Assessing the genetic structure of teak from Southeast Sulawesi and its implication for genetic conservation and utilization in Indonesia

Abstract Genetic relationships among teak (Tectona grandis) seed sources have been found to be low, thus genetic materials from other sources are required to maintain broad genetic diversity. This study here is therefore aimed to assess the potency of teak genetic structure in Southeast Sulawesi, Indonesia. Leaf materials were sampled from six populations: the villages of Angondara, Anduna (Konawe), Napabalano, Matakidi, Wakuru (Muna) and Wakonti (Buton). One population from Java Island (Kepek) was used as an outgroup. Genetic structures were assessed by using six microsatellite markers. The genetic diversity within populations was moderate (mean expected heterozygosity = 0.544; mean allelic richness = 3.752) as well as the genetic differentiation among the populations (mean F ST = 0.085). Dendrogram analysis revealed that the populations were separated into two clusters; the first is Kepek-Angondara, Anduna-Napabalano-Matakidi, Wakuru, and the second is Wakonti. AMOVA showed that the genetic variation was insignificant between regions, but significant among populations. The structural analysis demonstrates the division of populations into two lineages (Java and Southeast Sulawesi populations). Differences in genetic structures indicated that the teak from Southeast Sulawesi may have originated from other sources. Thus, those populations are promising for broadening the genetic base of commercial teak in Java.


Introduction
Teak (Tectona grandis Linn. f) is a valuable and durable species for timber with excel-lent wood quality, color, and texture. The price of teak wood is largely dependent on its quality, which is mainly determined not only by its esthetic value but also their growth. Teak wood has very high price worldwide, with the value of round log of teak reaching up to USD 250-321 per m 3 (Wanneng et al. 2021). Originally, teak is native to India, Laos, Myanmar, and Thailand (FAO 2012), covering an area of approximately 29,035 million hectares in natural forests. It is also widely planted in other countries, including Indonesia (Verhaegen et al. 2010;Fofana et al. 2013). Teak was introduced to Java Island, Indonesia between the period of 14th and 16th centuries, wherein it was widely distributed as plantations due to Hinduism and Buddhism practices for missionary purposes. Since its early introduction, the seeds and seedlings of this exotic teak represent the genetic basis of teak plantations in Indonesia leading to the formations of various landraces after decades (Wibowo 2014). There is also information indicating that Indonesian teak was originally introduced from the natural forests of East Thailand or Central Laos. Nevertheless, additional seed sources from South India were also introduced recently into Java to increase the diversity of the gene pool for provenance trials and conservation areas (Fofana et al. 2013).
Teak domestication and improvement has been initiated since the 1800s, within the Dutch colonial era (Wibowo 2014). More intensive forest tree-breeding program was initiated in the 1930s (Kollert and Cherubini 2012), with provenance tests were established in 1932 by Forestry Institution (Bosbouw Proefstation) and grown in 4 Forest Management Units of Ngawi, Randublatung, Bojonegoro dan Blitar regions (Poerwokoesoemo 1965;Susanto 2020). After a stagnant period, because of the governance changes from the Dutch, Indonesia government initiated its tree improvement activity, in which the teak seed orchards were finally established in the 1980s (Kollert and Cherubini 2012). Since 1981, further teak tree improvement was planned by Perum Perhutani which ending up with a collaboration with Faculty of Forestry, Gadjah Mada University in 1988 (Wirabuana et al. 2022). Combination trials of progeny and provenance trials were then established in 2 locations, Jember (East Java) and Wanagama (Yogyakarta) (Wirjodarmodjo and Subroto 1983;Wibowo 2014). The teak plantations, at present cover the areas of 1.3 million hectares (7% the total land area of Java), and represent the second-largest area of teak plantations worldwide after India (Kollert and Cherubini 2012).
Globally, the recent teak growing under natural forests has declined by 385.000 hectares (FAO 2012). Therefore, the patterns of its genetic diversity and structure are becoming more important to consider for many purposes. This includes in establishing and managing seed sources, translocation or deploying seeds, implementing conservation and breeding strategies as well as in controlling the existing gene pool. Similarly in Indonesia, even the teak protected forests in Java experienced a significant decline within the 5-year period of 2015 to 2020, from 155.009 ha into 152.951 ha (Perhutani 2015(Perhutani , 2020. There have been previous studies assessing the genetic diversity and structure of teak, both in natural and exotic plantations. The genetic diversity in natural populations is reportedly moderate (Fofana et al. 2009;Verhaegen et al. 2010;Win et al. 2015), while its genetic differentiations identified among natural forests in India, Laos, Myanmar, and Thailand has the value of F ST ¼ 0.22 (Fofana et al. 2009;Verhaegen et al. 2010). Recent report informing that teak was distributed to Java (Indonesia) from North Thailand or Central Laos, and from North India (Verhaegen et al. 2010). This has been confirmed with the simple sequence repeat (SSR) analysis which revealed that Java teak was similar to that of Myanmar (Win et al. 2015). Further, by using random amplified polymorphic DNA (RAPD) markers, Indonesian teak was also found to be close to that of Thailand, India, and Myanmar (Widyatmoko et al. 2013).
Beside in Java, teak is also widely distributed outside Java, especially in Sulawesi. Its distribution is mainly in Southeast part of Sulawesi (Konawe and Muna districts, and Buton Island). Whether they are as natural forests or landraces is still under debate (Widyatmoko et al. 2013). Konawe is located in the mainland of Southeast Sulawesi, while Muna and Buton islands represent a series of disjunctive islands isolated by straits in the Banda Sea. Teak in Sulawesi does not apply adequate silvicultural management, rotation and harvesting such as in Java. However, due to its cheaper price compared to Java, Sulawesi teak is marketed there. In southeast Sulawesi, non-teak log production reached up to 74% whereas teak barely reached 26% (BPS 2013). A number of viewpoints contend that Java teak is superior to Sulawesi teak in terms of quality. Teak itself is a type of timber tree that needs a lot of phosphate and calcium to grow. Additionally, teak needs lots of sunlight. The coastal regions of Java have an abundance of these soil types, which produce high-quality teak. Blora Regency-Central Java produces even the best teak. There are traits that set Sulawesi teak apart from Javanese teak, including sapwood, or the heartwood's outermost layer, which is thicker in color in Sulawesi teak than Javanese teak. Sulawesi teak needs to be fully dried before processing since it shrinks more quickly than Javanese teak, according to certain furniture artisans. Sulawesi teak is therefore unable to compete in trade compared to teak in Java. Because of this, teak business in Sulawesi is not as good as in Java (Anonim 2018).
Meanwhile in Muna, where teak is preferred due to its best quality (Lempang and Asdar 2007), the logging of teak wood for entrepreneurs holding (IPTKM) is mostly carried out in state forest areas which is strictly prohibited. The weakness of local regulations encourages the exploitation of forest resources in state forest areas, as well as in community plantation forest areas (Faqih 2018). So that Jati Muna's popularity is said to be gone which also causes a decrease in water storage in the area (Rusdianto 2006).
However, so far, the patterns of distribution and the origins of the teak of forests within those areas have never been assessed. This study is therefore projected to assess the genetic diversity and structure of teak by using microsatellite markers, specifically from Southeast Sulawesi (Konawe, Muna, and Buton) with Java as an additional accession. The findings might reveal the genetic information of the unique Southeast Sulawesi teak to support strategies of the Indonesia and global genetic conservation and tree improvement, for its sustainable utilization in the future.

Sample collection
Leaf and cambium samples of six landraces were randomly collected from teak populations in Southeast Sulawesi Province (Angondara and Anduna at Konawe district; Napabalano, Matakidi, and Wakuru at Muna district; and Wakonti at Buton island). A teak plantation in Java Island from Kepek, Yogyakarta Province, was also included as an outgroup ( Figure 1 and Table  1). The reason to choose Kepek to represent Java is because this area of plantation is a forest community plantation which its seeds are from many unknown sources. However, within the area, there are teak trees which are estimated to be hundreds of years old representing the initial condition of gene structure. Those trees become teak seeds sources for planting over large area within the Java Island.

Dna extraction and SSR analysis
Leaf and cambium samples were collected from individuals (Table 1), dried using silica gel and stored at room temperature (25 C) up to the process of DNA extraction. Total genomic DNA was extracted by using CTAB modification (Shiraishi and Watanabe 1995). Six (6) SSR markers (CI-RAD1TeakH10, CIRAD3TeakA11, CIRAD3TeakF01, CIRAD4TeakH9, CIRAD4TeakF02, and CIRAD4TeakDa12) which have been used for teak (Verhaegen et al. 2005) were used for screening. Pol-ymerase chain reaction (PCR) was performed into 10 mL samples containing 1-10 ng genomic DNA, master mix, and 0.3 mM for each primer. Samples were amplified in a GeneAmp PCR System 9700 (Applied Biosystem, USA) under the following conditions: denaturation at 94 C for 5 min, followed by 30 cycles at 94 C for 30 s, annealing for 30 cycles at 94 C for 30 s, annealing at 56 C and then 72 C for 30 s, and the final extension at 72 C for 5 min. The amplified PCR products were loaded on an ABI 3100 Avant Genetic Analyzer (Applied Biosystems, USA), and their sizes were determined by using Gene-Mapper analysis software (Applied Biosystems).

Data analysis
Genetic diversities within teak populations were evaluated for number of detected alleles (N A ), allelic richness (A R [16] ) and private allelic richness (P A [16] , Nei unbiased expected heterozygosity (H E ), and fixation index (F IS ). The N A , H E and F IS were calculated by using FSTAT software version 2.9.3.2 (Goudet 2001 [16] were calculated by using rare faction method at a locus for fix samples size (Kalinowski 2005). We used minimal samples size for all population i.e. 16 genes (8 individual diploid trees) and were calculated using HP-Rare 1.1. software (Takezaki 2001).
The genetic differentiation (F ST ) among populations was calculated by using FSTAT (Goudet 2001). The genetic relationships among the populations were evaluated by using a neighbor-joining tree based on the D A genetic distance and estimated by using POPTREE2 software (Takezaki 2001). The significance was estimated with 1000 bootstrap replicates, and for constructing consensus neighbor-joining trees. Genetic variation was evaluated hierarchically through analysis of molecular variance (AMOVA) and performed PCoA (principal coordinate analysis) by using GenAlEx 6.5 software (Peakall and Smouse 2012). In addition, using the STRUCTURE program, we carried out Bayesian cluster analysis to gain understanding of the genetic composition of Southeast Sulawesi populations (version 2.3; Pritchard et al. 2000). Ten replications were performed, each with K ranging from 1 to 10, a burning length of 100,000, and 100,000 Markov Chain Monte Carlo (MCMC) iterations. Using the software STRUCTURE HARVESTER (Earl and VonHoldt 2012), we plotted the log probability (L(K)) and K the method of Evanno et al. (2005) to identify the most likely number of K. The outputs from STRUCTURE and STRUCTURE   (Piry et al. 1999). It was conducted under the infinite allele mutation model (IAM) assumption, which is the only precise model specifically used for few numbers of SSR markers.

Genetic diversity within populations subsection
The genetic diversity parameters within populations are shown in

Bottleneck detection
Under the IAM assumptions, almost all populations show significant mutation-drift equilibrium with excess heterozygosity and significant bottlenecks, according to the Wilcoxon signed-rank test (Piry et al. 1999; Table 2).

Population differentiation and genetic relationship among populations
The mean F ST value of the seven populations was 0.085 with confident intervals of 0.040-0.123 (95%) and 0.032-0.135 (99%), indicating that the population genetic differentiation is relatively low. Additionally, most pairwise F ST values are significant (p < 0.05). The neighbor-joining tree presented in Figure 2 identifies two large population clusters: the first is in Java, Konawe (Angondara and Aduna), and Muna (Napabalano, Matakidi and Wakuru), and the second is in Buton (Wakonti). These two clusters showed high bootstrap values, while the sub-cluster with high bootstraps (>50%), are sub-cluster Java and Angondara (Konawe), sub-cluster Konawe and most Muna populations, sub-cluster Wakuru, and sub-cluster Wakonti. Clustering is likely related to the geographic positions. The PCoA result demonstrated a distinct difference between the Angondara-Kepek populations and the other Southeast Sulawesi populations (Figure 3). Although the results of hierarchical AMOVA (Table 3) show that the genetic variant partitioning between the two regions of Wakonti and the other populations) is insignificant, it is significant among populations and individuals within the population (12% and 88%, respectively). Thereby it is indicating significant genetic variation among populations.
In the structural analysis, the delta K shows the highest values at DK ¼ 4 ( Figure 4). Further, the six populations and the outgroup population are visualized  into four genetic clusters ( Figure 5). In DK ¼ 4, they show that Kepek from Java and Angondara differ from the other populations from Southeast Sulawesi. Moreover, the genetic composition of Kepek is similar to that of Angondara. Meanwhile, the genetic compositions of the remaining five populations from the Southeast Sulawesi site (Anduna, Napabalano, Matakidi, Wakuru, and Wakonti) are similar.

Discussion
Analyzed data were carefully read, and it was discovered that the percentage polymorphic of the SSR markers employed in this study ranged from 83% to 100% with an average 93%. Other SSR markers on teak have information content (PIC) from 0.81 to 0.92 with an average of 0.88 (Mohammad et al. 2022). SSR markers were effective and acceptable for evaluating genetic diversity in teak, as evidenced by high estimations of PIC value, a genetic marker's informativeness metric. The results show that the high and significant coefficient of inbreeding value (F IS ) found in the Angondara and Matakidi populations indicate inbreeding. This might be caused by the characteristics of either the primers or the populations. However, the primer aspect may not greatly influence the results because those SSR have been tested by prior study (Verhaegen et al. 2005;Fofana et al. 2009). In term of population, optimum pollination should be maximized when seeds from those two populations will be collected.
The SSR markers in this study are used because those markers have been evaluated previously (Verhaegen et al. 2005;Fofana et al. 2009) and reported with confirmation that all their loci constitute useful tools for population genetic studies. By using those markers, we induce from the results of this study that the genetic diversity parameters found for Southeast Sulawesi teak, described in parameter values, are still high. These are similar to those of teak from natural forests of India, Thailand, and Myanmar, as well as to plantations in Africa and Java (Indonesia) (Verhaegen et al. 2010;Win et al. 2015).    . A visualization of the first two primary coordinates from the PCO analysis using DA distances for the seven populations (Nei et al. 1983). The populations of Southeast Sulawesi grouped into one cluster, while those of the Kepek and Angondara formed another cluster.
High genetic diversity for teak is expected, because teak is mainly an insect-pollinated and outcrossing species (Palupi and Owens 1998). An insect might carry a large number of pollen grains at least 500 m away from the pollen sources (Pu et al. 2014;Nurtjahjaningsih et al. 2020). The study of outcrossing rate estimation for teak by using alloenzym demonstrate the value of 0.89 for single-locus and 0.95 for multilocus; which describing that teak is predominantly outcrossing (Kjaer and Suangtho 1995;Hansen et al. 2015). With high genetic diversity shown in this study, the teak of Southeast Sulawesi might provide precious sources of more complete genetics. This should be advantageous for infusion into populations of plantations in Java that have been intensively exploited for decades. Long and intense teak production as well as the use of clones in Java (Prasetyo et al. 2020), might cause the genetic structure deleterious. Moreover, according to genetic evidence, teak was likely introduced to Indonesia from a restricted region with a limited genetic base in Central Laos or eastern Thailand (Verhaegen et al. 2010). Hansen et al.'s (2017) more recent genomic analysis similarly revealed that Indonesian landraces primarily come from a certain region of the native distribution. This is because, it has been well known that a species with long term utilization, especially very highly demanded commercial timber and domestication, might encounter significant loss of its many essential unique alleles (Porth and El-Kassaby 2014). Meanwhile, those specific alleles should be fundamentally useful and required for maintaining its diversity in the future (Tsuda and Ide 2005;Hadiyan et al. 2022).
In previous study using the RAPD analysis, it shows that the genetic diversity of Indonesian teak (both Java and Southeast Sulawesi) was relatively low (Widyatmoko et al. 2013), which in contrast with the result of SSR used in this study which demonstrating high value. This might because the SSR is considered to be more precise compared to RAPD due to higher SSR stability character and polymorphism (Ahmed et al. 2012). However, the dissimilarities in the genetic diversity of both methods might also be complemented by the differences in population and in DNA regions.
Both the H E and A R[16] values are the lowest in Anduna (Konawe) and Napabalano (Muna). This might be because Anduna comprises of stands planted by local individuals with very limited seed sources and very few trees. In addition, the teak in Anduna is also influenced by human intervention in which common people may access those precious trees with minimum control. Meanwhile, the teak area in Napabalano (Muna) has been becoming the oldest teak conservation area in Southeast Sulawesi. It is located in the regions with high tree density and canopy closure which cause low humidity. These conditions affect teak with lack of synchronized flowering and pollinator foraging, causing high rates of self-pollination and endosperm abortion. Insect biodiversity has been known to influence the success of pollination in teak (Gunaga and Vasudeva 2003).
While in turn, the abundance of insects is greatly influenced by sunlight (Nurtjahjaningsih et al. 2020). In addition, the Napabalano population is isolated, and it has no other neighborhood stands. Thus, no interactions with the surrounding populations, which particularly after centuries, might create genetic and geographic barriers. This leads to the low geneflow exchange which causes low genetic diversity (Schierenbeck 2017).
Consequently, the Napabalano population shows the evidence of declining in effective population size, due to low gene interchange among populations. This is clearly supported by the bottleneck analysis of this study under IAM assumptions. Fluctuations in effective population size were due to many factors including tree competition and adaptation to many environmental factors (Tsuda and Ide 2005).
Moderate or relatively high values of genetic differentiation have been reported in tree species with restricted and fragmented distributions (Tsuda and Ide 2005). However, for teak in this study, the genetic differentiation among the seven teak populations is relatively low (mean F ST ¼ 0.085). This little value of difference is most likely caused more by the limited origins of the teak seed sources (Prasetyo et al. 2020) rather than geographical barriers among populations used in this study. This confirms previous studies of teak in Africa and India which also indicated low level of F ST values with F ST ¼ 0.08 and F ST ¼ 0.03 respectively (Verhaegen et al. 2010). Meanwhile, in India germplasms, due to teak's predominately cross pollination mode, the pairwise population F ST values were likewise low (F ST ¼ 0.023) but significant, indicating adequate gene movement among locations (Mohammad et al. 2022). The genetic trend of teak in Indonesia performs similarly to the natural dispersion from India, which exhibits very low differentiation and high genetic variability. By treating it as a single germplasm source and disregarding the location of the genotypes within India seed origin, it exhibits high polymorphism because of significant genetic variance among genotypes (Vaishnav and Ansari 2019). Combining seed sources from Southeast Sulawesi in this approach may also be useful for preserving Indonesia's genetic diversity.
The neighbor-joining tree analysis based on D A distances revealed a geographical genetic structure, in which the populations are divided into two main groups (Java-Konawe-Muna and Buton). It was also supported by PCoA results. Interestingly, it is very noticeable that the Angondara (Konawe) population was genetically close to the Kepek population (Java) (Figure 2). The analysis also showed that other Southeast Sulawesi populations, except Buton, were genetically close to each other. However, based on the results of the clustering analysis of AMOVA, it shows that the genetic variation between Buton population and other Southeast Sulawesi populations was insignificant. The genetic variation which was significant among the other populations (12%), indicating a strong genetic structure between the Java-Angondara populations and the other Southeast Sulawesi populations.
Further, structure analysis in this study also supports the AMOVA analysis, wherein Kepek (Java) population and only one population of South Sulawesi (Angondara), showed a similar genetic structure ( Figure 5). While with the other Southeast Sulawesi populations (population 3-7), differences in genetic structures with Java occur obviously. There is very interesting information which approves this similarity, in which the Angondara population was actually established in 1997-1998, by using genetic materials mainly from Java sources (Mahfudz 2021).
Meanwhile, genetic structure between Java and other Southeast Sulawesi populations shows significant differences. This has also been reported previously by using the RAPD markers (Widyatmoko et al. 2013). It reveals that the Javanese teak was found to cluster with the teak populations of Thailand, India, and Myanmar, while the Southeast Sulawesi teak was separately grouped into different clusters. Remarkably, both results of those genetic structural analyses were in concordance with the historical origin, in which the teak introduced to Java was from natural forests of India, Thailand, Laos, and Myanmar (Verhaegen et al. 2010). Further, the fact that genetic structures differed between the Javanese and Southeast Sulawesi teak, specify that the teak forest in Southeast Sulawesi might have been originated from other sources. There is confirmation that teak has been introduced to Southeast Sulawesi during the British colonial from Java around 1811 sh , and so it is called Kulidawa (wood from Java) (Azhar 2007).
All in all, genetic materials within Southeast Sulawesi regions are considered to be similar. Whenever, in-situ conservation will be determined, Wakonti in Buton is the best representing the genetic of Southeast Sulawesi. This is demonstrated from its high expected heterosigozity value (H E ¼ 0.586). Meanwhile for ex-situ conservation, genetic materials from Southeats Sulawesi and Java should be separated due to their genetic differences (Figure 3).
Considering the results that Southeast Sulawesi teak has different genetic structure from that of Java, specific and unique genetics might likely exist. Therefore, genetic materials from Southeast Sulawesi will be very advantageous for genetic infusion into Java populations for breeding and large-scale plantations. Comparing both regions, there has been very much differences in the way teak is commercially harvested. Heavily populated Java Island has been recorded to manage teak in such an intensive way for decades, with the total commercial production increased dramatically from 2016 to 2020 from 250.638 m 3 to 432.817 m 3 per year (Wirabuana et al. 2022). Meanwhile, the annual new plantations are established routinely and increasing from 1.617 ha in 2016 to 13.469 ha in 2020 (Perhutani 2020). This practice of intensive teak wood consumption would significantly deplete and decline the gene diversity for teak, which in decades might lead to the bottleneck effect (Tsuda and Ide 2005). This will culminate into the remaining populations with very narrow gene diversity in the future. Meanwhile, in Southeast Sulawesi, commercial exploitation of teak is relatively lower and slower compared to Java. This is partly due to common Indonesia teak trade that has been fulfilled and dominated by teak timber yielded from Java (Wirabuana et al. 2022). Less competitive and less intensive harvest from Southeast Sulawesi is also because of no specific semi-private company such as PT Perhutani in Java which has been managing teak for decades. Further, some populations in Southeast Sulawesi have been appointed to become teak conservation areas. In addition, it is also supported with a cultural influence of spreading belief, in which harvesting teak from that areas would cause death or illness, and this affects teak populations to remain untouched (Azhar 2007).
Although the results of this study confirmed several previous findings, further investigation by using organelle DNA markers would be commendable to carry out.

Conclusion
This study reveals the presence of genetic diversity and structure of teak in Southeast Sulawesi using SSR DNA markers in which the parameters are found to be similar to those of natural teak populations in Southeast Asia. The genetic structure results suggest that Java and Southeast Sulawesi comprised of teak populations of different genetic origins, whereas those among populations at Southeast Sulawesi have similarity in genetic structure. In term of breeding and large-scale plantations, genetic materials of teak from Southeast Sulawesi will be very valuable for genetic infusion into Java populations as well as for conservation. In evaluating the total genetic diversity more distinctly, a large-scale assessment of the organelle DNA of teak, especially for teak distributed in Indonesia, should be conducted for future study.

Data availability statement
Not applicable.