Immunogenetics

, Volume 55, Issue 10, pp 674–681

Gm and Km immunoglobulin allotypes in Sicily

Authors

  • Nicoletta Cerutti
    • Department of Animal and Human BiologyUniversity of Turin
  • Jean Michel Dugoujon
    • Anthropology CenterUMR 8555
  • Evelyne Guitard
    • Cellular and Molecular Physiopathology UnitCNRS
    • Department of Animal and Human BiologyUniversity of Turin
Original Paper

DOI: 10.1007/s00251-003-0628-z

Cite this article as:
Cerutti, N., Dugoujon, J.M., Guitard, E. et al. Immunogenetics (2004) 55: 674. doi:10.1007/s00251-003-0628-z

Abstract

The aim of this study was to evaluate the intra- and inter-population variability of the Gm/Km system in the Madonie Mountains, one of the main geographical barriers in north-central Sicily. We analysed 392 samples: 145 from Alia, 128 from Valledolmo, 25 from Cerda and 94 from Palermo. Serum samples were tested for G1m (1,2,3,17), G2m (23), G3m (5,6,10,11,13,14,15,16,21,24,28) and Km (1) allotypes by the standard agglutination-inhibition method. We found the typical genetic patterns of populations in peripheral areas of the Mediterranean basin, with a high frequency of haplotypes Gm5*;3;23 and Gm5*;3;... The frequency of Gm21,28;1,17;... (about 16%) is rather high compared with other southern areas. Of great importance is the presence of the common African haplotype Gm 5*;1,17;..., ranging in frequency from 1.56% at Valledolmo to 5.5% at Alia. The presence of this haplotype suggests past contacts with peoples from North Africa. The introduction of African markers could be due to the Phoenician colonization at the end of the 2nd millennium b.c. or to the more recent Arab conquest (8th–9th centuries a.d.).

Keywords

SicilyGenetic markersPopulation geneticsImmunoglobulin allotypes

Introduction

Because of its central position in the Mediterranean Sea, Sicily has been the meeting place of numerous peoples. They not only exchanged goods and traditions but also their genes. Since the arrival of the first human groups (Sicani, Siculi, Elymi), the island has been subjected to numerous migratory flows. Archaeological remains and important monuments discovered on the island testify to the successive passage of Greeks and Phoenicians, Etruscans and Romans, Vandals and Goths, Byzantines and Arabs, Aragonese and Normans.

In agreement with the complex history of dominions and colonizations, recent anthropogenetic studies have demonstrated strong heterogeneity in the genetic structure of Sicilian populations (Rickards et al. 1992, 1998;Giambona et al. 1995; Cittadella et al. 1997; Vona et al. 1998, 2001; Ghiani et al. 2002; Romano et al. 2003). Various studies have found differences between western and eastern Sicilian populations. This diversity is shown by historical and cultural data, surnames (Guglielmino et al. 1991), dialect isoglosses (Ruffino 1997) and classical genetic markers (Piazza et al. 1988). However, a genetic analysis by Rickard et al. (1998) failed to find this geographical pattern.

Recent data on DNA mutations (Piazza et al. 1998; Romano et al. 2003) have confirmed the differentiation, referable to Greek colonization in the eastern part of the island (8th century b.c.) and Phoenician-Carthaginian dominion in the triangle of western Sicily.

Studies of immunoglobulin allotypes have been conducted, but they are still rather patchy (Marziano 1968; Walter et al. 1997).

To broaden our knowledge of the genetic characteristics of Sicily, we conducted an anthropogenetic investigation in Palermo and three small rural towns (Alia, Cerda, Valledolmo). The aim of the study was to evaluate the intra- and inter-population variability of the Gm/Km phenotypes and haplotype frequencies in the Madonie Mountains. We then compared our results with other data to analyse the genetic relationships among Mediterranean populations. The occurrence of certain haplotypes permits the reconstruction of migrations to the island, establishing ancient contacts of the Madonie people with other Mediterranean populations.

Historical background

Alia (700 m), Cerda (274 m) and Valledolmo (750 m) are situated in the heart of the Madonie Mountains, one of the main geographical barriers in north-central Sicily (Fig. 1). The highest peak, Pizzo Carbonara, is 1,975 m. In the past, it was difficult to reach this area, which remained isolated on account of its geological structure and various historical events.
Fig. 1

Geographical locations of the Sicilian communities and reference populations

The history of Alia, Cerda and Valledolmo is generally linked to the complex pattern of foreign domination of Sicily. The first historical evidence shows that, at the end of the 9th century a.d., several Arab villages flourished in the wide valley that now contains these towns. Thanks to agriculture and livestock breeding, the inhabitants reached a considerable degree of prosperity.

Alia, originally Lalia (Guccione 1991), was cited in documents as early as 1296, the year in which Frederick II of Aragon officially became King of Sicily. Graves and various objects discovered in the countryside around Alia demonstrate the presence of Roman, Byzantine, Arab and Norman settlements. Of particular historical interest is the archaeological complex known as the Gurfa Caves, containing Phoenician inscriptions (Runfola 1978). This is the first evidence of the presence of Phoenicians in an inland area of Sicily. Alia currently has around 3,500 inhabitants.

Valledolmo originated as a feudal estate called Castrum Normandum (11th century a.d.). According to a Rivelo of 1682, the population numbered 243 inhabitants, belonging to about 60 families. Contemporary documents indicate that more than half of them were of local origin. The current number of Valledolmo residents is around 4,400.

Complete lists of the names of all family heads from 1682 to 1811 have been published for Alia (Guccione 1991) and Valledolmo (Granata 1982). They show that many of the surnames of the subjects analysed in our study appeared in the old Riveli.

The site of Cerda was also known as Fondaco Nuovo, as it was a rest and refreshment stop for travellers to the Madonie area. The population of Cerda is currently around 5,000 inhabitants.

Gm and Km systems

Of the four human immunoglobulin systems (Gm, Am, Em and Km), the Gm system (containing 18 allotypes) shows the greatest degree of polymorphism and is thus most often studied in population genetics. However, the Gm system has not been fully explored in some anthropological studies due to the difficulty in finding good typing reagents for certain specificities. Gm allotypes are restricted to one subclass and located on the heavy chains (constant domains CH1, CH2 and CH3) of the IgG1:G1m(1,2,3,17), IgG2:G2m(23) and IgG3:G3m(5,6,10,11,13,14,15,16,21,24,26,27,28). As the genes coding for heavy chains are closely linked, the allotypes G1m, G2m and G3m are inherited in fixed combinations or haplotypes (Lefranc and Lefranc 1990). Their composition and frequency differ among human ethnic groups (Schanfield 1980; Steinberg and Cook 1981). Therefore, Gm allotype polymorphism is considered a good genetic marker for population studies.

Three Km allotypes have been defined: Km (1), Km (2) and Km (3). They are all situated on the constant region (CL) of the kappa light chain and thus occur in all Ig classes and subclasses. Gm and Km allotypes are associated with amino acid substitutions on the constant region of heavy and light chains. The gene cluster coding for heavy chains of Ig is assigned to Chromosome 14q32.3 (Lefranc et al. 1982) and the kappa gene is located on Chr. 2p12 (McBride et al. 1982).

Materials and methods

We analysed 392 samples: 145 from Alia, 128 from Valledolmo, 25 from Cerda and 94 from Palermo. Information provided by the subjects indicated the presence of their families in the town for at least three generations.

Serum samples with 0.2% sodium azide were kept at −20 °C until typing. All the samples were tested for G1m (1,2,3,17), G2m (23), G3m (5,6,10,11,13,14,15,16,21,24,28) and Km (1) allotypes by the standard agglutination-inhibition method (Field and Dugoujon 1989). All the antisera were of human origin except for anti-G2m (23), a mouse monoclonal antibody (Dugoujon et al. 1989).

The Gm haplotype and Km allele frequencies were estimated by the maximum likelihood method (Edwards 1984). The analysis of heterogeneity among the Madonie, Sicilian and continental Italian populations was performed with the χ2 test (using the Yates correction) or the G test (using the Williams correction) when χ2 could not be calculated because of the small number of cases.

We constructed a dendrogram directly from the haplotype frequencies according to the principle of maximum likelihood (Weir 1990). The reliability of the dendrograms was evaluated by the bootstrap technique (1,000 replications). Principal components analysis was applied to the haplotype frequencies to provide a synthetic picture of the Gm system in different Mediterranean populations.

The statistical analyses were performed with Genepop v. 3.1 (Raymond and Rousset 1995), Systat and Phylip v. 3.6a (Felsenstein 2000).

Results

The phenotype frequencies observed in the populations of the Madonie area and in the Palermo group are in agreement with Hardy-Weinberg equilibrium. The χ2 test is not significant in each of the groups tested. The Gm phenotypes are reported in Table 1. One subject (in the Alia and Madonie populations) with a probable Gm5*;-;23 haplotype was excluded from the statistical analysis. The frequencies of the Gm haplotypes are reported in Table 2.
Table 1

Gm phenotype frequencies in the Alia, Valledolmo, Cerda (Madonie) and Palermo populations (observed and expected). One subject with a probable Gm5*;-;23 haplotype was excluded from the statistical analysis (in the Alia and Madonie populations)

Alia

Valledolmo

Cerda

Madonie

Palermo

Gm phenotypes

Genotype

Obs.

Exp.

Obs.

Exp.

Obs

Exp.

Obs

Exp.

Obs.

Exp.

5*,21,28;1,3,17;23

14 18

25

24.6

33

28.9

6

4.5

64

57.3

15

16.5

5*;3;23

34 44

76

74.7

69

73.2

13

14.6

158

163.1

50

49.3

5*,21,28;1,2,17;...

25

2

0.7

0

0.2

0

0.0

2

1.0

0

0.0

5*;1,17;..

55

1

0.4

0

0.03

0

0.04

1

0.4

0

0.0

5*,21,28;1,3,17;...

13

8

8.8

6

3.3

2

1.9

16

14.6

6

5.8

5*,21,28;1,2,3,17;23

24 28

7

7.3

13

8.9

0

0.0

20

15.9

2

4.0

21,28;1,2,17;...

12 22

2

2.4

0

2.5

0

0.0

2

4.8

2

1.3

5*,21,28;1,2,3,17;...

23

2

2.6

0

1.0

0

0.0

2

4.1

2

1.4

5*;3;...

33

7

5.6

0

0.8

2

1.4

9

7.1

4

3.6

5*;1,3,17;...

35

1

3.0

1

0.3

0

0.5

2

3.2

0

0.0

5*;1,3,17;23

45 58

6

8.2

2

2.7

2

1.1

10

12.7

5

3.3

5*,21,28;1,17;...

15

4

2.3

1

0.7

0

0.3

5

3.4

1

1.0

21,28;1,17;..

11

3

3.5

1

3.4

0

0.6

4

7.6

3

2.4

5*,15,16;1,3,17;23

47 78

0

0.0

1

0.7

0

0.0

1

0.6

1

0.6

5*;1,3;23

38 48 88

0

0.0

1

1.0

0

0.0

1

1.03

0

0.0

5*,15;1,3,17;23

46

0

0.0

0

0.0

0

0.0

0

0.0

1

0.6

21,28;1,2,3,17;...

29

0

0.0

0

0.0

0

0.0

0

0.0

1

0.1

5*;1,2,3,17;23

49

0

0.0

0

0.0

0

0.0

0

0.0

1

1.1

5* = 5,10,11,13,14

 

 

 

 

 

 

 

 

 

 

 

Total

144

128

25

297

 

94

χ2

1.08

2.67

0.17

4.91

0.16

d.f.

8

 

9

 

6

 

9

 

10

 

Table 2

Gm haplotype frequencies at Alia, Valledolmo, Cerda (Madonie) and Palermo

GM haplotypes

Genotype no.

Alia

Valledolmo

Cerda

Madonie

Palermo

Freq.

SE

Freq.

SE

Freq.

SE

Freq.

SE

Freq.

SE

21,28;1,17;...

1

0.1554

0.0213

0.1641

0.0233

0.1600

0.0517

0.1595

0.0151

0.1584

0.0268

21,28;1,2,17;...

2

0.0460

0.0124

0.0508

0.0139

0.0000

-

0.0442

0.0085

0.0384

0.0144

5*;3;...

3

0.1970

0.0304

0.0778

0.0272

0.2400

0.0788

0.1545

0.0207

0.1948

0.0381

5*;3;23

4

0.5495

0.0352

0.6824

0.0362

0.5600

0.0870

0.6025

0.0248

0.5552

0.0438

5*;1,17;...

5

0.0521

0.0134

0.0156

0.0078

0.0400

0.0277

0.0353

0.0076

0.0319

0.0141

10,11,13,15;1,17;...

6

0.0000

-

0.0000

-

0.0000

-

0.0000

-

0.0053

0.0069

10,11,13,15,16;1,17;...

7

0.0000

-

0.0039

0.0039

0.0000

-

0.0017

0.0017

0.0053

0.0068

5*;1,3

8

0.0000

-

0.0054

0.0053

0.0000

-

0.0023

0.0023

0.0000

-

5*;-;... and -;1,2,17;...

9

0.0000

-

0.0000

-

0.0000

-

0.0000

-

0.0107

0.0090

The small number of Cerda inhabitants tested (n=25) precluded the calculation of heterogeneity and the phylogenetic and multivariate analyses for this population.

We found the typical genetic patterns of populations in peripheral areas of the Mediterranean basin, with high frequencies of haplotypes Gm5*;3;23 and Gm5*;3;...

The frequency of Gm21,28;1,17;... (about 16%) is rather high compared with other southern areas, even though it shows high variability in different Eurasian populations, as well as a clinal distribution from northwest to southeast. Haplotype Gm21,28;1,2,17;... has a low frequency (around 5%), similar to that of populations in continental Italy. This haplotype is rare in Sardinia (Piazza et al. 1976).

Of great importance is the presence of the common African haplotype Gm 5*;1,17;..., with a frequency of 3.19% at Palermo, 1.56% at Valledolmo, 4.00% at Cerda and 5.52% at Alia.

The Km phenotype and allele frequencies are reported in Table 3. The Km (1) phenotype includes the Km (1), Km (1,2), Km (1,2,3) and Km (1,3) phenotypes, since sera were tested for the Km(1) allotype only. In fact, positive sera for the Km (1) allotype are always positive for Km (2) as well, while negative sera for Km (1) are always positive for Km (3). The typing of Km (2) and Km (3) does not provide additional information. The Km1 allele corresponds to Km1 plus Km1,2, whereas Km3 corresponds more exactly to Km−1,−2. The frequency of the Km3 allele is 93.22% at Alia, 84.85% at Cerda, 86.60% at Valledolmo and 90.51% at Palermo.
Table 3

Observed Km phenotypes and allele frequencies. Sera were tested for the Km(1) allotype only, since the Km(1) phenotype includes the Km (1), Km (1,2), Km (1,2,3) and Km (1,3) phenotypes

Alia

Valledolmo

Cerda

Madonie

Palermo

Km phenotypes

  1

19

32

7

58

17

  −1

126

96

18

240

77

  Total

145

128

25

298

94

Km allele frequencies

  Km1 and Km 1,2

0.0678

0.1340

0.1515

0.1026

0.0949

  Km3 (Km−1,−2)

0.9322

0.8660

0.8485

0.8974

0.9051

The results of the analysis of Gm and Km phenotype heterogeneity in the Madonie, Sicilian and continental Italian populations are reported in Table 4. The comparisons are not all independent.
Table 4

Analysis of heterogeneity for the Gm and Km phenotypes. Madonie Alia + Cerda + Valledolmo, Sicily Alia + Cerda + Valledolmo + Palermo, Italy Ferrara (Vierucci 1965)

Gm system

d.f.

G test

(Williams correction)

P

 

Between Alia and Valledolmo

9

25.391

<0.01*

Madonie v. Palermo

9

12.333

>0.1

Between Alia and Palermo

9

9.816

>0.1

Between Valledolmo and Palermo

9

24.269

<0.01*

Within Sicily (Cerda excluded)

18

38.584

<0.01*

Between Sicily and Italy

9

21.800

<0.01*

Km system

d.f.

χ2 test

(Yates correction)

G test

(Williams correction)

P

Between Alia and Valledolmo

1

5.575

6.293

<0.02*

Madonie v. Palermo

1

0.021

0.087

>0.1

Between Alia and Palermo

1

0.751

1.073

>0.1

Between Valledolmo and Palermo

1

1.131

1.513

>0.1

Within Sicily (Cerda excluded)

2

6.407

6.307

<0.05*

Between Sicily and Italy

1

0.214

0.384

>0.1

The phenotypes considered for calculation of the Gm system heterogeneity are reported in Table 1. The phenotypes corresponding to haplotype Gm5*;1,17;... (refer to the presence of no. 5 in the Genotype column of Table 1) were pooled to obtain a matrix with a statistically acceptable number of observations. The χ2 test (even with the Yates correction) cannot be calculated for all the Gm phenotypes because of the still very low number of some observations. To avoid further pooling (with an inevitable loss of information about heterogeneity), we applied the G test with the Williams correction.

The difference in the Gm system between Sicily and continental Italy (Ferrara) is highly significant (P<0.01). In general, heterogeneity is very high in Sicily. We found a highly significant difference in genetic structure among the Alia, Valledolmo and Palermo groups. The difference also remains within the Madonie area, between Alia and Valledolmo. This is primarily due to the high frequency of phenotypes referable to the common African haplotype Gm5*;1,17;... in the Alia sample (Table 2), while Valledolmo presents Gm allotypes nearer to the continental populations. These towns have probably had different distributions of the Gm genes since their foundation. The χ2 and G tests also indicate a significant difference between the populations of Alia and Valledolmo for the Km system.

No significant differences in either the Gm or Km system were found between Alia and Palermo.

We did not record haplotype Gm5,*,26;1,3 reported by Walter (1997) in a previous study in Sicily (even at a frequency of up to 40%). Haplotype Gm5,*,26;1,3 is common in Asia but very rare in Mediterranean populations. The present study seems to exclude an important influence of Asian groups on the genetic structure of the population of west-central Sicily, contacts that would be very difficult to explain from the historical and geographical points of view.

Phylogenetic and multivariate analyses

The main human groups are characterized by specific Gm haplotypes. Hence, it is possible to reconstruct the biological history, origin and gene flows that have affected a population.

Unfortunately, information about the distribution of Gm haplotypes in the Mediterranean basin is still patchy and often incomplete. Moreover, the literature data are sometimes discordant. In particular, the G2m(23) allotype has not been tested in many studies, despite the fact that the presence or absence of this marker is fundamentally important in the study of Mediterranean populations, since it is strongly discriminating for the Gm5*;3 haplotype.

Considering all these limitations, we selected data available in the literature (Vierucci 1965; Ropartz et al. 1966, 1972; Durante and Ronchi 1967; Piazza et al. 1976; Lefranc et al 1976, 1979) and we supplemented them with unpublished data.

In Table 5, the results are compared with the frequencies of Gm haplotypes from other Italian populations (Emilia, Latium, Apulia, Sardinia), Greece and populations with Phoenician, Arab and African contributions to their gene pools (Algeria, Tunisia, Lebanon). Of these populations, the most interesting are the Lebanese and Tunisian, because they have Phoenician ancestry. The Gm10,11,13,15,16;1,17 haplotype (present at polymorphic frequencies in Palermo, Valledolmo, Sardinia, Tunisia, Algeria and Lebanon) might have been introduced by Phoenicians, as suggested by Lefranc and co-workers (1979).
Table 5

Mediterranean populations used for the phylogenetic and multivariate analyses. Data for the Lower Kabylie population are unpublished results (J.M. Dugoujon)

Populations

Number of individuals

21,28;1,17

21,28;1,2,17

5*;3

5*;1,17

10,11,13,15,16;1,17

5*;1,3

Others

References

Alia

144

15.43

04.57

74.14

05.52

00.00

00.00

00.00

-

Valledolmo

128

16.41

05.08

76.02

01.56

00.39

00.54

00.00

-

Palermo

94

15.84

03.84

75.00

03.19

00.53

00.00

01.60

-

Emilia (Ferrara)

86

18.80

03.60

77.60

00.00

00.00

00.00

00.00

Vierucci 1965

Apulia (Bari)

144

15.50

03.10

74.30

00.00

00.00

00.00

07.00

Ropartz et al. 1966

Latium

180

22.10

04.80

73.10

00.00

00.00

00.00

00.00

Durante and Ronchi 1967

Sardinia

1208

11.10

00.70

82.70

02.30

00.20

00.00

03.00

Piazza et al. 1976

Greece (Athens)

297

16.30

01.70

82.00

00.00

00.00

00.00

00.00

Ropartz et al. 1966

Algeria (Lower Kabylie)

124

24.20

00.80

52.80

16.00

01.60

00.00

03.80

-

Lebanon

160

17.76

02.24

72.49

03.65

03.36

00.00

00.50

Lefranc et al. 1976

Tunisia (Mahdia, Sfax)

378

23.40

03.80

61.00

09.20

00.60

00.00

02.00

Lefranc et al. 1979

We estimated the genetic relationships among these populations by constructing a dendrogram according to the maximum likelihood principle (Fig. 2). The robustness of this tree was tested by the bootstrap technique. The variables (Gm haplotype frequencies) were randomly resampled 1,000 times with replacement. The bootstrap values were higher than 50% for each branch of the tree. Any percentage higher than 50% was considered to give a “robust” splitting.
Fig. 2

Dendrogram obtained from Gm haplotype frequencies with the maximum likelihood method. The bootstrap values are higher than 50% for each branch of the tree

It is possible to identify two main groups. The first contains populations from Lebanon, Tunisia and Algeria, while the second consists of Greece and Italy (Emilia + Latium + Apulia). The Gm haplotype characterization is very peculiar in Sardinia, which occupies a separate branch of the dendrogram. Palermo, Alia and Valledolmo are intermediate between the populations with Phoenician, Arab and African contributions to their gene pools (Algeria, Tunisia, Lebanon) and the populations from Greece and continental Italy.

Figure 3 shows the first two axes of a principal components analysis based on the Gm haplotype frequencies in the Mediterranean populations listed in Table 5. The first two principal components explain 49.86% and 22.32% of the total genetic variance. Unfortunately, most of the populations were not tested for allotype G2m(23). In spite of the loss of this very discriminant marker, analysis of the Gm system is very useful for population genetic studies.
Fig. 3

Principal components analysis of Gm haplotype frequencies in various Mediterranean populations

Lebanon, Tunisia and Algeria present positive values for the first component, confirming that the African genetic traits in their gene pools are so marked and peculiar that they can be discriminated even with the first principal component. The cause of this difference could be related to the recent African origin of our species.

Alia and Palermo are very close to each other, while Valledolmo is very far from the Italian groups. Because of its special genetic peculiarities, Sardinia is also some distance from the Italian populations.

Interestingly, the principal components analysis indicates genetic affinities between Lebanon and Palermo. This is an important observation, perhaps due to genetic traces of the Phoenicians in western Sicily. They retained their hegemony until the Arab conquest in the 9th century a.d. Therefore, the Phoenicians, and later also the Arabs, likely played an important role in the genetic makeup of Sicily, since they settled in the region for quite a long time and probably mixed with the local people.

Discussion

In different eras, the need to find new commercial routes compelled Mediterranean peoples to embark on new explorations. These were favored and encouraged by the climatic and morphological uniformity of the Mediterranean coasts. Because of its central position in the Mediterranean Sea, Sicily has long been the meeting place of ancient civilizations and cultures. Hence, various populations have contributed to the very heterogeneous genetic structure of the island’s population, making Sicily very interesting from the anthropological point of view.

On the other hand, the differentiation of Sicilian communities could have been influenced by pathogenic selective pressure, which probably had a direct affect on the variability of the immunoglobulin constant domains (Gm phenotypes). After studying the distribution of Gm haplotypes in Sardinia, Piazza and co-workers (1976) concluded that the altitudinal variation in Sardinia was consistent with malarial selection of the Indo-European Gm 5*;3;23. In Lebanon and Tunisia, the difference in G2m(23) frequency between highlanders and lowlanders was highly significant and suggested a past adaptation to malaria (Lefranc et al. 1978, 1979).

Although Gm haplotypes are very useful in determining population affinities, the possibility of malarial selection of this locus could affect studies of population relationships (Schanfield et al. 2002). Therefore, a selective effect of malaria on the Gm5*;3;23 haplotype cannot be excluded in Mediterranean populations. Sicily, including the Madonie area, was once a zone of high malarial risk. Many genetic markers associated with high resistance to malaria (i.e., G6PD deficiency, haemoglobin defects) have been identified in all parts of the island. Several haemoglobin defects (Hb S, Hb G San José, thalassaemias) characteristic of southern Mediterranean areas historically affected by malaria have also been found in Alia, Cerda and Palermo (Cerutti et al. 2001).

The high frequency of the Gm5*;3;23 haplotype in the Madonie area and at Palermo suggests an adaptive response to malarial selection. Nevertheless, interpretation of the distribution of the Gm5*;3;23 haplotype is rather difficult. In some cases, populations with a high frequency of this haplotype belong to areas where malaria was endemic (i.e., Sicily, Sardinia, Emilia), while in others the malarial pattern was very variable (Latium, Algeria, Tunisia, etc.).

Moreover, STR data provide evidence that selection was probably not a major factor causing genetic heterogeneity in Sicily (Romano et al. 2003). Although it cannot be excluded that some of the STR alleles may be in linkage disequilibrium with selectively non-neutral coding mutations, migration and genetic drift seem to have played a more effective role.

The analysis of the Gm system has demonstrated that Palermo and the communities of the Madonie Mountains are genetically heterogeneous. The genetic structure is characterized by great intra- and inter-population variability.

In 1837, Sicily was struck by a cholera epidemic that affected many victims. It developed first at Palermo but soon became epidemic, spreading rapidly to several nearby towns, including Alia. The demographic trend at Alia was fairly constant until 1837, when there was a peak in mortality and a decrease in births. Although a surname analysis at Alia revealed an extremely low number of surnames, Alia presents great genetic heterogeneity. The bottleneck effect observed in the distribution of surnames is not reflected in an excess of homozygous individuals. During the cholera epidemic, about 30% of the surnames disappeared from Alia (Chiarelli et al. 2002), while in the following periods, new surnames were introduced, mainly from the surrounding areas.

The presence of a typical African marker (haplotype Gm5*;1,17;...), especially in the genetic structure of Alia and Palermo, highlights the possibility of past contacts with peoples from Africa. Indeed, Palermo was originally a Phoenician-Carthaginian colony (toward the end of the 2nd millennium b.c.) and remained so for a long time. The Phoenician colonization is also demonstrated by inscriptions found in the Gurfa Caves near Alia (Runfola 1978). Therefore, the introduction of an African polymorphism could have been due to the Phoenician colonization or to the more recent Arab conquest of the territory (9th century a.d.).

A study (Semino et al. 1989) carried out with restriction enzymes on mtDNA indicated the presence of African haplotypes (4.4%) in a sample of Sicilians. The authors hypothesized an input of genes from Africa to Sicily (estimated at about 10%) brought by Phoenician migrations. More recent studies on classical polymorphisms have demonstrated the Phoenician influence on the genetic structure of northwestern Sicily, which would have helped differentiate the local populations from the rest of the island (Piazza 1998; Romano et al. 2003).

The Gm haplotypes at Valledolmo are more similar in typology and frequency to those of continental populations (Table 5). The foundation of this village as Castrum Normandum (as shown by the ancient coat of arms) during the Norman rule (11th century a.d.) could explain the very peculiar genetic structure of the population with respect to other towns in the Madonie Mountains. The latter mainly have distant Arab roots related to the original Saracen villages (9th century a.d.), which only later became feudal towns. Valledolmo is genetically different, but whether this difference is due to gene flow from Normans or to genetic drift is difficult to tell.

Our results indicate that, despite the difficulties of access to the central mountainous portion of Sicily, the Madonie area was still affected by intense and diversified gene flow from Mediterranean and northern European populations, just like the coastal areas. Although it cannot be excluded that selection was a factor causing genetic heterogeneity in Sicily, migration and genetic drift seem to have played a more important role. However, the various contributions of foreign dominions to the genetic composition of Sicily’s current population must still be clarified, since anthropogenetic studies in Sicily, unlike Sardinia and Corsica, are still rather patchy.

Acknowledgements

We gratefully acknowledge the assistance of Prof. B. Chiarelli, Department of Animal Biology and Genetics, University of Florence (Italy), co-ordinator of the Trinacria Project. We are grateful to Dr. M.G. Andollina and the Thalassa Association of Palermo (Italy) for their help in collecting samples. We gratefully acknowledge the assistance of Dr. A. Sevin, Anthropology Centre, Toulouse (France), and Dr. A. Roggero, Department of Animal and Human Biology, University of Turin (Italy), for the data analysis. We are grateful to Prof. A. Piazza, Department of Genetics, Biology and Biochemistry, University of Turin (Italy), for the useful discussion of results and to Dr. P. Christie for the careful revision of the manuscript.

Copyright information

© Springer-Verlag 2004