Genetic Variation Within and Among Populations of a Dominant Desert Tree Haloxylon ammodendron (Amaranthaceae) in China

Abstract

• Background and Aims China is one of the countries most severely affected by desertification. Haloxylon ammodendron (Amaranthaceae) is an ecologically important component of the desert ecosystem and is one of the main tree species used for restoration, yet we know little about its genetic structure.

• Methods Genetic variation within and between nine populations of H. ammodendron from two regions of China was investigated using ISSR (inter-simple sequence repeat) markers.

• Key Results Eight primers used in this study amplified 219 reproducible bands of which 184 (84 %) were polymorphic. Analysis of molecular variance (AMOVA) revealed high genetic variation within populations (97·63 %) and low genetic differentiation between regions (0·62 %) and among populations (1·75 %).

• Conclusions It is suggested that the present genetic structure could have arisen by high levels of gene flow. The gene flow among populations observed here is probably mainly attributable to pollen movement. The genetic structure also has important implications in ecological restoration practice.

INTRODUCTION

China is one of the countries most severely affected by desertification. The total area of desertified land in China is approx. 2 622 300 km², or 27·32 % of the total land territory. Desertification is still occurring at a rapid pace (CCICCD, 2000), and rehabilitation and ecological restoration are major tasks. A primary goal of restoration is the rapid establishment of long-term viable populations that will help restore ecosystem functions and processes, prevent erosion and protect biological diversity (Bradshaw, 1987; Lesica and Allendorf, 1999). The success of a vegetation restoration project might depend on choosing the appropriate genetic sources, and the use of native plants in ecological restoration has received much attention recently (Knapp and Rice, 1994; Linhart, 1995; Lesica and Allendorf, 1999), especially for use in stressful environments.

It is often stated that only by possessing genetic variation will a given species be able to respond to environmental pressure, evolve, and survive in the long term (e.g. Ayala and Kiger, 1984). The amount and partitioning of genetic variation among and within populations is correlated with the life history characteristics and population history of a species, and environmental factors (Hamrick and Godt, 1989). Hoffmann and Parsons (1991) stated that severe environmental stress could be regarded as an environmental probe, leading to useful evolutionary insights that are difficult to reveal under less extreme conditions. Despite extensive ecological research on plants from arid regions, there is a lack of information concerning levels of genetic variation in these systems (Hu and Wang, 1998). Knowledge of the genetic structure of desert plants not only permits us to estimate the importance of some evolutionary forces, such as selection, mutation, migration and genetic, drift under stressful environmental conditions, but also provides fundamental information for plans for restoration and rational exploitation.

Haloxylon ammodendron is a xerophytic desert tree. Due to its great drought resistance and saline tolerance, H. ammodendron occurs naturally in various habitats, including gravel desert, clay desert, fixed and semi-fixed sand, and saline land in the Asian and African deserts (Chen et al., 1983; Tobe et al., 2000). In China, about 56 % of H. ammodendron is found in Xinjiang, 40 % in Inner Mongolia, and the remaining 4 % in Qinghai and Gansu Provinces (Ma et al., 2000). As a dominant desert plant, H. ammodendron plays an important role in the maintenance of the structure and function of the whole ecosystem, reducing wind speed and ameliorating the forest microclimate, thus facilitating the settlement and growth of other desert plants (Shamsutdinov and Ubaidullaev, 1988). During the last five decades, land reclamation and cultivation, over-grazing, over-cutting and digging have resulted in the destruction of H. ammodendron vegetation and, as a result, have increased mobility of sand dunes that are no longer stabilized by root systems. Compared with a field survey in 1958, over 50 % of the natural populations of the species have become extinct or degraded (Huang, 2002). Meanwhile, the genetic diversity and population genetic structure of H. ammodendron are largely unknown, making the conservation and utilization in ecological restoration of this important desert species difficult.

In recent years, the technique of inter-simple sequence repeats (ISSR) has been used in the detection of genetic diversity in many species (Robinson et al., 1997; Wolfe et al., 1998a, b; Esselmann et al., 1999; Qian et al., 2001). The ISSR technique has several advantages including a relatively unbiased sampling of the genome, simplicity, relatively low cost, and the requirement for only a small amount of plant material. In the present study, ISSR markers were used to examine the genetic structure of natural populations of H. ammodendron in China. Of particularly interest were the level of genetic diversity within populations and the level of genetic differentiation among populations, information that might provide important insights into the population genetics of the species and facilitate its use in vegetation restoration.

MATERIALS AND METHODS

Plant material

Nine populations of Haloxylon ammodendron (C. A. Mey.) Bunge (Saxaul; Amaranthaceae), which could be grouped into two regions, Xinjiang (Fig. 1A–G) and Inner Mongolia (Fig. 1H and I), were sampled. Because the leaves of H. ammodendron are reduced to scales (Pyankov et al., 1999), the young annual cylindrical shoots were used as material for DNA extraction. In each of the nine populations, shoot tissue from 20 mature individuals was collected (Table 1). All tissue was dried at room temperature (Thomson and Henry, 1993) and transported back to the Institute of Botany, Beijing.

Fig. 1.

Map showing Haloxylon ammodendron populations analysed in this study. See Table 1 for reference to the population locations.

Table 1.

Populations of Haloxylon ammodendron for ISSR analysis

Population no.	Location	Latitude (°N)	Longitude (°E)	Samples size	Voucher*
A	Beishawo, Fukang City, Xinjiang	44·24	87·52	20	Qin RC 623
B	Xingonglu, Fukang City, Xinjiang	44·13	87·50	20	Qin RC 623
C	Beishawo, Fukang City, Xinjiang	44·25	87·55	20	Qin RC 623
D	Wucaiwuan, Jimusaer County, Xinjiang	44·50	88·55	20	Qin RC 623
E	Dongdaohaizi, Miquan City, Xinjiang	44·41	87·33	20	Qin RC 623
F	Baikouquan, Kelamayi City, Xinjiang	45·53	85·24	20	Li AR, Zhu AN 5877
G	Mosuowan, Manasi County, Xinjiang	45·03	86·09	20	XZ 0054
H	Wulatehou, Bayannaoer, Inner Mongolia	42·10	108·35	20	0027
I	Alashanzuo, Alashan, Inner Mongolia	39·30	104·59	20	0066

^*Vouchers are deposited in the Chinese National Herbarium, Institute of Botany, Beijing.

Genomic DNA extraction

Genomic DNA was extracted from approx. 1 g of dried shoot material using a modification of the 2× CTAB method of Doyle and Doyle (1987). The tissue was ground to fine powder in liquid nitrogen and incubated at 65 °C for 60 min in 2 mL of 2× CTAB isolation buffer [100 mM Tris–HCl, pH 8·0, 1·4 m NaCl, 20 mm EDTA, 2 % hexadecyltrimethyl-ammonium bromide (CTAB), 0·2 % β-mercaptoethanol]. The sample was mixed with an equal volume of Tris–phenol, centrifuged at 10 000 g for 10 min and then the supernatant was collected in a clean tube. An equal volume of chloroform–isoamyl alcohol (24:1) was added and mixed by inversion for 10 min. Then the mixture was centrifuged at 10 000 g for 10 min. The supernatant was mixed with 0·5 mL ice-cold isopropanol and 0·1 mL 2 m potassium acetate to precipitate the DNA. The DNA was pelleted by centrifugation at 10 000 g for 5 min, washed twice with 0·6 mL of 70 % ethanol, air-dried at room temperature and resuspended in 0·2 mL of 0·1× TE buffer (10 mM Tris–HCl, 1 mm EDTA; pH 8·0). All DNA samples were purified using the Wizard DNA Clean-Up System (Promega, Madison, WI, USA). After visual quantification by comparison with standard DNA concentrations, DNA samples of approx. 50 ng µl⁻¹ were prepared.

ISSR amplification

Twenty-one primers from Shanghai Sangon Biologic and Engineering Technology and Service Co. Ltd (Shanghai, China) were screened initially against four plants from four populations. Primer design was based on SSR motifs reported for flowering plants (reviewed in Wolfe and Liston, 1998; Wolfe et al., 1998a, b) and recommended on at http://www.biosci.ohio-state.edu/~awolfe/ISSR/ISSR.html. To avoid biasing estimates of polymorphism, the selection of primers for band scoring was dependent only on the clearness and reproducibility of ISSR fragments, not on the level of polymorphism. Reactions were carried out in a volume of 25 µl consisting of 1·5 mm MgCl₂, 0·2 mM dNTPs, 1·25 μm of primer, 1·25 units of Taq DNA polymerase (Promega), 1× Taq DNA polymerase buffer and approx. 20 ng of template DNA. PCR amplifications were performed in a PTC-200 thermocycler (MJ Research, Watertown, MA, USA) under the following conditions: 94 °C for 1·5 min; 35 cycles of 94 °C for 40 s, 49 °C for 45 s, 72 °C for 1·5 min; linked to 94 °C for 45 s, 44 °C for 45 s, 72 °C for 5 min. PCR products were then stored at 4 °C. The amplification products were separated by electrophoresis on 2 % agarose gels in 1× TBE (Tris–EDTA–borate) buffer, stained with ethidium bromide. After running for approx. 6 h at 50 V (2·5 V cm⁻¹) the gel was photographed by an Alpha Ease FC Imaging System (Alpha Innotech Corporation, San Leandro, CA, USA). Scoring of the bands was performed by visual analysis of the gel photographs. Molecular size of the fragments was estimated using a 200-bp DNA ladder.

Data analysis

Amplified fragments were scored for presence (1) or absence (0) of putatively homologous bands and a matrix of ISSR phenotypes was assembled. Homology was assessed based on co-migration of bands on the gels. Only data from intensely stained, unambiguous, clear bands were used for statistical analysis. An ISSR marker was determined to be polymorphic when it was found in <95 % of the individuals sampled (i.e. absent in eight or more individuals). Genetic diversity was measured as the percentage of polymorphic loci. Analysis of molecular variance (AMOVA) was conducted using DCFA 1.1 (Zhang and Ge, 2002) and WINAMOVA 1.55 (Excoffier et al., 1992; Excoffier, 1993) to calculate variance components and their statistical significance levels for variation among and within the nine populations. Euclidean squared distances () instead of Jaccard similarity coefficients were used in AMOVA as suggested by Zhang and Ge (2002) and Excoffier (1993). The was obtained by the following equation:

where for two individuals x and y, x_i and y_i represent presence or absence of a particular molecular marker, and s is the total number of markers in the AMOVA. Pairwise genetic distances (Φ_ST) among the nine populations were obtained from the AMOVA and were used to construct a UPGMA tree (routine UPGMA in the NTSYSpc package; Rohlf, 1994). A nonparametric test for homogeneity of molecular variance (HOMOVA), based on the Bartlett statistic, was also applied to test whether the populations have different amounts of ISSR variation (Stewart and Excoffier, 1996). Both HOMOVA and AMOVA were performed using the WINAMOVA 1.5 program (obtained at ftp: 129.194.113.1). The Mantel test (Mantel, 1967) was used to test whether matrices of genetic distances between populations were significantly correlated with matrices of geographic distances (1000 permutations; routine MXCOMP in NTSYSpc).

RESULTS

ISSR diversity

Of the 21 ISSR primers screened, eight produced clear and repeatable fragments and were selected for further analysis (Table 2). These primers consistently amplified a total of 219 scorable markers of 165–2300 base pairs (bp) in size. Individual primers generated 18–36 bands, with a mean of 27·4. Of the 219 markers, 184 (84 %) were polymorphic across the 180 plants, and each individual had a unique multilocus ISSR phenotype. Some primers were more efficient in differentiating between individuals than others. For example, ISSR-7 and ISSR-12 identified 179 unique phenotype profiles, whereas ISSR-6 only revealed 124 profiles (Table 2).

Table 2.

Primers for ISSR analysis

Primer	Primer sequence*	Band size range (bp)	Tm (°C)	Total no. of loci	Total no. of polymorphic loci (%)	Total no. of distinct phenotypes
ISSR-4	(AC)8YT	190–1000	44	22	19 (86·4)	160
ISSR-6	(CA)6RY	180–1200	44	18	15 (83·3)	124
ISSR-7	(CA)6RG	180–2300	43	36	30 (83·3)	179
ISSR-9	(CTC)4RC	210–2150	48	32	24 (75·0)	169
ISSR-11	BDB(ACA)5	185–1050	49	24	21 (87·5)	146
ISSR-12	BBB(GAAA)3GAA	165–1800	43	33	27 (81·8)	179
ISSR-13	(AC)8YG	190–1700	53	30	27 (90·0)	178
ISSR-14	(CAC)4RC	180–1000	60	24	21 (87·5)	171

^*Y = G/C; R = A/T; B = C/G/T; D = A/G/T.

Percentages of polymorphic loci per population and per region are shown in Table 3. The percentage of polymorphic loci within populations ranged from 48·9 % (A) to 74·9 % (E), with the mean percentage of 57·7 %. Within regions, the percentage of polymorphic loci of the populations from Xinjiang was 78·5 %, a little higher than that from Inner Mongolia (70·8 %).

Table 3.

Percentage of polymorphic ISSR loci in each Haloxylon ammodendron population and region (see Fig. 1)

	A	B	C	D	E	F	G	Xinjiang Province	H	I	Inner Mongolia	Total
No. of PL	107	110	110	109	164	146	122	172	127	142	155	184
Percentage of PL	48·9	50·2	50·2	49·8	74·9	66·7	55·7	78·5	58·0	64·8	70·8	84·0

PL, polymorphic loci.

Genetic structure

The AMOVA of the distance matrix for the 180 individuals permitted a partitioning of the overall variation into three levels (Table 4). The proportion of variation attributable to within-population differences was high (97·63 %), whereas only 0·62 % and 1·75 % occurred between regions and among populations, respectively. Genetic differences between Xinjiang and Inner Mongolia were not significant (P = 0·202), whereas those between populations were significant (P < 0·05). Bartlett's test for the homogeneity of the ISSR variance indicated significant differences in the amount of genetic variability present in the nine populations (B = 0·194, P < 0·001).

Table 4.

Analysis of molecular variance (AMOVA) for 180 individuals in nine populations of Haloxylon ammodendron

Source of variance	d.f.	SSD	MSD	Variance component	%Total	P
Among regions	1	0·6079	0·608	0·002	0·62	0·202
Among populations	7	3·2983	0·471	0·006	1·75	<0·05
Within populations	171	59·2825	0·347	0·347	97·63	<0·001

SSD, sum of squared deviation; MSD, mean squared deviation.

Pairwise Φ_ST values between populations and their significance are presented in Table 5. Values of Φ_ST ranged from −0·0069 (D–F) to 0·07 (C–H), indicating that populations D and F were the most similar and populations C and H were the most different. An UPGMA dendrogram of the nine populations based on the pairwise Φ_ST values is shown in Fig. 2.

Table 5.

Pairwise genetic distances (Φ_ST) among nine populations of Haloxylon ammodendron

Population	A	B	C	D	E	F	G	H	I
A	–
B	0·0374	–
C	0·0453***	0·0154	–
D	0·0311*	0·0061	0·0030	–
E	0·0365*	0·0310***	0·0196	0·0051	–
F	0·0143	0·0106	0·0111*	−0·0069	−0·0035	–
G	0·0152	0·0092	0·0280	0·0150	0·0206*	−0·0052	–
H	0·0030	0·0556***	0·0700***	0·0610***	0·0474***	0·0224	0·0128*	–
I	0·0224*	0·0094	0·0293***	0·0159***	0·0278***	0·0050	–0·0134	0·0229	–

P values indicate the probability that a random genetic distance is larger than the observed distance and are based on 1000 permutations.

^*P < 0·05;

^***P < 0·001.

Fig. 2.

Dendrogram of the nine Haloxylon ammodendron populations obtained by UPGMA cluster analysis of a matrix of pairwise Φ_ST values.

Correlation of geographical and genetic (ISSR) distance

The matrix of pairwise genetic distances was not significantly correlated with the corresponding matrix of geographic distances (Mantel tests; r = 0·2230, P = 0·8448; Fig. 3). Thus, differentiation observed among populations (Table 4) did not directly correspond to the geographic distance.

Fig. 3.

The relationship between pairwise geographic and pairwise genetic distances (Φ_ST) among nine populations of H. ammodendron. Distances >1000 km denote pairs with one population in Xinjiang and the other at Inner Mongolia.

DISCUSSION

Genetic diversity and population structure

Using ISSR markers, 84 % of loci were found to be polymorphic in H. ammodendron. Levels of genetic variation within populations established for H. ammodendron can be compared with other woody plants with similar life histories. Ge and Sun (1999) revealed a surprisingly low genetic variation with ISSR (P = 16·18 %) in Aegiceras corniculatum (Myrsinaceae), a widespread, long-lived, woody, perennial species with mixed-mating to outcrossing systems. Lacerda et al. (2001) using RAPD markers, found considerable levels of intrapopulation genetic variation (P = 70·8 %) for Plathymenia reticulata (Fabaceae), a tree from the Brazilian Cerrado. Shrestha et al. (2002) used 18 RAPD primers to detect the genetic diversity of 12 populations of Acacia raddiana (Mimosaceae), a keystone tree species in the Negev desert of Israel. Of the 290 markers revealed, 90·69 % were polymorphic. By comparison, it can be estimated that H. ammodendron presents high levels of intrapopulation genetic variation.

In H. ammodendron, ISSR markers revealed that the majority of the variation was found within populations rather than among populations. The genetic differentiation among regions was not significant. Long-lived, woody plants typically harbour a greater percentage of their variation within populations (Hamrick and Loveless, 1989; Hamrick et al., 1992). According to Hamrick and Godt (1989), reproductive biology is the most important factor in determining the genetic structure of plant populations. They showed that outcrossing plant species tend to have 10–20 % of the genetic variation among populations, whereas selfing species have, on average, 50 % of the variation among populations. Studies of the biology of flowering and pollination in H. ammodendron indicate that it is an outcrosser (Tursunov et al., 1989). Despite several possible pollination mechanisms, such as xenogamy, geitonogamy and autogamy, the presence of three groups of bisexual flowers that differ in time of maturation of the anthers and stigma promotes cross-pollination by anemophily (Tursunov et al., 1989). However, Hamrick and Godt (1996) have pointed out that life history traits alone explain a relatively low amount of the variation in genetic structure that is seen among species. The high intrapopulation variability and genetic homogeneity across populations could have arisen by high levels of gene flow. Alternatively, it may be that these populations have simply not been separated long enough to accumulate detectable genetic differences.

Haloxylon ammodendron generally reproduces sexually (Song and Jia, 2000), although in the young forest, vegetative propagation from buds on the trunk or branches has been reported when the plant is damaged (Huang, 2002). In the mature natural populations studied, however, vegetative propagation is infrequent. Seeds are small (2·5mm in diameter, 1000 seeds weighing 2·99 ± 0·04 g; Huang, 2002), and bear wing-like membranes. Occasionally seeds can fall >5 miles away from the canopies of their maternal trees (Mamedov and Osipov, 1985; Y. Sheng et al., unpubl. res.). Most of the seeds are consumed by ants and small mammals, which is a major cause of seed mortality in the desert (Brown and Davidson, 1977). Seed longevity of H. ammodendron is short under natural conditions, and few seeds survive in the soil for >10months (Huang et al., 2002). In deserts, the water necessary for seedling establishment is only available transiently following precipitation (Ellner and Shmida, 1981), and the dispersed seeds must geminate quickly after precipitation during April and May, or they will die the following summer (Yang et al., 1995). Dispersal and germination investigations, as presented above, suggest that long-distance seed dispersal is rare in H. ammodendron. Absence of adaptation for long-distance dispersal in desert plants has long been recognized. Ellner and Shmida (1981) argue that atelechory (the absence of dispersal-enhancing characters) in desert species is an adaptive response to the extremely low benefit of long-range dispersal mechanisms in deserts. Pollen is small (19·1µm in diameter; Pan, 1993), round or nearly round, and wind-dispersed (Wu et al., 1992). Due to the frequently strong winds and the low coverage of vegetation in the desert, long-distance dispersal of pollen by wind is possible. After pollination and fertilization in May, the ovaries do not begin to develop until September, when the temperature and moisture levels are suitable (Song and Jia, 2000). This special habit, summer dormancy, ensures effective pollen flow. Therefore, the high levels of gene flow among populations observed here might be mainly attributed to pollen movement. Further studies of pollination biology and parentage analysis using co-dominant markers are needed to clarify this point. For most biological systems, the most powerful genetic tools for parentage analysis will be microsatellite markers, and most of the recent advances in techniques of data analysis have been aimed at studies employing microsatellites (Jones and Ardren, 2003). Using microsatellite analysis, Dow and Ashley (1998) characterized pollen dispersal in a stand of 62 adult bur oaks (Quercus macrocarpa), providing direct evidence for high levels of long-distance pollination in this wind-pollinated species. Konuma et al. (2000) using microsatellite polymorphism suggested that long-distance gene flow and seed migration were responsible for the poorly developed genetic structure of the tropical rainforest tree Neobalanocarpus heimii (Dipterocarpaceae). Recent review papers (Estoup and Angers, 1998; Squirrell et al., 2003; Zhang and Hewitt, 2003) have addressed the advantages and disadvantages of the approach and have suggested methods for isolation of microsatellite markers at relatively low cost.

Restoration implications

The genetic structure of H. ammodendron revealed in this study is important in ecological conservation and restoration practice. The fact that an overwhelming proportion of the variability is present within populations suggests that smaller numbers of populations will be required for effective conservation compared with island endemics with many, strongly isolated populations. Likewise, the high genetic homogeneity among populations of H. ammodendron suggests that the exact population used as a source of material for restoration may be less crucial in this species, compared with other species with strong genetic structure and a high potential for local adaptation. Population size is also an important restoration consideration in H. ammodendron. It is thought that small populations can lose large amounts of genetic variation due to of genetic drift and thus have reduced probabilities of long-term viability (Lande and Barrowclough, 1987). Moreover, small populations are more prone to extinction from random environmental fluctuations (Allendorf, 1986). Therefore, maintaining a proper population size should be emphasized in planning restoration and rational exploitation of H. ammodendron.