotype) were harvested and taken to the laboratory. A digital caliper was used to measure the morphometric traits fruit length (FL), width (FW), and thickness (FT). After the size was measured, the weights of fruit (FWE), seeds and pulp were determined using an analytical precision balance. Next, the fruit volume was obtained based on the volume of water displaced after it was immersed in a beaker containing water, following the methodology proposed by Basso (1999). After the seeds were removed, the pulp was manually extracted, and then seed length, width, and pulp thickness were measured. All of these measurements were taken with a digital caliper. For the assessment of the sugar content, based on the degrees brix (DB ou Brix), 4 mL of distilled water were added to the pulp of each fruit, using a syringe, to better visualize the Brix with a refractometer. Multivariate analyses were undertaken through principal components and cluster analyses, according to UPGMA (Unweighted Pair Group Method with Arithmetic Mean) and Tocher’s hierarchical methods, using the standardized mean Euclidean distance as a dissimilarity measure. The relative importance of the traits for phenotypic divergence was estimated by the method proposed by Singh (1981). R software version 3.2.4 (R Core Team, 2016) was used in the analysis of correlations among traits.

RESULTS

Four distinct groups were formed. Group 1 included accessions 4, 3, 15, 9, 8, 14, 1, 19, 11, 16, 10, 7, 5, and 6; Group II contained only accession 18; Group 3 had accessions 12 and 17. Lastly, Group 4 allocated accessions 20, 2, and 13 (Figure 2). The number of groups formed indicated the existence of variability among the accessions, which was confirmed by the UPGMA clustering method, considering a cut-off point as 2.7% distance.

Figure 2. Dendrogram representing the phenotypic dissimilarity among 20 Byrsonima crassifolia genotypes, obtained by the UPGMA method, based on the mean Euclidean distance, estimated from 12 fruit traits. The cophenetic correlation coefficient (CCC) was 0.77.

Genetics and Molecular Research 19 (2): gmr18620 ©FUNPEC-RP www.funpecrp.com.br

Variation of fruits and seeds of Byrsonima crassifolia 5

Overall, results indicated a trend of accessions from the same collection site being placed in different groups, suggesting variation within each collection site. Ganga et al. (2010) found that variations across populations can be caused by environmental factors such as type of soil, rainfall regime, temperatures, and altitudes. The cophenetic correlation coefficient, which was applied to the clustering method by the t test, showed a significant value for the method (CCC = 0.77) (P < 0.01), suggesting reliability in the relationship between the original dissimilarity means and the dendrogram generated by the UPGMA method. This fit is considered well-accepted, since it allows for inferences to be drawn through a visual assessment of the dendrogram. The closer it is to unity, the better will be the representation of dissimilarity in the form of a dendrogram (Monteiro et al., 2010). Cluster analysis of the genotypes via Tocher’s optimization method based on the mean Euclidian distance resulted in four groups formed. Group I contained genotypes 5, 7, 10, 4, 3, 8, 15, 9, 1, 6, 14, 19, 17, 11, and 13; Group II had genotypes 16, 20, and 2; Group III was composed only of genotype 18; and, lastly, Group IV was only represented by genotype 12 (Table 1).

Table 1. Clustering by Tocher’s optimization method based on the mean Euclidean distance estimated from 12 characterístics in fruits from 20 Byrsonima crassifolia genotypes.

Group Genotype I 05. 7. 10. 4. 3. 8. 15. 9. 1. 6. 14. 19. 17. 11. 13 II 16. 20. 2 III 18 IV 12

In Tocher’s clustering, it is common for a larger number of genotypes to be concentrated in the first groups. This type of analysis is aimed at maintaining homogeneity within the groups and heterogeneity across groups. A higher number of individuals being allocated to a given group indicate that they have greater genetic similarity, while individuals placed in the last groups are more divergent in relation to those in the first group (Cruz et al., 2012). The correlation analysis referring to the characteristics evaluated and presented in the heatmap below, showed the orange colors refer to values close to zero, as those presented in the variables: degree brix and length of the fruit, degree brix and width of the seed, degree brix and mass of the seed, among others. The colors that approached of red indicated values of negative correlations, such as almost all the variables that correlated with the degree brix and finally, white colors that determined positive and highly significant values, taking as an example: mass of the pulp and thickness of the fruit, fruit mass and fruit thickness, mass of the fruit and volume of the fruit (Figure 3). Knowing the correlation between variables helps in the selection process, as it makes it possible to define the interference from selection performance on one trait in relation to another as well as in the selection of traits that are difficult to measure (Gonçalves et al., 2013). Principal component analysis (PCA) revealed that the first three components (PCA1, PCA2, and PCA3) explained 87.51% of the total variation. The first component

Genetics and Molecular Research 19 (2): gmr18620 ©FUNPEC-RP www.funpecrp.com.br

J.F.L. Santos et al. 6 explained 58.68%, the second explained 16.88%, and the third explained 11.95% of the total variation (Table 2).

Figure 3. Biometric correlations between 12 variables of Byrsonima crassifolia: Brix, seed mass (SM), width (SW), thickness (ST) and length (SL), pulp thickness (PT), fruit length (FL), thickness (FT), width (FW) and volume (FV), pulp mass (PM) and fruit mass (FM).

Table 2. Estimates of eigenvalues associated with the principal components analysis corresponding to 12 traits of fruits of 20 genotypes of Byrsonima crassifolia.

Principal component Eigenvalue Root (%) (%) Accumulated PCA1 7.04 58.68 058.68 PCA2 2.03 16.88 075.55 PCA3 1.43 11.95 087.51 PCA4 0.64 05.34 092.84 PCA5 0.30 02.49 095.33 PCA6 0.27 02.24 097.57 PCA7 0.13 01.07 098.64 PCA8 0.11 00.90 099.54 PCA9 0.04 00.31 099.85 PCA10 0.02 00.13 099.98 PCA11 0.00 00.02 099.99 PCA12 0.00 00.00 100

These values are satisfactory for the study of genetic diversity, since they are higher than the 80% threshold (Cruz et al., 2012). The principal components are independent of

Genetics and Molecular Research 19 (2): gmr18620 ©FUNPEC-RP www.funpecrp.com.br

Variation of fruits and seeds of Byrsonima crassifolia 7 each other and are estimated with a view to retaining as much information as possible in terms of the total variation contained in the initial data. For this reason, it makes it possible to evaluate the importance of each studied trait in the total variation, allowing for the discard of redundant (less discriminant) variables, which are correlated with other variables, due to their invariance or because they are a linear combination of other traits (Barbosa et al., 2006). Results were similar to those found by Silva et al. (2013), who evaluated the physical and physicochemical profile of Hancornia speciosa fruits and observed an accumulation of 83.25% of the total variations in the first two principal components. However, the present findings can be used in studies aimed at the selection of genotypes to compose programs for the pre-breeding and conservation of species and at the identification of contrasting genotypes for promising crosses. The scatter plot of the genotypes, developed based on the first three principal components using the three-dimensional model (Figure 4), resulted in four groups formed. Group I contained genotypes 5, 7, 10, 4, 3, 8, 15, 9, 1, 6, 14, 19, 17, 11, and 13; Group II included genotypes 16, 20, and 2; Group III had only genotype 18 in it; and Group IV only contained genotype 12. Some groups included the same genotypes in relation to the Tocher and UPGMA method.

Figure 4. Scatter plot of the 20 Byrsonima crassifolia genotypes based on the three principal components.

As in Tocher’s method and in relation to UPGMA, the genotypes were ordered in similar groups, with the most divergent genotypes placed in distinct groups, which we can perceive in group III with individual 18, confirming the phenotypic divergence among the in this population. The relative contribution of each variable to the expression of divergence, according to Singh’s method (1981), indicated degrees brix as the variable that most contributed (36.68%) to the differentiation of the genotypes. Together, the variables fruit length (11.59%), fruit width (12.51%), and fruit thickness (13.58%) accounted for 37.68% of the

Genetics and Molecular Research 19 (2): gmr18620 ©FUNPEC-RP www.funpecrp.com.br

J.F.L. Santos et al. 8

total contribution. Seed length, seed width, seed thickness, fruit volume, pulp thickness, fruit weight, and pulp weight, in turn, contributed with values close to each other for the distinction of genotypes. However, seed weight contributed with less than 1% (Table 3).

Table 3. Estimates of relative contribution of each variable (Sj) to divergence across Byrsonima crassifolia genotypes via the method of Singh (1981).

Variable Sj Value (%) Fruit length 173.96 11.59 Fruit width 187.65 12.51 Fruit thickness 203.71 13.58 Seed length 109.01 07.27 Seed width 061.33 04.09 Seed thickness 075.08 05.00 Fruit volume 052.39 03.49 Degree brix 550.29 36.68 Pulp thickness 036.89 02.46 Fruit weight 027.41 01.83 Seed weight 000.64 00.04 Pulp weight 021.88 01.46

These results demonstrate the importance of fruit length, width, and thickness, as well as of the sugar content (degrees brix), in the discrimination of B. crassifolia genotypes, indicating that these traits should not be discarded in future evaluations. Such analyses and evaluations are similar to those of studies of Hancornia speciosa accessions (Silva et al., 2013).

CONCLUSIONS

Analysis of the fruits and seeds characterization of B. crassifolia resulted in four distinct groups, using the UPGMA and Tocher methods, based on the average Euclidean distance. The brix degree variable was negatively correlated with the mass and width of seeds, fruit and seed volume, and pulp and fruit thickness. The first three principal components explained 87.51% of the total variation. Degrees brix was the trait that most contributed to the phenotypic divergence across the genotypes.

ACKNOWLEDGMENTS

We thank the Municipal Department of Education, the Postgraduate Program in Genetics and Plant Breeding for their support. We also thank the Laboratory of Genetics and Molecular Biology (GenBioMol), the State University of Mato Grosso (UNEMAT), the Herbarium of the Southern Amazon (Herbam) and the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) - Financing Code 001.

CONFLICTS OF INTEREST

The authors declare no conflict of interest.

Genetics and Molecular Research 19 (2): gmr18620 ©FUNPEC-RP www.funpecrp.com.br

Variation of fruits and seeds of Byrsonima crassifolia 9

REFERENCES

Barbosa L, Lopes OS, Regazzi AJ, Guimarães SEF, et al. (2006). Avaliação de características de qualidade da carne de suínos por meio dos componentes principais. Rev. Bras. Zootec. 35: 1639-1645. Brasil (2015). Ministério da Saúde. Alimentos regionais brasileiros. 2ª Edição. Brasília, Distrito Federal. Costa DO, Cardoso GR and Silva GMV (2013). A evolução do setor produtivo e a comercialização de polpa de fruta no Brejo Paraibano: Estudo de caso na coaprodes. Encontro nacional de engenharia de produção. Anais eletrônicos. 33, Salvador. BA. Available at [http://www.abepro.org.br/biblioteca/enegep2013_tn_stp_177_007_22751.pdf]. Accessed May 02, 2020. Cruz CD, Regazzi AJ and Carneiro PCS (2012). Modelo biométrico aplicado ao melhoramento genético. 5ª edn. Universidade Federal de Visçosa, Minas Gerais. Cruz CD (2013). Programa Genes: a software package for analysis in experimental statistics /and quantitative genetics. Acta Sci., Agron. 35: 271-276. Gadelha AJF, Rocha CO and Vieira FF (2009). Avaliação de Parâmetros de Qualidade Físico-Químicos de Polpa Congeladas de Abacaxi, Acerola, Cajá e Caju. Rev. Caatinga. 22: 115-118. Ganga RMD, Ferreira GA, Chaves LJ, Naves RV, et al. (2010). Caracterização de frutos e árvores de populações naturais de Hancornia speciosa Gomes do Cerrado. Rev. Bras. Frutic. 32: 101-113. Gonçalves LGV, Andrade FR, Marimon Junior BH, Schossler TR, et al. (2013). Biometria de frutos e sementes de mangaba (Hancornia speciosa Gomes) em vegetação natural na região leste de Mato Grosso, Brasil. Rev. Cienc. Agrar. 36: 31-40. GreenMel (2017). Farei bem à Terra. Murici - todos os benefícios para sua saúde e muito mais. Available at [https://www.greenme.com.br/usos-beneficios/6064-murici-beneficios-saude-muito-mais]. Accessed May 1, 2020. Guimarães MM and Silva MS (2008). Valor nutricional e características físicas e químicas dos frutos de murici-passa, (Byrsonima verbascifolia). Ciênc. Tecnol. Aliment. 28: 817-821. Inmet (2009). Instituto Nacional de Meteorologia. Available at [http://www.inmet.gov.br/portal]. Accessed April 26, 2020. Monteiro ER, Bastos EM, Lopes ACA, Gomes RLF, et al. (2010). Diversidade genética entre acessos de espécies cultivadas de pimentas. Ciênc. Rural. 40: 288-293. Alberto PS, Silva FG, Cabral JSR, Fátima Sales J, et al. (2011). Methods to overcome of the dormancy in murici (Byrsonima verbascifolia Rich) seeds. Semin. Ciênc. Agrár. 32: 1015-1020. Oliveira KAM, Ribeiro LS, Oliveira GV, Mendonça RCS, et al. (2008). Desenvolvimento de formulação de iogurte de araticum e estudo da aceitação sensorial. Aliment. Nutr. 19: 277-281. Pereira JOP (2001). O papel de Abelhas do gênero Centris na Polinização e Sucesso Reprodutivo do Muricizeiro (Byrsonima crassifolia). Masters thesis. Universidade Federal do Ceará, Fortaleza. 59f. RCore TEAM (2016). R: a language and environment for statistical computing. R Foundation for Statistical Computing. Vienna, Austria. Silva SS, Cruz EMO, Reis RV, Ferreira CP, et al. (2013). Caracterização morfológica e molecular de genótipos de mangaba. Rev. Bras. Frutic. 35: 1093-1100. Souza VC and Lorenzi MH (2005). Botânica sistemática: guia ilustrado para identificação das famílias de Angiospermas da flora brasileira (baseado em APG II). Instituto Plantarum de Estudos da Flora. Nova Odessa, São Paulo.

Genetics and Molecular Research 19 (2): gmr18620 ©FUNPEC-RP www.funpecrp.com.br