Genetic Diversity and Population Structure of Black Pine (Pinus nigra Arn.) in Mt. Athos, Northern Greece

Abstract

European black pine (Pinus nigra Arn. subsp. nigra) persists in scattered montane stands across Greece, where isolated populations harbour genetic variation shaped by local environments and demographic history. In this study, we assessed the genetic diversity and population structure of P. nigra using nuclear microsatellite markers (nSSRs) across four populations: Mt. Athos, Sithonia, Thassos, and Perama. A total of 67 individuals were genotyped, and seven high-quality polymorphic loci were retained after rigorous filtering. The Mt. Athos population exhibited the highest allelic richness and heterozygosity, with all loci being polymorphic and a low inbreeding coefficient after null allele correction. In contrast, the Perama population displayed reduced diversity, fewer polymorphic loci, and persistent heterozygote deficits. Principal Component Analysis (PCA) and Discriminant Analysis of Principal Components (DAPC) revealed weak overall population structure, with Perama genetically distinct from the other sites. Spatial Principal Component Analysis (sPCA) further uncovered an east–west cline within Athos and localized structure potentially shaped by both natural isolation and human influence. These findings highlight regional variation in genetic diversity within P. nigra and identify Athos as a genetically rich population of particular interest. The results provide a foundation for long-term monitoring and support informed strategies for the management and conservation of P. nigra in Greece.

1. Introduction

European black pine (Pinus nigra Arn. subsp. nigra) is one of seven native pine species in Greece and reaches its southern European limit in the Hellenic mountains, where it occupies montane belts from c. 400 m to 1500 m a.s.l. in scattered pure or mixed forests, in regions such as the Mediterranean basin, Black Sea, Southern and Central Europe, Western Asia, and parts of Northern Asia [1,2]. It can live more than 500 years, tolerates poor soils and drought, and is therefore pivotal for erosion control and restoration projects across the Mediterranean. Genetic differentiation among its five recognized subspecies reflects adaptation to contrasting continental and insular climates, yet rangewide studies show that most diversity is partitioned within, rather than among, populations because of long-distance pollen flow [3,4,5,6,7]. In Greece, black pine persists as a network of small, isolated demes on Mt. Parnon, Euboea, Lesbos, Samos, Thasos, and the Chalkidiki peninsulas [1,8,9].

Such fragmented southern-European populations are predicted to harbour unique alleles but are also exposed to genetic erosion when effective population sizes fall below viability thresholds. An interesting case of such fragments in north Greece can be found on Mt. Athos, an autonomous monastic peninsula whose ridge still bears priority habitat 9530* “(Sub-)Mediterranean pine forests with endemic black pines”. Recent field inventories describe the Athos population as naturally small, patchy and isolated, with trees occurring singly or in minute groves along steep, fire-prone slopes [10,11]. Long-term threats include genetic drift, inbreeding, stochastic storms, and a lack of silvicultural attention relative to chestnut and holm oak [6,12,13].

Several genetic studies have shown pronounced differentiation between northern and southern European populations, yet the rear-edge stands in southern Europe remain comparatively understudied [6,14]. The Genetic marker-based insights provided by recent studies [15,16] reinforce this knowledge gap and stress the need for a finer-scale assessment.

Four interrelated questions drive the present study: (i) How distinct is the Mt. Athos black-pine gene pool from other north-eastern Greek populations? (ii) Do the genetic patterns of forest patches on Mt. Athos reveal genetic uniformity and high diversity? (iii) Is its genetic diversity sufficient for long-term viability? (iv) Do the data reveal ongoing genetic erosion or signals of past bottlenecks? and (v) Which conservation and management priorities follow from the answers? Answering these questions is important as, in view of projected warming and increasing aridity in the Mediterranean Basin, these small peripheral populations may harbour genetic variants associated with enhanced tolerance to adverse environmental conditions such as heat and drought, thereby providing a robust scientific foundation for targeted conservation and restoration efforts.

2. Materials and Methods

2.1. Sample Collection and Origin

Needle samples from black pine were collected from four distinct regions in Greece: the peninsula of Athos and Sithonia (Chalkidiki), Perama (Eastern Macedonia and Thrace), and the island of Thasos. A total of 93 samples were gathered: 43 from the Athos peninsula, 20 from Psaria in Sithonia, 20 from Perama, and 10 from Maries in Thasos (Figure 1). Geographic coordinates were recorded for each sampled tree (Table S1). All samples were stored at −20 °C prior to DNA extraction. Sample nomenclature employed the initial letter of each collection region followed by a unique identifier for each individual. Only mature trees were selected, and individuals were separated by at least 50 m. Sampling was carried out extensively in Mt. Athos, which is the focal population of this study. There, individuals sampled were representative of all locations where black pines occurred. The other three populations were chosen because of their geographical proximity to Mt. Athos. In these cases, adult trees were selected randomly, again with a minimum distance of 50m from each other.

2.2. DNA Extraction and Quantification

Therefore, fresh DNA extraction was performed for all samples. Genomic DNA was extracted from 70–100 mg of needle tissue using the NucleoSpin^® Plant II kit (Macherey-Nagel, Düren, Germany) according to the manufacturer’s protocol. DNA purity and quantity were assessed using a NanoDrop™ OneC Spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). DNA integrity was further verified by electrophoresis on a 1% agarose (UltraPure Agarose, Invitrogen, Carlsbad, CA, USA) gel, pre-stained with GreenSafe Premium Nucleic Acid Gel Stain (NZYtech, Lisbon, Portugal).

2.3. Microsatellite Loci and Multiplex Design

Eleven nuclear microsatellite (nSSR) primer pairs, previously validated for related Pinus species (P. taeda, P. halepensis, P. sylvestris), were selected from Giovannelli et al. [17] for amplification fragment analysis (Table S2).

Primers were grouped into three multiplex sets based on expected PCR product sizes and potential for primer dimer formation. Forward primers in each set were labelled with distinct fluorescent dyes. Prior to empirical testing, in silico PCR analysis was performed using Primer-BLAST (NCBI) against available Pinus spp. genomic sequences (as a complete P. nigra reference genome was unavailable) to check for potential non-specific products from primer combinations within each multiplex.

2.4. PCR Amplification

PCR amplifications were performed in a VeritiPro Thermal Cycler (Applied Biosystems©, Foster City, CA, USA). Gradient PCR was initially conducted for each primer pair to determine optimal annealing temperatures for the multiplex sets. Each 25 µL reaction contained 12.5 µL of OneTaq^® 2X Master Mix with Standard Buffer (New England Biolabs, Ipswich, MA, USA), 0.2 µM of each primer, and 2 µL of genomic DNA. The PCR cycling conditions were as follows: initial denaturation at 94 °C for 30 s, followed by 35 cycles of 94 °C for 30 s (denaturation), annealing at 62 °C (M1), 56 °C (M2), or 59 °C (M3) for the three multiplex sets for 30 s, and 68 °C for 1 min (extension), with a final extension step at 68 °C for 5 min. Amplified products were visualized on 2% agarose gels stained with GreenSafe Premium (NZYtech, Lisbon, Portugal) using a 100–1000 bp DNA ladder (NZY-Ladder V, NZYtech, Lisbon, Portugal).

2.5. Data Processing

Fluorescently labelled PCR products from each multiplex reaction were subjected to fragment analysis by an external service provider (CeMIA, www.cemia.eu) using an ABI 3730 DNA analyzer (Applied Biosystems). In total, 67 samples were genotyped. Products were mixed with GeneScan™ 600 LIZ™ Size Standard (Applied Biosystems) prior to capillary electrophoresis. The analyzer detects 6-FAM, VIC, NED, PET, and LIZ dyes (Applied Biosystems G5 dye set).

Raw data were provided as FSA files. Allele calling was performed using Geneious Prime 2024.0 software. The software automatically fitted the size standard to determine fragment sizes using the Third-Order Least Squares sizing method. Peaks corresponding to microsatellite alleles were manually inspected and scored to ensure accuracy and differentiate true alleles from background noise or artifacts. Dye wavelengths used in this study were matched by Geneious Prime to its registered dyes for accurate signal interpretation.

Allele binning was performed manually in Geneious Prime, considering the repeat motif length for each locus to group observed fragment sizes into discrete allele categories (e.g., fragments ranging from 156.48–156.9 bp for primer AF333785 were binned as allele 157). Loci with more than 30% missing data were excluded from the analysis. Subsequently, individuals with successfully genotyped data for fewer than three loci were also removed to ensure data quality and reliability.

2.6. Genetic and Statistical Analysis

Genetic diversity was quantified per locus and population in GenAlEx v6.503 with the following parameters: number of alleles (Na), number of effective alleles (Ne), observed (Ho) and expected (He) heterozygosity, and the inbreeding coefficient (F). Allele frequencies and pairwise Nei’s genetic distances (standard and unbiased) were also obtained, an AMOVA based on Φ_PT estimates partitioned variance among and within populations [18].

To control for potential scoring bias, each locus was screened for null alleles with MICRO-CHECKER v2.2.3 [19] (Oosterhout algorithm, 95% CI). Where nulls were indicated, allele frequencies were adjusted under Hardy–Weinberg expectations and Ho, He, and Fis were recalculated; these corrected metrics were carried forward into all subsequent analyses.

Multivariate structure was explored in R (adegenet): an initial PCA assessed clustering without a priori groups, followed by DAPC to maximise between-group discrimination and validate PCA patterns. sPCA added the spatial dimension, mapping global and local genetic structure across the study area. All maps in this study were produced using QGIS [20].

3. Results

Out of the nine loci used in this research, two were removed since fewer than 30% of the individuals were genotyped across them. In a similar manner, 26 out of 93 were excluded, because they produced genotypes for three loci or fewer. Across the seven polymorphic nSSR loci retained, black pine in the populations studied displayed moderate levels of genetic diversity and differentiation among populations. Mean allelic richness was 13.7 ± 5.9 alleles per locus, with gene diversity (He) averaging 0.620 ± 0.057; however, observed heterozygosity was markedly lower (Ho = 0.453 ± 0.075), producing an overall inbreeding coefficient of F = 0.320 ± 0.106 (

Table 1. Summary statistics (mean and standard error) over populations for each locus, including number of different alleles (n), observed heterozygosity (Ho), adjusted observed heterozygosity (Ho_adj), expected heterozygosity (He), adjusted expected heterozygosity (He_adj), fixation index (F), and adjusted fixation index (F_adj).

Locus	n	Ho	Ho_adj	He	He_adj	F	F_adj
KU953385	15.25 ± 6.945	0.176 ± 0.104	0.524	0.717 ± 0.062	0.807	0.785 ± 0.127	0.383
KU953379	13 ± 6.868	0.347 ± 0.13	0.492	0.513 ± 0.171	0.576	0.322 ± 0.109	-
AF333785	13.25 ± 5.721	0.064 ± 0.04	0.139	0.189 ± 0.063	0.226	0.669 ± 0.156	-
AF286619	14.75 ± 6.142	0.705 ± 0.066	0.705	0.717 ± 0.014	0.717	0.017 ± 0.089	0.017
KU953384	13.25 ± 6.047	0.456 ± 0.041	0.587	0.626 ± 0.047	0.677	0.264 ± 0.074	0.141
KU953382	14.25 ± 5.949	0.643 ± 0.079	0.752	0.829 ± 0.012	0.850	0.224 ± 0.096	0.116
KR779884	12.25 ± 4.008	0.778 ± 0.063	0.778	0.749 ± 0.033	0.749	−0.043 ± 0.092	−0.042
Mean	13.714 ± 5.954	0.453 ± 0.075	0.568	0.62 ± 0.057	0.658	0.32 ± 0.106	-

). Heterozygosity deficits were locus-specific. KR779884 exhibited a slight heterozygote excess (F ≈ −0.04), whereas KU953385 showed strong, population-dependent deficits (F = 0.785 ± 0.127) (Table 1).

Population-level patterns mirrored local demographic histories. Athos retained 100% of polymorphic loci and harboured roughly three times as many alleles per locus (31.3 ± 1.4) as any other stand, yet still showed an overall heterozygote deficit (Ho = 0.486 ± 0.080 < He = 0.701 ± 0.075, F = 0.308 ± 0.081) (

Table 2. Summary statistics (mean and standard error—SE) of genetic diversity in four populations, including population size after data processing, percentage of polymorphic loci (%p), number of different alleles per locus (n/L), observed heterozygosity (Ho), adjusted observed heterozygosity (Ho_adj), expected heterozygosity (He), adjusted expected heterozygosity (He_adj), fixation index (F), and adjusted fixation index (F_adj).

Populations	Population Size	p (%)	n/L	Ho	Ho_adj	He	He_adj	F	F_adj
Athos	37	100.00	31.286 ± 1.409	0.486 ± 0.08	0.732	0.701 ± 0.075	0.757	0.308 ± 0.081	0.023
Sithonia	12	100.00	9.857 ± 0.508	0.436 ± 0.078	0.55	0.666 ± 0.075	0.72	0.375 ± 0.076	0.271
Perama	10	71.43	7.143 ± 1.033	0.388 ± 0.149	-	0.485 ± 0.128	-	0.215 ± 0.188	-
Thassos	8	100.00	6.571 ± 0.481	0.502 ± 0.139	0.604	0.628 ± 0.077	0.668	0.301 ± 0.185	0.181
Mean	-	92.86 ± 7.14	13.714 ± 0.858	0.453 ± 0.112	-	0.62 ± 0.089	-	0.23 ± 0.133	-

). Correcting for null alleles reconciled this imbalance almost completely (Ho_adj = 0.732, F_adj ≈ 0.023) (Table 2). In contrast, Perama—a small, fragmented mainland population—was polymorphic at only 71% of loci and contained just 7.1 ± 1.0 alleles per locus (Table 2). Its observed heterozygosity remained the lowest recorded (0.388 ± 0.149) and could not be adjusted owing to scant data, indicating persistent inbreeding (F ≈ 0.22) (Table 2). Sithonia and Thassos were intermediate, each polymorphic at all loci but with reduced allelic richness (9.9 and 6.6 alleles locus per locus, respectively) and moderate fixation indices, even after adjustment (F_adj ≈ 0.27 and 0.18) (Table 2). Notably, some of these differences may, in part, reflect variations in sample size among populations, as smaller samples are more likely to underestimate allelic richness and within-population diversity, potentially biasing certain parameter estimates in Table 1 and Table 2.

Among-population differentiation remained low but significant (Fst = 0.042, p = 0.001), while Fis was at 0.494 (p = 0.001). AMOVA therefore attributed 95.8% of the total variance to individuals and only 4.2% to populations, yielding a global Fit of 0.515 (Table S3).

PCA (Figure 2) did not provide clear insights into the population structure. All populations seem to be genetically close, with no distinct pattern or clustering. DAPC (Figure 3) recovers weak among-population structure (FST ≈ 0.04) with a single mainland outlier, Perama, which diverges sharply from the largely overlapping Mount Athos, Sithonia, and Thassos groups.

Spatial DAPC membership coefficients (Figure 4) corroborate this: Perama is assigned almost exclusively to one cluster, whereas Athos exhibits extensive admixture and the two island stands are intermediate, supporting a metapopulation in which Athos acts as a genetically diverse core while Perama is a bottlenecked, isolation-prone fragment.

The spatial Principal Component Analysis (sPCA) reveals both broad-scale and fine-scale genetic structure within the Athos population. The eigenvalue plot (Figure 5) shows several large positive eigenvalues, indicating strong global structure—genetic similarity among geographically close individuals—while smaller negative eigenvalues point to the presence of local structure, or fine-scale genetic differentiation among neighbours. In Figure 6A, three geographically coherent genetic clusters are evident: individuals at the western tip (24.20° E, 40.25° N) exhibit negative scores, those in the central area (24.25° E, 40.20° N) show intermediate-to-positive scores, and individuals in the southeastern tip (24.35° E, 40.20° N) have uniformly high scores. This pattern suggests a clear east–west genetic cline with minimal overlap among clusters, likely driven by isolation by distance or spatially restricted gene flow. In contrast, the local structure (Figure 6B) highlights fine-scale spatial patterns: while the western and southeastern groups remain genetically homogeneous (scores ≈ 0), the central region displays sharp, metre-scale contrasts between positive (black) and negative (white) scores.

Figure 7 further illustrates this spatial genetic differentiation using a colour plot of the first two principal components based on the RGB colour system. Each point represents an individual positioned according to geographic coordinates, and the colour reflects genetic similarity: red, green, and blue intensities correspond to the first three principal components. The resulting pattern reveals three primary clusters matching those in Figure 6A—green in the northwest, red in the central region, and yellow in the southeast—supporting the presence of both broad-scale and localized genetic structure.

The geographic distribution of individuals on Mt. Athos reveals two spatially distinct clusters (Figure S1). Individuals sampled in close proximity—such as those from the central and southeastern regions—form tight, separate geographic clusters, reflecting the same genetic structure and limited gene flow inferred from the sPCA analyses (Figure 8).

4. Discussion

Genotyping 67 trees from four black pine stands with nuclear SSRs showed that genetic diversity is concentrated in Athos. Every locus was polymorphic there, and the population showed nearly triple the regional average in allele diversity, while the smaller samples of Sithonia, Perama, and Thassos carried notably fewer alleles per locus. Analysis of molecular variance showed that the vast majority of genetic variations were found among individuals within populations, with only a small proportion attributed to differences between populations, a result that is expected when hypervariable markers, such as microsatellites, are used [21]. These results identify Athos as the primary reservoir of allelic richness in our study and highlight a possible genetic erosion causing heterozygote deficits in the peripheral populations—especially the isolated Perama forest. Inbreeding was detected in the three smaller populations, while no such pattern emerged in Athos. However, it is important to note that these peripheral populations were represented by relatively small samples, which may have influenced the apparent levels of inbreeding or limited the detection of rare alleles [22].

The genetic patterns found within Mt. Athos correspond to the intense demographic fragmentation of the population, as it is not genetically uniform presenting moderate levels of genetic diversity. The spatial genetic patterns demonstrated a strong global component among patches, especially on the south–north axis, while a local component remained profound as well. These patterns suggest a possible fragmentation of an initially larger population, followed by several subsequent founder effects [23], due to the ability of black pine to colonize degraded areas [24]. This could also indicate the idea of a naturally established population, with a large group of trees showing characteristics of a natural population and other smaller groups differing from it. An alternative hypothesis suggests that the populations may have originated from a smaller forest stand that later expanded [23]. Finally, another scenario includes the idea that some smaller groups of trees may have derived from reforestation efforts [25].

Despite the small size of the population and its fragmented nature, the black pine population of Mt. Athos is diverse and demonstrates spatial genetic levels of differentiation at both global and local scale. This diversity may prove important for adaptation to future environmental changes in a region of critical importance from an ecological and cultural point of view. Since small and fragmented populations are often vulnerable to stochastic events, including wildfires, targeted measures of conservation and management should be planned and implemented against both exogenous and endogenous threats to these ecosystems [26]. An effective strategy should involve both in situ and ex situ measures, utilizing on site protection measures against threats with the reintroduction and population restoration for bridging the gaps between patches and creating gene flow corridors [22,27,28] among the fragments of the black pine population on Mt. Athos. This can be facilitated by systematically collecting, storing, testing, and maintaining a representative pool of reproductive material (seed). Seedlings raised from this material can be carefully introduced in selected areas. This restoration strategy will allow seedlings to grow in local nurseries, establishing a controlled environment that promotes genetic diversity [29]. These management strategies are essential for preserving and restoring black pine populations in Mt. Athos and other sites in Greece. Combining these strategies with a broader study of species’ genetic diversity across Greece could enrich biodiversity and safeguard genetic heritage in the long term.