Insect Molecular Biology (1997) 6(4), 377±383

Molecular identi®cationof sympatric chromosomal forms of Anopheles gambiaeand further evidence of their reproductive isolation

G. Favia,1 A. della Torre,1 M. Bagayoko,2 A. Lanfrancotti,1

N'F. Sagnon,3 Y. T. Toure 2 and M. Coluzzi1

1Istituto di Parassitologia, Fondazione Pasteur ± CenciBolognetti, UniversitaÁ 'La Sapienza', Rome, Italy;2Ecole Nationale de Me decine et de Pharmacie,Bamako, Mali; and 3Centre National de Luttecontre le Paludisme, Ouagadougou, Burkina Faso

Abstract

Three chromosomal forms of Anopheles gambiae s.s.,designated as Bamako, Mopti and Savanna, werestudied for diagnostic PCR assays based on the analy-sis of the X-linked ribosomal DNA (rDNA). The studywas performed on a 1.3 kb fragment containing part ofthe 28S coding region and part of the intergenic spacerregion. The ampli®ed material was cut with fourteenrestriction enzymes to detect Restriction FragmentLength Polymorphisms (RFLPs). The enzymes Tru9Iand HhaI produced patterns of DNA bands whichdi�erentiatedMopti fromSavanna and Bamako; more-over, a distinct `hybrid' pattern was recognized in theF1 female progeny from the cross of Mopti with eitherone of the other two chromosomal forms. The diagnos-tic signi®cance of the PCR-RFLP assay was veri®ed on203 karyotyped females from®eld samples collected intwo villages in Mali and one village in Burkina Faso.Agreement was observed between the chromosomaland the molecular identi®cations. No `hybrid' molecu-lar patternswere detected even among carriers of rareheterokaryotypes hypothetically produced by crossesbetween Mopti and Savanna. The results con®rmprevious observations indicating barriers to gene ¯owwithin An. gambiae s.s. and supporting the speci®cstatus of the taxonomic units proposed on cytogeneticground.

Keywords: Anopheles gambiae, chromosomal forms,PCR-RFLP, rDNA.

Introduction

The recognition of recent cryptic speciation has special

interest in Anopheles mosquitoes since the character-ization of the malaria vector system in terms of single

panmictic units is a basic requirement within epi-

demiological investigations and control programmes,

and it becomes even more important when pursuingthe development of antivector measures based on

genetic manipulations and population replacements

(Crampton et al., 1990; Coluzzi, 1992; Collins, 1994).One of themost challenging cases of cryptic speciation

refers to the Anopheles gambiae complex. This con-

sists of six well-established sibling species with addi-

tional intrinsic barriers to gene ¯owwithin the nominal

taxon, An. gambiae s.s., that is themost anthropophilicmember of the complex and themainmalaria vector in

Subsaharan Africa. Based on the lack of expected

intergradation between alternative arrangements of

polymorphic paracentric chromosome inversions,

West African An. gambiae s.s. was tentatively split

into ®ve taxonomic units, designated by non-Linnean

nomenclature as chromosomal forms Forest, Bissau,

Bamako, Savanna and Mopti (Bryan et al., 1982; ToureÂ

et al., 1983; Coluzzi et al., 1985; Petrarca et al., 1983,1986, 1987; Robert et al., 1989; Appawu et al., 1994;Toure et al., 1994, 1997).The Forest and Bissau chromosomal forms prevail

in humid and coastal areas of West Africa and could

both belong to a single potentially panmictic unit

widely distributed also in East Africa. The other three

taxa are more adapted to dry environments, are oftensympatric with evidence of assortative mating, and

overlap the distribution of An. arabiensis, the second

anthropophilic member of the An. gambiae complex.

The Bamako chromosomal form appears to have a

riverine ecology in the upper Niger river basin, from

SouthernMali toNorthernGuinea; it generally coexists

with Mopti and, less frequently, with Savanna. The

Mopti chromosomal form is the dominant taxon in theinner delta of theNiger river and inall irrigatedareas in

SouthernMali and eastwards in Burkina Faso up to the

middle Niger river basin. The Savanna chromosomal

Received 18 February 1997; accepted 30 May 1997. Correspondence:Professor Mario Coluzzi, Istituto di Parassitologia, UniversitaÁ 'La Sapienza',Piazzale Aldo Moro 5, 00185 Roma, Italy. e-mail: coluzzi@axrma.uniroma1.it.

# 1997 Blackwell Science Ltd 377

form has a wider distribution overlapping both the

Bamako and Mopti ranges. It is generally absent from

the sahelian zones and the inner part of large irrigationschemes wherever Mopti becomes dominant.

Savanna has also various contact zones with Bissau-

Forest populations and studies carried out in some of

these areas showed incomplete intergradation (Bryan

et al., 1982; Petrarca et al., 1983).We are speci®cally dealing here with the Bamako,

Mopti and Savanna chromosomal forms of An.gambiae s.s. These are characterized, according toToure et al. (1983), by di�erent 2R inverted arrange-

ments, namely jcu and jbcu for Bamako, bc and u for

Mopti, and b, cu and bcu for Savanna. The standard

arrangement (+) is common toSavanna andMopti. All

types of crosses and backcrosses were obtained in the

laboratory indicating the absence of postmating repro-

ductive barriers and the potential for introgression and

fusion between the chromosomal forms (Toure et al.,1983). Very low genetic distances and no discriminant

alleles were found by multilocus analysis based on

protein electrophoresis (Cianchi et al., 1983). The studyby gas chromatography showed distinct patterns of

cuticular hydrocarbons (Milligan et al., 1993).Various unclassi®ed low-frequency heterokaryo-

typeswere reported from ®eld samples and tentatively

interpreted as the product of sporadic crosses andbackcrosses involving di�erent chromosomal forms

(Toure et al., 1983). The only polytene rearrangementsnot recorded were those expected from crosses

between the Mopti and Bamako chromosomal forms.

Heterokaryotypes corresponding to hypothetical

Mopti6Savanna hybrids were repeatedly scored,

and even more frequent were those expected from

crosses between Savanna and Bamako. However, thehybrid origin of these heterokaryotypes has not been

proven and their possible classi®cation as rare chro-

mosomal polymorphisms cannot be discarded. This

interpretationwas actually the one favoured in a recent

analysis of chromosomal data from Mali (Toure et al.,1997). Consequently, none of the inverted karyotypes

would be exclusive to one taxon only, so that the

chromosomal forms can be diagnosed with high prob-

ability but not unequivocally.The lack of completely reliable diagnostic markers

for Bamako, Mopti and Savanna as well as for their

putative hybrids, makes it di�cult to progress in the

study of the degree of reproductive isolation existing

between these taxa which remain taxonomically unde-

®ned and incompletely understood in their bionomics.

A valuable approach to the problem was expected

through the development of DNA based diagnosticmethods. PCR assays based on random primers dis-

played di�erences between the chromosomal forms

(Favia et al., 1994a, b), but the low reproducibility of the

technique did not allow its implementation for a routine

assay. Following the lead of Collins and collaborators

(see the recent review by Collins & Paskewitz, 1996),

we decided to focus on ribosomal DNA (rDNA) whose

comparative study provided taxonomic tools forvarious cryptic taxa including the six main sibling

species of the Anopheles gambiae complex (Scott etal., 1993). We performed Restriction Fragment LengthPolymorphism (RFLP) analysis on a PCR-ampli®ed

fragment of 1.3 kb (Fig. 1) that includes the 3' end of

the 28S coding region and part of the adjacent inter-

genic spacer (IGS). This PCR-RFLP assay revealed

DNA bands which distinguish the Mopti chromosomalform from both Savanna and Bamako forms.

Results

Samples and sampling sites

The reference material for the study consisted of

karyotyped female specimens of An. gambiae s.l.collected in 1995 between 27 August and 14 September

in the villages of Banambani and Moribabougou near

Bamako (Mali), and on 5 September in the Goundry

village near Ouagadougou (Burkina Faso). More than

3000 specimens were preserved in Carnoy's ®xative

and part of them were later processed for ovarian

Figure 1. Restriction map of the rDNA region of Anopheles gambiae containing the PCR amplified 1.3 kb fragment. The open arrows identify theprimers used in the PCR amplification experiments. Single letters denote restriction endonuclease sites: T= Tru9I; H= HhaI. The sites indicatedby an asterisk are polymorphic and used for the identification of the chromosomal forms (see text for details). The second Tru9I site can not bemapped and the two alternative sites are indicated by dashed lines.

378 G. Favia et al.

# 1997 Blackwell Science Ltd, Insect Molecular Biology 6: 377±383

polytene chromosome complement analysis. More

than 900 half-gravid indoor-resting females (about 300

from each of the three villages) were scored and thecarriers of selected karyotypes were utilized for the

molecular study. Further samples were collected in

Mali in 7±10 October 1996 to establish various labora-

tory colonies from single female ovipositions. The 1996

collections were again from the villages of Banambani

and Moribabougou and from the village of Pimperena,

about 350 km south of Bamako in the Sikasso area.

Three of the sampling sites (Moribabougou, Banam-bani and Goundry) lie within the Sudan Savanna belt

characterized by 800±1000 mm of annual rainfall dis-

tributed from June to October. Pimperena is in the

south-Sudan Savanna zone with an annual rainfall

around 1200 mm. Moribabougou has more intense dry

season breeding since it is on the edge of the Niger

river, whereas only temporary rainy season streams

occur in the other localities. These have limited dryseason breeding associated to long-lasting ponds or

water reservoirs.

The frequencies of the species of the An. gambiaecomplex and their chromosomal variants have been

repeatedly studied at the sampling sites (Toure et al.,1983; Coluzzi et al., 1985; Petrarca et al., 1986; Merza-gora, unpublished) and these data have been con-

®rmed in the present investigation. The taxa recordedin Banambani are An. arabiensis and the chromo-

somal forms Bamako, Mopti and Savanna of An.gambiae, which generally coexist during the rainy

season with wide ¯uctuations in their relative frequen-

cies. The situation is rather di�erent in Moribabougou

where both An. arabiensis and An. gambiae Savannaare very rare and the prevailing taxa are Bamako and

Mopti, usually occurring in ratios near 1:1 in the rainyseason. The Goundry village is out of the range of

distribution of the Bamako chromosomal form and the

samples from this village show An. arabiensis sympa-tric with An. gambiae Mopti and Savanna. The latter

taxon increases its frequency in the second half of the

rainy season even if generally is not dominating.

Pimperena has an almost pure population of the

Savanna chromosomal form. Hybrid material wasobtained by crossing laboratory colonies of the three

chromosomal forms fromMali.

PCR-RFLP comparative analysis

We ampli®ed by PCR a fragment of 1.3 kb using the

primers A0 and Ag1.3 by Paskewitz & Collins (1990).

The primer A0 sequence is identical in An. gambiaes.s. and An. arabiensis whereas the Ag1.3 sequence isspeci®c for An.gambiae. The ampli®ed fragment con-

tains the 3'endof the 28Sandpart of the IGSsequences(Fig. 1). We used fourteen di�erent four-base recogni-

tion restriction enzymes to detect RFLPs in an analysis

of Banambani specimens bearing homokaryotypes

typical of the three chromosomal forms, namely 2Rb/bfor Savanna, 2Rbc/bc for Mopti and 2Rjcu/jcu for

Bamako. Two of these enzymes, HhaI and Tru9I,provided RFLP patterns which were found di�erent in

Mopti versus Savanna and Bamako due to two poly-

morphic sites for Tru9I and one for HhaI (Fig. 2).Laboratory crosses between Mopti and the other two

chromosomal forms produced F1 females showing a

distinct hybrid pattern whereas the F1 males showedthe pattern of the female parent (Figs 3 and 4), as

expected from the fact that the rDNA cluster is X-linked

(Collins et al., 1987). The diagnostic PCR-RFLP assay

was also successfully applied to all preimaginal

stages, including embryonated eggs and ®rst-stage

larvae.

Validation of PCR-RFLP di�erences

The diagnostic signi®cance of the above molecular

di�erences was validated in 203 ®eld-collected karyo-

typed females (70 from Banambani, 73 from Moriba-

bougou and 60 from Goundry) chosen to represent

each chromosomal form as well as some of the

putative hybrids. The karyotypes were included in the

analysis independently from their actual frequencies

so that some rare types are over-represented in thesubsample tested. As expected, no polytene comple-

ment was scored which could be interpreted as

Bamako6Mopti hybrid. A few karyotypes possibly

Figure 2. Diagnostic patterns for different chromosomal forms ofAnopheles gambiae obtained by a PCR-RFLP assay which cuts the1.3 kb rDNA amplified fragment with Tru9I and HhaI. Lanes 1 and 2 =Bamako; lanes 3 and 4 =Mopti; lanes 5 and 6 = Savanna;M=100 bp ladder.

Molecular identi®cation of An. gambiae chromosomal forms 379

# 1997 Blackwell Science Ltd, Insect Molecular Biology 6: 377±383

originating from crosses between Bamako and

Savanna were observed in Banambani but were not

included in the analysis because of the lack of aspeci®c diagnostic tool discriminating between these

two forms. The heterokaryotypes showing three con-

tiguous heterozygous loops (2R inversions b, c and u)

were only tested when the scoring of the pairing

con®guration allowed the distinction of the 2Rbc/u of

Mopti from the Savanna 2Rb/cu or 2Rbcu/+.

All karyotypes tested are listed in Table 1 with the

corresponding classi®cation in terms of chromosomalforms. The results of their PCR-RFLP classi®cation as

Mopti (MM PCR-RFLP pattern) or Bamako/Savanna

(SS PCR-RFLP pattern) are shown by village. The

karyotypes typical of Bamako (nos. 1 and 2), Mopti

(nos. 3, 4 and 6) and Savanna (nos. 9, 10, 11 and 12)

provided molecular results completely consistent with

the chromosomal identi®cation with only one excep-

tion recorded out of 106 cases tested. This refers to a2Ru/u homozygous specimen from Banambani which

showed the SS pattern. Karyotype 5 (2R homozygous

standard) is shared by both Savanna and Mopti chro-

mosomal forms and both MM and SS PCR-RFLP

patterns were recorded as expected. Heterokaryo-

types 7, 8 and 13 all involve the standard arrangement

and the identi®cation is tentativelymadeon thebasis of

the alternative arrangements. In the case of 2Rbc/+(no. 7) and 2Ru/+ (no. 8) heterokaryotypes, 24 speci-

mens out of 25 and 17 specimens out of 18, respec-

tively, showed the expected MM PCR-RFLP pattern,

whereas one specimen in each case showed the

alternative SS pattern. In the case of the 13 2Rb/+

heterokaryotypes analysed (no. 23), seven of them

showed the SS PCR-RFLP pattern whereas the other

six were classi®ed asMopti.The heterokaryotypes 14±18 are rare combinations,

which can be tentatively interpreted as the F1 or Fn

product of crosses between Mopti and Savanna. This

would apply particularly to the 2Rbc/b (no. 14) and 2Rb/

u (no. 15) heterokaryotypes, the latter being microsco-

pically indistinguishable from 2Rbu/+. However, none

of these heterokaryotypes showed the expected hybrid

MS PCR-RFLP pattern. Out of thirteen 2Rbc/b, twelvehad the Mopti pattern and one the SS; similarly, both

PCR-RFLP patterns were represented among the car-

riers of the 2Rb/u or 2Rbu/+ heterokaryotypes and

those scored as 2Rbu/b; the 2Rbc/bu and 2Rbu/u were

Mopti.

Discussion

The present results con®rm that the analysis of rDNA is

a valuable approach for the detection of cryptic Ano-pheles taxonomic units (Collins & Paskewitz, 1996).

Figure 3. PCR-RFLP assay performed with Tru9I on the F1 progenyfrom both reciprocal crosses between the Mopti and Bamakochromosomal forms of Anopheles gambiae. The F1 females show thespecific hybrid pattern whereas the males show the same pattern asthe female parent. M = 100 bp ladder.

Figure 4. PCR-RFLP assay performed with HhaI on the F1 progenyfrom both reciprocal crosses between the Mopti and Bamakochromosomal forms of Anopheles gambiae. The F1 females show thespecific hybrid pattern whereas the males show the same pattern asthe female parent. M= 100 bp ladder.

380 G. Favia et al.

# 1997 Blackwell Science Ltd, Insect Molecular Biology 6: 377±383

The PCR-RFLP patterns that distinguish the chromoso-

mal form Mopti from both Savanna and Bamako

appear to be a reliable guide to a very recent specia-

tion process previously detected only by the absence

of intergradation of polymorphic paracentric inver-

sions (Toure et al., 1983; Coluzzi et al., 1985). The caseof the 2Ru homokaryotype showing the Savanna PCR-

RFLP pattern is the only one apparently misleading.However, once the rare occurrence of 2Ru in the

Savanna population has been established (see the

molecular identi®cation of nos. 8, 15, and 16) the

corresponding homokaryotype would be extremely

rare but not unexpected. Moreover, the possibility of a

mistake during themanipulations in either the chromo-

somal or the molecular identi®cation can not be de®-

nitely ruled out.Thedi�erences in the rDNAmake it possible to score

the female hybrids thus representing a powerful tool to

evaluate the actual hybridization frequency.Moreover,

the molecular approach overcomes the limitations of

the cytotaxonomic identi®cation that is not 100% diag-

nostic and can be performed only to half-gravid

females and fourth-stage larvae.

The application of the PCR-RFLP diagnosis to karyo-typed females sampled in areaswhere theMopti taxon

is sympatric with Bamako and/or Savanna provided

data in agreement with the hypothesis of di�erent

reproductive units as indicated by the chromosomal

study. Rare unassigned heterokaryotypes, that could

be tentatively classi®ed as F1±Fn hybrid products, did

not show any hybrid PCR-RFLP pattern. Consequently,

they should be interpreted as low-frequency inversion

polymorphisms in either one of the taxa. The 2Rb

chromosomal arrangement, although characteristic of

the Savanna taxon, is also shared by Mopti, where itgives the bc/b, b/u and b/+ heterokaryotypes. The

reciprocal applies to the 2Rbc and 2Ru arrangements

typical of Mopti but ¯oating at low frequencies in the

Savanna population. The fact that not a single hybrid

PCR-RFLP pattern was detected in the analysis of the

®eld-collected material is particularly meaningful

when considering both the sympatric and synchronic

occurrence of the chromosomal forms in the studyareas and when taking into account that the molecular

identi®cation was speci®cally performed on speci-

mens bearing karyotypes of a possible hybrid origin

between Savanna andMopti.

The reproductive isolation between Bamako and

Mopti, already indicated by the chromosomal studies,

is con®rmed by the results of the present PCR-RFLP

method which additionally support isolation betweenSavanna and Mopti. It has to be stressed that these

convergent conclusions are obtained with totally inde-

pendent genetic markers: X-linked rDNA in one case

Table 1. Chromosomal and molecular identi®cations of female specimens of An. gambiae sampled in two villages in Mali (Banambani andMoribabougou) and one village in Burkina Faso (Goundry). The specimens are grouped according to their chromosome 2R karyotypes and thechromosomal form identi®cation currently utilized (BAM=Bamako, MOP=Mopti, SAV=Savanna, HYB=hybrid, ?=hypothetical). Themolecular identi®cation is presented by village. The PCR-RFLP pattern shared by Bamako and Savanna is designated as SS, the alternativeMopti pattern as MM. The double letter nomenclature was adopted to facilitate the designation of the possible hybrid pattern (SM) not recorded inthis material.

Identi®cation based on PCR-RFLP patternsIdenti®cation based onchromosome 2R scoring Moribabougou Banambani Goundry All villages

No. Karyotype Form Total SS MM SS MM SS MM SS MM

1 jcu/jcu BAM 11 8 ± 3 ± ± ± 11 ±2 jbcu/jcu BAM 9 3 ± 6 ± ± ± 9 ±3 bc/bc MOP 23 ± 8 ± 7 ± 8 ± 234 u/u MOP 14 ± 10 1 3 ± ± 1 135 +/+ MOP/SAV 11 ± 5 2 1 ± 3 2 96 bc/u MOP 13 ± 6 ± 1 ± 6 ± 137 bc/+ MOP? 25 1 9 ± 4 ± 11 1 248 u/+ MOP? 18 ± 9 1 2 ± 6 1 179 b/b SAV 24 2 ± 13 ± 9 ± 24 ±10 bcu/b SAV 3 ± ± 3 ± ± ± 3 ±11 b/cu SAV 7 ± ± 7 ± ± ± 7 ±12 bcu/cu SAV 2 ± ± 2 ± ± ± 2 ±13 b/+ SAV? 13 1 3 6 1 ± 2 7 614 bc/b HYB? 13 ± 7 1 1 ± 4 1 1215 b/u or bu/+ HYB? 6 ± 1 3 ± ± 2 3 316 bu/b HYB? 6 ± ± ± 1 1 4 1 517 bc/bu MOP? 4 ± ± ± 1 ± 3 ± 418 bu/u MOP? 1 ± ± ± ± ± 1 ± 1

Total 203 15 58 48 22 10 50 73 130

Molecular identi®cation of An. gambiae chromosomal forms 381

# 1997 Blackwell Science Ltd, Insect Molecular Biology 6: 377±383

and inversion on the second chromosome in the other.

Further application of the molecular identi®cation on

samples from other geographical areas should estab-lish the consistency of the positive assortative mating

and its taxonomic signi®cance. The available evidence

of reproductive isolation of the chromosomal form

Mopti versus both Bamako and Savanna supports thehypothesis of intrinsic premating barriers to gene ¯ow

that would ®t with a status of at least incipient species,

as suggested by Coluzzi et al. (1985). The lack of

di�erences between Bamako and Savanna could betentatively related to the hypothesis of their closer

relationship, but the present limited rDNA comparison

hardly allows evolutionary inferences. The ®nal sys-

tematic assessment of the taxonomic units shown

within the major Afrotropical malaria vector, An.gambiae s.s., needs further investigation to evaluate

the relationships between both Bamako and Savanna,

and between the dry savanna taxa and those describedfrom humid savanna and forest areas, namely Bissau

and Forest chromosomal forms.

Experimental procedures

Mosquito breeding and crossing

Adult mosquitoes were kept at T=26+18C and r.h.=75+5%

in cubic cages (25 cm), supplied with 1% sugar solution adlibitum and allowed to spontaneously mate for 4±6 days.

Females were blood-fed on immobilized guinea-pigs and

provided with cups half-®lled of water for egg laying. The eggs

were then transferred to a second room (T=28+ 18C and

r.h.=65+ 5%) and, after hatching, the larvae were kept to a

density of about 1 larva/3 cm3 inplastic basins (356 256 8 cm)

containing distilledwater toa depth of 4±5 cm, 0.1%marinesalt

and cat biscuits as food. After emergence adultswere collected

directly from thebasinswith electric aspirators and transferred

in new cages in the ®rst room. Both climatic rooms had a 12-h

photoperiod.

The mosquito colonies utilized for the crosses were estab-

lished from single ovipositions of An. gambiae s.s. collected inMali in October 1996 and classi®ed by their chromosomal

complement as follows: GABBMOR (Xag, 2Rjcu, 2La, 3R, 3L),

a Bamako colony from the village of Moribabougou;

GAMMMOR (Xag, 2Rbc/u, 2La, 3R, 3L), a Savanna colony from

the village of Moribabougou; GASSPIN (Xag, 2Rb, 2La, 3R, 3L),

a Mopti colony from the village of Pimperena. The so-called

`hybrids' were obtained by mixing in standard breeding cages

twenty-®ve specimens of GAMMMOR of one sex with the same

number of GABBMOR or GASSPIN of the other sex.

Polytene chromosomes

The ovaries of both laboratory and ®eld-collected mosquitoes

were ®xed in Carnoy's solution (3:1 ethanol:acetic acid) and

later processed for polytene chromosome preparations (della

Torre, 1997). Paracentric inversion karyotypes were scored by

phase-contrast microscopic examination and classi®ed

according to Coluzzi et al. (1979, 1985).

PCR primers

The primers are those described by Paskewitz & Collins (1990)

as A0 and Ag1.3. The former is a 20mer oligonucleotide

belonging to the 28S sequence. The latter, a 16mer, is located

in the IGS region, 1.3 kb downstream from the 3' end of the 28S

gene (Fig. 1).

DNA extraction and PCR ampli®cation

Genomic DNA was extracted from single mosquito as

described in Favia et al. (1994b). About 10 mg were used as

template for PCR ampli®cation performed as follows: 25ml totalreaction containing 50 pmol of each primer, 1 unit of Perkin

Elmer Taq polymerase, 200 mM of each deoxynucleotide, 2.5

mMMgCl2, 16 reaction bu�er provided from themanufacturer.

The reaction was cycled twenty-®ve times through the follow-

ing scheme: 30 s at 948C, 30 s at 588C, 30 s at 728C. Subse-quently, the samples were set at 728C for 7 min.

Restriction reaction

Once the band of the expected size (1.3 kb) had been identi®ed,

restriction analyses were carried out in order to detect speci®c

RFLPs. The following four-base recognition restriction

enzymes were tested on the ampli®cation product: Tru9I,HhaI, MspAI, MspI, RsaI, AluI, Hsp92I, Hsp92II, Sau3AI, HaeII,HaeIII, HpaII, TaqI, HinfI. The assay was based on 10 ml of theampli®edproduct added to 1.25U enzyme in 16 enzymebu�er

in 20 ml total reaction. The samples were restricted 3 h at 378Cwith the exception of Tru9I and TaqI that were incubated for 3 hat 658C. After digestion, samples were resolved on a 2%

MetaPhor-agarose TBE gel containing ethidium bromide 0.5

mg/ml.

Acknowledgements

The processing of the material for this study was

carried out at the Istituto di Parassitologia of the

University of Rome 'La Sapienza', and we acknowl-

edge the collaboration of M. Calzetta, G. Croce, S. Del

Signore, C. Fanello, D. Petrangeli and particularly G.

Petrangeli for his skilful assistance in polytene chro-mosome scoring. Field sampling was organized by the

De partement d'Epide miologie des A�ections Parasi-

taires of the Ecole Nationale de Me decine et de

Pharmacie of Bamako and by the Centre National de

Lutte contre le Paludisme of Ouagadougou, and we

thank the entomology teams of the two institutions. We

are grateful to N. Besansky, F. Collins, C. Louis, V.

Petrarca and J. R. Powell for reading and commentingthe manuscript. Financial support was provided by the

EU INCO project No. TS3*CT930236, by WHO CTD, and

by UNDP/World Bank/WHO Special Programme for

Research and Training in Tropical Diseases.

References

Appawu, M.A., Aba Ba�oe-Wilmot, Afari, E.A., Nkrumah, F.K. andPetrarca, V. (1994) Species composition and inversion poly-

382 G. Favia et al.

# 1997 Blackwell Science Ltd, Insect Molecular Biology 6: 377±383

morphism of the Anopheles gambiae in some sites of Ghana,West Africa. Acta Tropica 56: 15±23.

Bryan, J.H., Di Deco, M.A., Petrarca, V. and Coluzzi, M. (1982)Inversion polymorphism and incipient speciation in Anophelesgambiae s.str. in The Gambia, West Africa. Genetica 59: 167±176.

Cianchi, R., Villani, F., Toure , Y.T., Petrarca, V. and Bullini, L. (1983)Electrophoretic study of di�erent chromosomal forms withinAnopheles gambiae s.s. Parassitologia 25: 239±241.

Collins, F.H. (1994) Prospects for malaria control through thegenetic manipulation of its vectors. Parasitol Today 10: 370±371.

Collins, F.H., Mendez, M.A., Rasmussen, M.O., Meha�ey, P.C.,Besansky, N.J. and Finnerty, V. (1987) A ribosomal RNA geneprobe di�erentiatesmember species of the Anopheles gambiaecomplex. Am J Trop Med Hyg 37: 37±41.

Collins, F.H. and Paskewitz, S.M. (1996) A review of the use ofribosomal DNA (rDNA) to di�erentiate among cryptic Anophelesspecies. Insect Mol Biol 5: 1±9.

Coluzzi, M. (1992) Malaria vector analysis and control. ParasitolToday 8: 113±118.

Coluzzi, M., Petrarca, V. and Di Deco, M.A. (1985) Chromosomalinversion intergradation and incipient speciation in Anophelesgambiae. Boll Zool 52: 45±63.

Coluzzi, M., Sabatini, A., Petrarca, V. and Di Deco, M.A. (1979)Chromosomal di�erentiation and adaptation to human envir-onments in the Anopheles gambiae complex. Trans R SocTrop Med Hyg 73: 483±497.

Crampton, J.M., Morris, A., Lycett, G., Warren, A. and Eggleston, P.(1990) Transgenic mosquitoes: a future vector control strategy?Parasitol Today 6: 31±36.

della Torre, A. (1997) Polytene chromosome preparation fromanopheline mosquitos. Molecular Biology of Insect DiseaseVectors: a Methods Manual (Crampton, J.M., Beard, C.B. andLouis, C., eds), chap. 28, pp. 329±336. Chapman & Hall,London.

Favia, G., Dimopoulos, G., della Torre, A., Toure , Y.T., Coluzzi, M.and Louis, C. (1994a) Polymorphisms detected by random PCRdistinguish between di�erent chromosomal forms of Anophelesgambiae. Proc Natl Acad Sci USA 91: 10315±10319.

Favia, G., Dimopoulos, G. and Louis, C. (1994b) Analysis of theAnopheles gambiae genome using RAPD markers. InsectMolec Biol 3: 149±157.

Kumar, A. and Rai, K.S. (1990) Chromosomal localization and copynumber of 18S + 28S ribosomal RNA genes in evolutionarilydiverse mosquitoes (Diptera, Culicidae). Hereditas 113: 277±289.

Milligan, P.J., Phillips, A., Broom®eld, G., Molyneaux, D.H., Toure ,Y. and Coluzzi, M. (1993) A study of the use of gas chromato-graphy of cuticular hydrocarbons for identifying members of theAnopheles gambiae (Diptera: Culicidae) complex. Bull Ent Res83: 613±624.

Paskewitz, S.M. and Collins, F.H. (1990) Use of the polymerasechain reaction to identify mosquito species of the Anophelesgambiae complex. Med Vet Entomol 4: 367±373.

Petrarca, V., Carrara, G.C., Di Deco, M.A. and Petrangeli, G. (1983)Il complesso Anopheles gambiae in Guinea Bissau. Parassito-logia 25: 29±39.

Petrarca, V., Petrangeli, G., Rossi, P. and Sabatinelli, G. (1986)Etude chromosomique d'Anopheles gambiae et Anophelesarabiensis aÁ Ouagadougou (Burkina Faso) et dans quelquesvillages voisins. Parassitologia 28: 41±61.

Petrarca, V., Vercruysse, J. and Coluzzi, M. (1987) Observations onthe Anopheles gambiae complex in the Senegal River Basin,West Africa. Med Vet Entomol 1: 303±312.

Robert, V., Petrarca, V., Carnevale, P., Ovazza, L. and Coluzzi, M.(1989) Analyse cytoge ne tique du complexe Anopheles gambiaedans la re gion de Bobo-Dioulasso (Burkina Faso). Annls ParasitHum Comp 64: 290±311.

Scott, J.A., Brogdon, W.G. and Collins, F.H. (1993) Identi®cation ofsingle specimens of the Anopheles gambiae complex by thepolymerase chain reaction. Am J Trop Med Hyg 49: 520±529.

Toure , Y.T., Petrarca, V. and Coluzzi, M. (1983) Nuove entitaÁ delcomplesso Anopheles gambiae in Mali. Parassitologia 25: 367±370.

Toure , Y.T., Petrarca, V., Traore , S.F., Coulibaly, A., Maiga, H.M.,Sankare , O., Sow, M., Di Deco, M.A. and Coluzzi, M. (1994)Ecological genetic studies in the chromosomal form Mopti ofAnopheles gambiae s.str. in Mali, West Africa. Genetica 94:213±223.

Toure , Y.T., Petrarca, V., Traore , S.F., Coulibaly, A., Maiga, H.M.,Sankare , O., Sow, M., Di Deco, M.A. and Coluzzi, M. (1997)Distribution and inversion polymorphism of chromosomallyrecognised taxa of the Anopheles gambiae complex in Mali,West Africa. Parassitologia (in press).

Molecular identi®cation of An. gambiae chromosomal forms 383

# 1997 Blackwell Science Ltd, Insect Molecular Biology 6: 377±383