Differences in Extent of Genetic Introgression Between Sympatric Culex pipiens and Culex quinquefasciatus (Diptera: Culicidae) in California and South Africa

POPULATION BIOLOGY/GENETICS

Differences in Extent of Genetic Introgression Between Sympatric Culex pipiens and Culex quinquefasciatus (Diptera: Culicidae) in California and South Africa ANTHONY J. CORNEL, RORY D. MCABEE, JASON RASGON,1 MATTHEW A. STANICH, THOMAS W. SCOTT,1 AND MAUREEN COETZEE2 Department of Entomology, University of California at Davis, Mosquito Control Research Laboratory, Parlier, CA 93648

J. Med. Entomol. 40(1): 36Ð51 (2003)

ABSTRACT Comparisons of Þve morphological characters, 12 enzyme electrophoresis proÞles, and Wolbachia pipientis infection rates were used to characterize populations of members of the Culex pipiens L. complex in California and South Africa. In South Africa, male phallosome DV/D ratio, male maxillary palp index, branching of siphonal seta 1a, the enzyme locus Mdhp-1, and W. pipientis infection rates proved highly diagnostic for separating Culex quinquefasciatus from Cx. pipiens phenotypes. In Johannesburg, where sympatric members of the Cx. pipiens complex were analyzed as one population, a signiÞcant Wahlund Effect was observed in the enzyme loci such as Ao, 6-Pgdh, Mdh-2, and Pgm. In California, all populations of the Cx. pipiens complex were in Hardy Weinberg equilibrium at all polymorphic enzyme loci examined. Additionally, in California, all populations had similar W. pipientis infection rates and appeared morphologically identical (except for DV/D ratio, in extreme north and south). These Þndings indicate that in South Africa, Cx. pipiens and Cx. quinquefasciatus remain as genetically distinct populations and behave as separate species. Conversely, in California, there is considerable genetic introgression between Cx. pipiens and Cx. quinquefasciatus, and they behave as a single species.

DISCUSSIONS AND DEBATES concerning the taxonomic status of the members of the Culex pipiens complex have not resulted in a consensus. Opinions on the relationship of members of the complex range from regarding them as distinct species (Miles and Paterson 1979, Jupp 1978), subspecies (Mattingly 1967, Barr 1982), forms with some degree of reproductive isolation depending on location (Urbanelli et al. 1985, Urbanelli et al. 1997, Byrne and Nichols 1998), to physiological forms with considerable genetic introgression (Harbach et al. 1984, Tabachnik and Powell 1983). To date, Þve members of the complex have been identiÞed based on morphology, behavior (summarized in Table 1), and enzyme allele frequencies. Names such as comitatus, dipseticus, and pallens have been used to identify stable hybrid forms between quinquefasciatus and pipiens in North America and East Asia. Cx. p. pallens is still used in East Asian literature. To avoid giving impressions of assumptions about their taxonomic status, we use the terms quinquefasciatus and pipiens to distinguish populations and families on the basis of their male genitalia DV/D ratios. Pipiens, quinquefasciatus, and molestus are the 1 Department of Entomology, University of California at Davis, Davis, CA. 95616 2 Medical Entomology, Department of Clinical Microbiology and Infectious Diseases, School of Pathology of National Health Laboratory Service and University of the Witwatersrand, Johannesburg.

most common and widespread of the members of the complex. Enzyme electrophoresis (Urbanelli et al. 1997, Byrne and Nichols 1998) and molecular assays (Bourguet et al. 1998, Crabtree et al. 1997) have been developed that distinguish between the three common members, suggesting that there is reduced introgression between members of the complex in some regions. In addition to constituting a fascinating system for studying sibling species evolution, there are important public health implications for resolution of systematic relationships among members of the Cx. pipiens complex. Collectively, members of the complex represent the most important Culex species from a medical and veterinary standpoint because they are vectors of St. Louis encephalitis virus in eastern and south central North America (Tsai and Mitchell 1989), West Nile (WN) virus in northeastern United States (Lanciotti et al. 2000) and Europe (Hubalek and Halouzka 1999), Rift Valley fever virus (Meegan 1979), Wuchereria bancrofti (Farid et al. 2001), Dirofilaria immitis (Lai et al. 2000), and bird malarias such as Plasmodium relictum (Atkinson et al. 1995). Considering the recent introduction of WN virus into North America, the role that members of the Cx. pipiens complex play as zoonotic and epizootic vectors of WN virus in the northeastern United States is of primary importance for understanding WN virus ep-

0022-2585/03/0036Ð0051$04.00/0 䉷 2003 Entomological Society of America

January 2003 Table 1.

CORNEL ET AL.: Cx. pipiens COMPLEX IN CALIFORNIA AND SOUTH AFRICA

37

Summary of traits used to distinguish the members of the Cx. pipiens complex

Trait

quinquefasciatus

pipiens

Key reference

Sirivanakarn and White 1978

Harbach et al. 1985

Distribution

Worldwide in tropical and subtropical climates

All temperate regions except Australia (Holoarctic)

Swarming: Stenogamy ⫺ Eurygamy ⫺ Diapuase Variable Autogeny Variable Host preference Mammophilic (⫹⫹)

Breeding site Male genitalia

molestus Harbach et al. 1984 and Byrne and Nichols 1998 Hypogeous and Epigeous locations in N. hemisphere and Australia

⫺ ⫺ Heterodynamic ⫺ Mammophilic (⫹)

⫺ ⫺ Homodynamic ⫹ Mammophilic (⫹ ⫹)

Ornithophilic (⫹)

Ornithophilic (⫹⫹)

Ornithophilic (⫺)

Epigeous

Epigeous

australicus

glabocoxite

Dobrotworsky 1967 Dobrotworsky 1967 Australia

S. Australia

⫺ ⫺

NA NA Mammophilic except man (⫹) Ornithophilic (⫹⫹) Epigeous

Epigeous and Hypogeous DV/D ratio ⬎0.4 DV/D ratio ⬍0.2 DV/D ratio ⬍0.2 DV/D ratio 0.2Ð0.4 Dorsal arm sharp tipped Dorsal arm blunt ended Dorsal arm blunt ended

⫺ ⫺ Heterodynamic ⫺ NA

Epigeous Enlarged coxit

(⫹), Denotes behavior present. (⫹⫹), Denotes behavior more prevalent than (⫹) marked traits. (⫺), Behavior or trait not present.

idemiology in the New World. WN virus has been recovered from members of the Cx. pipiens complex in the northeastern United States (Lanciotti et al. 1999, Anderson et al. 2000), suggesting that these mosquitoes, as in Europe (Hubalek and Halouzka 1999), are primary vectors. In Africa (with possible exception in Egypt), Cx. pipiens complex members are not considered important WN virus vectors (Jupp and McIntosh 1970). Differences in isolation rates and vector competencies of Cx. pipiens complex mosquitoes (Varma 1960, Jupp and McIntosh 1970) for WN virus between continents provide further reasons for an investigation of the relatedness of members of the complex on a global scale. Insecticide resistance complicates the role of these mosquitoes in pathogen transmission because it limits options for vector control. Members of the complex have a notorious reputation for developing resistance to insecticides, including organophosphates, carbamates (reviewed in Hemingway and Karunaratne 1998), pyrethroids (Bisset et al. 1991, Ben Cheikh et al. 1998, Chandre et al. 1998, Kasai et al. 1998), and Bacillus sphaericus (Nielsen-Leroux et al. 1997, Yuan et al. 2000). Given their ability to develop insecticide resistance, the evolution and spread of resistance genes among members of the Cx. pipiens complex across continents have become a topic of considerable applied and basic scientiÞc interest (Raymond et al. 1991, Callaghan et al. 1998, Small et al. 1999). Different levels of reproductive isolation among members of the complex should be taken into account when interpreting the signiÞcance of resistance gene distribution, their alleles, and changing geographic patterns of these traits. Another interesting issue is the unidirectional cytoplasmic incompatibility characteristics of cytotypes of the endosymbiont bacteria Wolbachia among members of the Cx. pipiens complex (Hertig and Wolbach

1924, Guillemaud et al. 1997). The existence of different Wolbachia cytotypes may be exploited for driving selected genes that are intended to interfere with or prevent pathogen transmission through Cx. pipiens complex populations (Handler and James 2000, Sinkins and OÕNeill 2000). Moreover, population differences in Wolbachia infection rate strain composition, or compatibilities between sympatric members of the Cx. pipiens complex are potentially important indicators of population structure. Two members of the complex, pipiens and quinquefasciatus, are known to occur in South Africa (SA) and are typically classiÞed as separate species. Jupp (1978) reported that both larvae and adult males of quinquefasciatus and pipiens from Johannesburg were different enough morphologically to warrant a clear distinction between the two forms, with very few intermediates being found in areas of sympatry. In North America, the situation differs as hybrids of pipiens and quinquefasciatus (based on DV/D ratio) occur between 36 and 39⬚N latitude. North of 39⬚N latitude DV/D ratios fall within measurements typical for pipiens, and south of 36⬚N DV/D ratios fall within measurements typical for quinquefasciatus (reviewed in Barr 1982). Enzyme studies performed on North American pipiens and quinquefasciatus populations corroborate the morphological Þndings of the latitudinal cline from pipiens to hybrids to quinquefasciatus (Pryor et al. 1980 a, b, Cheng et al. 1982, Tabachnik and Powell 1983, Urbanelli et al. 1997). The research described in this work is part of a larger research program to deÞne the global taxonomic and reproductive relationships among mosquitoes in the Cx. pipiens complex. Current contention, as stipulated by Barr (1982), is that the evidence presented by Jupp (1978) that pipiens and quinquefasciatus behave as separate species in SA is too meager. Therefore, this study was undertaken to conÞrm

38

JOURNAL OF MEDICAL ENTOMOLOGY Table 2.

Locations, date, and method of capture of Cx. pipiens complex members from California and South Africa

Code 1 2 3 4 5 6 7 8 9 10 a

Vol. 40, no. 1

Location a

Redding (CA) 40⬚45⬘N 122⬚5⬘W Fresno (CA) 36⬚50⬘N 120⬚W Centerville (CA) 36⬚50⬘N 119⬚50⬘W Reedley (CA) 36⬚40⬘N 119⬚50⬘W Chino (CA) 34⬚04⬘N 117⬚35⬘W Skukuza (SA) 25⬚01⬘S 31⬚35⬘E Johannesburg (SA) 26⬚06⬘S 27⬚50⬘E Johannesburg (SA) 26⬚06⬘S 28⬚04⬘E Johannesburg (SA) 26⬚06⬘S 28⬚04⬘E Boane (Mozambique) 26⬚06⬘S 32⬚42⬘E

Method of collection

Month collected

CDC trap Unkempt swimming pool Kings River seepage pool Horse ranch, drinking trough Dairy lagoon CO2 and goat-baited net traps and biting humans Chicken coop Resting outdoors at lights Bird-baited lard cans Resting indoors

July 2000 May 2000 June 2000 August 2000 September 2000 February 1989 March 2000 JanuaryÐMarch 1989 February 1989 March 2000

Degrees latitude.

JuppÕs conclusion using enzymes that distinguish pipiens from quinquefasciatus and Wolbachia infection patterns in addition to the morphological characters used by Jupp (1978) that separate pipiens and quinquefasciatus. The other purpose of this study was to use these same methods of investigation to show genetic introgression between pipiens and quinquefasciatus in California as a comparison. We chose California and SA as our study sites because they appeared to represent two extremes of the spectrum, from considerable introgression in California to, based on morphological and behavioral traits (Jupp 1978), little introgression in SA. Materials and Methods Mosquito Rearing and Collections. Females or larvae were collected from various locations in California and SA (Table 2). Three larval collections were made in Fresno county, two along the same latitudinal transect and one a little south (⬃9 miles), to examine the possibility of population structural differences in the hybrid zone between urban (dense housing community within city boundary), semiurban (in small holdings bordering city boundary), and rural areas (over 20 miles from nearest town or city) that experience similar climatic conditions. In the urban site, larvae were collected from a backyard swimming pool situated among a dense cluster of homes close to the center of Fresno. The rural site was ⬇15 miles east of Fresno, where there is a low density of housing and thick riverine vegetation surrounded by cattle and horse ranches. Larvae collected from a drinking trough on a horse ranch directly on the outskirts of Reedley (small town 20 miles southeast of Fresno) served as specimens from the semiurban site. Adult collections from Johannesburg originated from urban (within city limits) and rural (20 km out of town among small farm holds) locations. Collections from Skukuza, Kruger National Park, were made from CO2 and goat-baited net traps and from human night-biting catches in Skukuza village. Wild females were allowed to oviposit, and whenever possible each egg raft was reared separately so that the morphology, allozymes, and Wolbachia from isofemale lines could be traced to speciÞc mothers. Larvae were reared to adults and used for further analysis.

Morphological Analysis. When larvae were collected (collections 2, 3, 4, and 5), 40 Ð50 fourth instar specimens were mounted on slides in Hoyers medium, according to the method described in Belkin (1962) for morphological analysis. The rest of the immature mosquitoes were allowed to develop into adults. A sample (ranging from 49 to 94) of adult males, 24 h after emergence, was dissected and mounted for morphological examination. The remaining adults were frozen at ⫺80⬚C for enzyme and Wolbachia assays. From collections of wild-caught females (1, 6, 7, 8, 9, and 10), morphological analyses were performed on the offspring (F1 generation) reared from egg rafts. In all of the SA collections, egg rafts were reared as individual families. From each family, morphological index values were measured from Þve larvae and two adult males. To be most conservative, the value that was closest to the pipiens/quinquefasciatus cutoff points was taken to be representative of that family. In all instances, the methodology for morphological analyses followed that described by Jupp (1978). Measurements were taken with a graticule in the eyepiece of a Zeiss Axioskop compound microscope at either 100 or 400⫻ magniÞcation. SigniÞcance testing of DV/D ratios was performed using StatView software and analysis of variance Scheffe´ F test (SAS Institute 1998). The following characters were measured: larval siphonal index, larval siphonal setae, cross vein index, male maxillary palp index, and DV/D index of male genitalia. Spearman Rank Correlation CoefÞcient (rs) (Fowler et al. 1998) was used to measure levels of correlation between each of the four morphological characters to DV/d ratio. Enzyme Electrophoresis. Native 6% polyacrylamide gel electrophoresis was performed on single male and female specimens (Hoefer Electrophoresis Unit SE600 Series; Hoefer Pharmacia, Biotech, San Francisco, CA) following methodologies described by Black and Munstermann (1996). Each mosquito was homogenized with a pestle in 30 ␮l of loading buffer (20% sucrose, Triton X-100 [0.5%], Tris-citrate, pH 7.0, buffer, and a trace of bromphenol blue as a tracking dye) in a 1.5-ml Eppendorf tube. Numerous electrophoresis gel and staining buffers and staining recipes were evaluated for each enzyme system to determine the methodology that resulted in the best band resolution and consistency. It was crucial that Mdh and Mdhp were run in Tris-maleate-EDTA and

January 2003 Table 3. complex

CORNEL ET AL.: Cx. pipiens COMPLEX IN CALIFORNIA AND SOUTH AFRICA

39

Electrophoresis buffers and references used for isoenzyme-staining procedures for members of the Cx. pipiens mosquito

Proteins

E.C. no.

Aspartate aminotransferase

2.6.1.1

Aldehyde oxidase Diaphorase

1.2.3.1 1.6.2.2

B-Phosphogluconate dehydrogenase Hexokinase

1.1.1.43 2.7.1.1

Isocitrate dehydrogenase

1.1.1.42

Malate dehydrogenase Malate dehydrogenase (NADPⴱ)

1.1.1.37 1.1.1.40

Phosphoglucoisomerase Phosphoglucomutase Xanthine dehydrogenase

5.3.1.9 5.4.2.2 1.1.1.204

Encoding loci Aat-1 Aat-2 Ao Dia-1 Dia-2 6-Pgah Hk-1 Hk-2 Hk-3 Idh-1 Idh-2 Mdh Mdhp-1 Mdhp-2 Pgi Pgm Xdh

Buffer systema

Staining references

1 1 2

Manchenko 1994

1 2 2 2 2 2 3 3 3 1 1 1

Cheng and Hacker 1976 Shaw and Prasad 1970 Shaw and Prasad 1970 Shaw and Prasad 1970 Ayala and Powell 1972

Steiner and Joslyn 1979

Shaw and Prasad 1970 Ayala and Powell 1972 Shaw and Prasad 1970 Cheng and Hacker 1976 Steiner and Joslyn 1979

a Buffer systems were 1, continuous Tris-borate-EDTA, pII 8.9 (Black and Munstermann 1996): 2, continuous Tris-citrate, pH 7.1 (Black and Munstermann 1996): 3, continuous Tris-maleate-EDTA, pII 7.6 (Brewer and Singh 1970).

Idh and Ao in Tris-citrate buffers to produce consistent resolution. The other enzymes could be run in either Tris-borate-EDTA or Tris-citrate buffer. A total of 11 enzyme systems potentially coding for 16 loci were analyzed. Electrophoretic and staining procedures are summarized in Table 3. Enzymes are numbered in order of decreasing mobility from the origin. Alleles were assigned relative mobility values (Rf) relative to the most common allele (100). Observed and expected allele frequencies and FIS calculations were performed for each enzyme locus and population using GenePop software (Raymond and Rousset 1995). ␹2 tests with Yates correction factor were performed to calculate signiÞcance of deviations from Hardy Weinberg equilibrium (Sokal and Rohlf 1981). Very rare alleles (expected frequency ⬍0.02 in the homozygous state) were not included in ␹2 calculation. Wolbachia Infection. DNA was extracted according to methods described by Black and DuTeau (1997). Two to four individuals from each Cx. pipiens complex SA family and ⱖ50 wild individuals each from Anderson (20 miles south of Redding), Fresno, and Los Angeles in California were tested for Wolbachia infection using two separate Wolbachia-speciÞc polymerase chain reaction (PCR) assays. One assay ampliÞed a ⬇900-bp fragment from Wolbachia 16S rRNAencoding DNA (rDNA) (forward, 5⬘-TTG TAG CCT ATG GTA TAA CT-3⬘; reverse, 5⬘-GAA TAG GTA TGA TTT TCA TGT-3⬘) (OÕNeill et al. 1992). The other assay ampliÞed a 550-bp fragment from the Wolbachia surface protein (wsp) gene (Zhou et al. 1998) (forward, 5⬘-TGG TCC AAT AAG TGA TGA AGA AAC-3⬘; reverse, 5⬘-AAA AAT TAA ACG CTA CTC CA-3⬘). DNA template quality was tested for all specimens by successful ampliÞcation of a 400-bp fragment from insect mitochondrial 12S rDNA (forward, 5⬘CTA GGA TTA GAT ACC CTA TT-3⬘; reverse, 5⬘AAG AGC GAC GGG CGA TG-3⬘) (OÕNeill et al.

1993). Known Wolbachia-infected specimens and antibiotic-cured specimens were included in all reactions as positive and negative controls, respectively. A sample containing deionized water instead of template DNA was included in all reactions as a negative control. Ready-to-go PCR beads (Amersham-Pharmacia Biotech, Piscataway, NJ) were used for PCR reactions to ensure standardized ampliÞcation. Each 25-␮l reaction contained: 500 nM of each Wolbachia 16S, wsp or 12S mtDNA primer, 1 ␮l template DNA, and one PCR bead. PCR conditions were as follows. Wolbachia 16S rDNA, 95⬚C for 5 min, followed by 35 cycles of 95⬚C/1 min, 54⬚C/1 min, 72⬚C/1 min; Wolbachia wsp gene, 95⬚C for 5 min, followed by 40 cycles of 94⬚C/ 1 min, 55⬚C/1 min, 72⬚C/1 min; insect mitochondrial 12rDNA, 94⬚C for 5 min, followed by 30 cycles of 94⬚C/1 min, 55⬚C/1 min, 72⬚C/1 min. After ampliÞcation, reactions were held at 72⬚C for 10 min and stored at 4⬚C. PCR products were visualized by electrophoresis in a 1% agarose gel, stained with ethidium bromide, and photographed under UV light. Results Morphology. Ranges of DV/D ratios, male maxillary palp, siphonal and cross vein indexes, and siphonal setae branching for each collection are provided in Table 4. Histograms of DV/D ratios from the six Californian locations and three southern African sites are provided in Fig. 1. All of our collections from Johannesburg, regardless of method of collection (Table 2), consisted of both pipiens and quinquefasciatus, in which mostly pipiens were collected resting inside chicken coops (84%) and only slightly more quinquefasciatus were attracted to lights (56%). All offspring reared from females collected in Skukuza fell within DV/D ratios typical for pipiens, and all those from

40

JOURNAL OF MEDICAL ENTOMOLOGY

Vol. 40, no. 1

Table 4. Summary of morphological measurements obtained for various Cx. pipiens complex populations in California and South Africa Population

Codea

pipiensb DV/D ratio ⬍0.2

Johannesburg 1978 Johannesburg 1989 Skukuza 1989 Boane 2000 Johannesburg 2000 Redding 2000 Fresno 2000 Reedley Chino dairy 2000

Jupp 1978 8,9 6 10 7 1 2.3 4 5

(32)c (20) (20)

Johannesburg 1978 Johannesburg 1989 Skukuza 1989 Boane 2000 Johannesburg 2000 Redding 2000 Fresno 2000 Reedley 2000 Chino dairy 2000

Jupp 1978 8,9 6 10 7 1 2.3 4 5

2.3Ð1 (50) ND ND

Johannesburg 1978 Johannesburg 1989 Skukuza 1989 Boane 2000 Johannesburg 2000 Redding 2000 Fresno 2000 Reedley 2000 Chino dairy 2000

Jupp 1978 8.9 6 10 7 1 2.3 4 5

1.3Ð1.3 (50) 1Ð4, 1Ð3 (20) 2Ð3, 1Ð5 (20)

Johannesburg 1975 Johannesburg 1989 Skukuza 1989 Boane 2000 Johannesburg 2000 Redding 2000 Fresno 2000 Reedley 2000 Chino dairy 2000

Jupp 1978 8,9 6 10 7 1 2.3 4 5

4.8Ð6.1 (48) 3.6Ð6 (20) 3.6Ð4.53 (20)

Johannesburg 1978 Johannesburg 1989

Jupp 1978 8.9

Intermediatesb DV/D ratio 0.2Ð 0.4

Quinquefasciatusb DV/D ratio ⬎0.4

Male genitalia index

Skukuza 1989 Boane 2000 Johannesburg 2000 Redding 2000 Fresno 2000 Reedley 2000 Chino dairy 2000 a b c

6 10 7 1 2.3 4 5

(31) (27) (13) (5)

(19) (50) (42) (5)

(42) (23) Male maxillary palp index

(10) (21) (50) 3.2Ð8.7 (50) ND

2.25Ð3.1 (19) 2.5Ð5.5 (50) 2.72Ð7.75 (42) 3.75Ð4.7 (5)

3.2Ð7.5 (13) 3.5Ð7.5 (5) 2.53Ð7 (42) 3Ð4.84 (23) Siphenal Sets 1aS 1bS

2Ð3, 2Ð4 (19) 2Ð5, 2Ð5 (50)

4.11; 4Ð10 (50) 4.11; 3Ð12 (27) 6Ð9; 6Ð9 (13) 4Ð8; 4Ð10 (5) 2Ð6, 3Ð6 (50) 2.5Ð2.6 (41) Siphonal index

3.53Ð5.1 (19) 2.5Ð3.68 (50)

2.7Ð4.84 (10) Ð4.84 (21) 3.6Ð8.9 (50)

4.3Ð4.4 (2)

4.8; 4.9 (50) 3.5Ð4.7 (50) 2.08Ð3.77 (27) 2.36Ð3.31 (13) 2.78Ð3.31 (13)

3.1Ð6.9 (50) 4Ð6.9 (41) Cross vein index of mass

1.4Ð5.0 1.4Ð4.3 (17) 1.1Ð14 (3) 1Ð1.88 (17) 1.3 (3)

3.1Ð4.81 (50) 0.7Ð1.3 0.75Ð3.3 (24) 1.3Ð2 (3) 1.3Ð2 (3)

1.1Ð2.5 (19) 1.07Ð1.25 (5) 1.07Ð2.9 (50) 0.75Ð4.3 (42) 0.85Ð1.1 (5)

0.75Ð1 (13) 6.86Ð1 (5) 0.9Ð1.58 (42) 0.81Ð1.88 (23)

0.9Ð1.6 (10) 0.85Ð1.38 (21) 1.07Ð1.45 (50)

Population codes correspond to those in Tables 2 and 5. Range of morphological values categorized according to DV/D ratio indices within each population. Numbers in parentheses refer to numbers of individuals or families evaluated that fall into that range.

Boane fell within the range typical for quinquefasciatus. No males with a DV/D ratio intermediate between pipiens and quinquefasciatus (0.2Ð 0.4) were reared from females collected from the three sites in southern Africa. Spearman Rank Correlation CoefÞcient showing perfect correlation (rs ⫽ 1) to DV/D ratio was obtained with all Þve morphological characters in Boane. The same was observed with 1aS and maxillary

palp index in the Johannesburg population. Although still highly correlated to DV/D ratio at the 99% conÞdence level, 1bS (rs ⫽ 0.93), siphonal index (rs ⫽ 0.93), and cross vein index (rs ⫽ 0.70) showed some overlap in values because of a few outlying individuals. The Þve morphological characters also showed significant (99% conÞdence) correlation to DV/D ratio in the Redding and Chino populations, although much

January 2003

CORNEL ET AL.: Cx. pipiens COMPLEX IN CALIFORNIA AND SOUTH AFRICA

41

Fig. 1. Collecting sites and DV/D ratio histograms of Cx. pipiens complex from California and SA. White histogram bars, pipiens; grey, intermediates (hybrids); and black, quinquefasciatus DV/D ratio differentiating thresholds.

42

JOURNAL OF MEDICAL ENTOMOLOGY

more overlap was observed in some characters, as reßected by a lower Spearman Rank Correlation CoefÞcient. The rs values for these populations were as follows: Shasta 1aS ⫽ 0.69, 1bS ⫽ 0.76, siphonal index ⫽ 1, cross vein index ⫽ 0.63, maxillary palp ⫽ 0.79, Chino 1aS ⫽ 1, 1bS ⫽ 1, siphonal index ⫽ 0.65, cross vein index ⫽ 0.71, and maxillary palp ⫽ 1. In this study, we measured cross vein indexes of males only, as females were frozen for electrophoresis, although Jupp (1978) found overlap in cross vein indexes of both males and females. Approximately 25% of the females originating from Johannesburg 1989 and 2000 collections produced male offspring that had cross vein indexes falling within the ranges of quinquefasciatus families. Additionally, 15% (3 of 20) of the Skukuza 1989 pipiens population had a value of 1.3 that just falls within the quinquefasciatus range. Phallosome DV/D ratios of California collections from Redding were typically pipiens, those in the south from Chino were typically quinquefasciatus, and those from central California in Fresno and surrounding areas included a range from pipiens to intermediate to quinquefasciatus phenotypes (Fig. 1 and Table 4). Ranges of DV/D ratio from the two sites along the same degrees of latitude in Fresno County (collections 2 and 3) did not differ signiÞcantly from each other at the 95% level of conÞdence. However, the range of DV/D ratios from the Reedley site (collection 4) was signiÞcantly (at 95% level) more skewed toward quinquefasciatus-like ratios than the other two Fresno county sites. Because there was no signiÞcant difference in the DV/D ratios in the Fresno and Centerville sites, mosquitoes from both sites were used together to represent a single row of morphological data designated as Fresno in Table 4. None of the other four morphological characters showed a cline of values from north to south that paralleled that of the DV/D ratios in California (Table 4). Because both the extreme northern and southern sites (that had nonoverlapping DV/D ratios) produced some overlapping ranges of values for the other morphological features, we did not follow through with efforts to rear larvae separately to make measurements on the emerged adult male and their corresponding larval exuviae. Enzymes. All enzyme systems evaluated were polymorphic except for Dia-1, Dia-2, Idh-2, and the cathodal Mdh-1 locus. A single Xdh locus was excluded because of poor band resolution and lack of conÞdence in scoring despite being polymorphic in both California and SA. Allele frequencies, Yates-corrected ␹2 statistics, and FIS values are represented in Table 5. Below follows a synopsis of each of the enzymes evaluated, giving a description of our results and correlations to other published results. Images of portions of gels showing the isoenzyme proÞle phenotypes are provided in Fig. 2. Aat-1. We identiÞed Þve alleles at this locus (Aat-180, Aat-190, Aat-195, Aat-1100, and Aat-1110), which were three more than were observed by others in California and Africa (Urbanelli et al. 1985, Urbanelli et al. 1995, Urbanelli et al. 1997). The two

Vol. 40, no. 1

slowest alleles, Aat-180 and Aat-190, required 6 h of electrophoresis for Aat-190 to migrate out of the wells. The existence of a slower Aat-180 made interpretative sense, and was postulated when no banding (homozygous for Aat-180) and when weak banding of allele Aat-190 (Aat-180/Aat-190 heterozygote) and a strong intermediate band were observed (Fig. 2). Existence of Aat-180 was conÞrmed by horizontal electrophoresis and staining of both anodal and cathodally ran gels. Aat-180 migrated a short distance from the wells anodally. By comparing allele frequencies between our data and others, it appears that what we have scored as Aat-180 and Aat-1100 is equivalent to what others have scored as Aat-290 or Got-294 and Aat-2100 or Got-2100, respectively. Our results will then be in agreement with observation of Urbanelli et al. (1997) that Aat-290 is more frequent in quinquefasciatus DV/D ratio phenotypes in California than in the north, where pipiens DV/D phenotype and Aat-2100 predominate. A fast migrating allele (Aat-1110) was observed in populations from Fresno and in the F2 generation individuals from Redding, but not in samples from southern California and SA. In Johannesburg, Aat-1100 was the only allele observed in pipiens families, whereas Aat-180, Aat-190, Aat-195, and Aat-1100 were observed in quinquefasciatus families. Aat-2. We observed four alleles at this locus: Aat-290, Aat-2100, Aat-2105, and Aat-2110. Other authors have reported three alleles (Urbanelli et al. 1997). Aat-2105, which was not previously reported, was uncommon in both California and SA. All of the other alleles occur in both California and SA and show no differentiation between pipiens and quinquefasciatus. Ao. Three alleles were observed at the Ao locus (Ao100, Ao105, and Ao110) in all populations from California and in SA. A fourth allele (Ao95) occurred in one individual from Centerville. In Johannesburg, the frequencies of alleles differed between pipiens and quinquefasciatus. Ao100 was more common in pipiens (0.60) than in quinquefasciatus (0.23), and Ao105 and Ao110 were more frequent in quinquefasciatus than in pipiens. 6-Pgdh. Gels stained for both Pgm (faster migrating with double-band appearance) and 6-Pgdh (slower lighter staining bands) resembled those obtained by Cheng and Hacker (1976). Two alleles, 6-Pgdh90 and 6-Pgdh100, were found in both Californian and SA populations. Others observed three alleles in California and Africa, although the third and fastest migrating allele (6-Pgdh104), observed by Urbanelli et al. (1997), was uncommon in all populations they examined in California, ⬍10% in African quinquefasciatus (Urbanelli et al. 1985), and ⬍3% in Madagascar (Urbanelli et al. 1995). Hk. Our three-banded homozygous and six-banded heterozygous staining pattern for Hk was identical to that observed by Cheng and Hacker (1979) and Pryor et al. (1980b). Genetic analyses by these authors conclude that a pair of codominant alleles at a single locus controls the variants observed of the Hk enzyme zymograms. Hence, we have treated scoring the Hk variants as representing different alleles at one locus

January 2003

CORNEL ET AL.: Cx. pipiens COMPLEX IN CALIFORNIA AND SOUTH AFRICA

43

Table 5. Allele frequencies, FIS values, and Yates corrected ␹2 values for deviation from expected Hardy Weinberg equilibrium of polymorphic enzyme loci for populations of Cx. pipiens complex mosquitoes sampled in California and South Africa Loci Aat-1 80 90 95 100 110 n FIS ␹2 Aat-2 90 100 105 110 n FIS ␹2 Ao 95 100 105 110 n FIS ␹2 6-Pgdh 90 100 n FIS ␹2 Hk 90 100 110 n FIS ␹2 Idh-1 95 100 n FIS ␹2 Idh-2 100 n FIS ␹2 Mdhp-1 90 100 n FIS ␹2 Mdhp-2 90 100 n FIS ␹2 Mdh-2 90 100 110 n FIS ␹2 Pgi 90 100 n FIS ␹2 Pgm 90 100 110 n FIS ␹2 a

LOCATIONa 1

2

3

4

5

7

10

N/D N/D N/D N/D N/D -

0.196 0.022 0 0.674 0.109 48 ⫺0.219 3.56

N/D N/D N/D N/D N/D -

N/D N/D N/D N/D N/D -

0.417 0.125 0 0.458 0 24 0.188 1.74

0.024 0.071 0.012 0.893 0 42 0.045 3.54

0.194 0.556 0 0.25 0 18 ⫺0.195 1.64

0.019 0.231 0 0.75 52 0.008 0.02

0.071 0.898 0.02 0.01 49 ⫺0.075 3.35

N/D N/D N/D N/D N/D -

0 0.932 0 0.067 52 ⫺0.063 0.43

0 0.021 0 0.979 48 ⫺0.011 0.005

0.027 0.909 0.027 0.036 55 0.157 14.35

0 0.95 0 0.05 20 ⫺0.027 0.01

0.948 0.042 0.01 48 ⫺0.035 8.53

0.02 0.867 0.071 0.041 48 0.162 4.07

0.836 0.029 0.135 52 ⫺0.153 8.09

0.521 0.125 0.354 48 0.089 1.77

0.685 0.191 0.125 84 0.582 48.77a

0.05 0.4 0.55 20 0.642 18.54

0.106 0.964 48 0.115 0.032

N/D N/D N/D -

0.073 0.927 48 ⫺0.068 0.36

0.115 0.885 48 ⫺0.119 0.06

0.298 0.702 52 0.411 7.40

0.025 0.975 20 0 0.51

0 0.78 0.22 100 0.194 1.09

0 0.736 0.264 35 ⫺0.202 0.76

0 0.875 0.125 52 0.043 0.09

0 0.927 0.073 48 0.239 0.52

0.042 0.958 0 36 0.66 5.69

0 1 0 20

0 1 52

N/D N/D N/D -

0.214 0.786 42 ⫺0.119 0.17

0.433 0.327 0.24 52 0.208 2.94 0 1 48

-

0 0.716 0.284 51 ⫺0.098 0.19 0 1 52

1 52

0 1 52

0 1 51

-

0 1 50 -

0 1 36

-

1 50 -

1 36

-

0.01 0.99 50 0 0.51

0 1 36

-

0 1 50 -

0 1 36

0 1 0 36

-

-

1 52

-

1 48

-

-

-

0.11 0.89 82 0.008 0.21

0 1 20

-

0 0.862 0.138 47 ⫺0.15 0.14

0 0.761 0.239 46 ⫺0.184 0.86

0.035 0.862 0.104 58 0.583 51.11

0 0.775 0.225 20 0.022 0.23

-

0 1 52

-

0 1 48

-

-

0.005 0.801 0.194 98 ⫺0.236 5.25

0 0.98 0.02 49 ⫺0.011 0.01

0.76 0.24 52 ⫺0.308 3.69

0.582 0.418 49 0.046 0.02

0.431 0.569 36 0.169 0.54

0.4039 0.5962 52 ⫺0.109 0.32

0.208 0.792 48 0.377 5.22

0.328 0.672 58 0.069 0.09

0 1 20

0 0.202 0.798 52 ⫺0.124 0.30

0 0.633 0.367 49 ⫺0.131 0.47

0.015 0.632 0.353 34 0.271 2.09

0.022 0.597 0.38 46 ⫺0.169 0.90

0 0.87 0.13 46 0.055 0.06

0.0525 0.921 0.026 57 0.415 27.80

0 0.825 0.175 20 0.6 2.83

Codes as in Table 2. Underlined ␹2 values depict signiÞcant deviation from expected Hardy Weinberg frequencies at p ⬎ 0.05. N/D, no data. b

-

1 20

0 1 20

-

0 1 48

-

1 42

N/D N/D N/D -

0.77 0.23 57 1 54.20

-

0 1 52

-

-

-

44

JOURNAL OF MEDICAL ENTOMOLOGY

Vol. 40, no. 1

Fig. 2. Isoenzyme proÞles of 10 enzymes (12 loci) in Cx. pipiens complex members from California and SA.

rather than as three separate loci. Three alleles were observed, Hk90, Hk100, and Hk110, with Hk100 predominating in all populations. There was a slight latitudinal cline in frequency of Hk100 from 72% in northern California to 93% in the south. This agrees with the latitudinal clinal data for Hk reported by Urbanelli et al. (1997). Hk110 occurs only in Californian populations. Hk90 is rare in pipiens from Johannesburg. Idh. Two separate loci were observed for Idh. No variation was found at either locus for all Californian populations tested. The only variation observed was at the Idh-1 locus for the Johannesburg population, in which both pipiens and quinquefasciatus possessed equal frequencies of the two alleles. Mdhp-1. We were unable to reproduce the results reported by Urbanelli et al. (1997) at this locus. They observed that Mdhp-1 was highly diagnostic between pipiens and quinquefasciatus and signiÞcantly correlated with DV/D phenotypes in California. In our study, all Californian populations were monomorphic. Conversely, we found this locus to be diagnostic be-

tween pipiens and quinquefasciatus in SA, where all pipiens were monomorphic for Mdhp-190, and all quinquefasciatus were monomorphic for Mdhp-1100. No heterozygotes were observed. Mdhp-2. This enzyme locus was monomorphic for all populations of pipiens and quinquefasciatus in California. Similar monomorphism was observed in all SA and Boane (Mozambique) quinquefasciatus. Johannesburg pipiens was the only population that was polymorphic, with the presence of a slower allele (Mdhp-290) at a low frequency. Mdh-2. Three alleles were observed at this locus; a similar number was observed by others among members of the Cx. pipiens complex in California and Africa (Tabachnik and Powell 1983, Urbanelli et al. 1985, Urbanelli et al. 1997). Mdh-2100 was the most common allele in Californian and SA populations. Mdh-290 was the least common, occurring only in Redding, Reedley, and in pipiens from Johannesburg. Mdh-2110 occurred in north, central, and south California, and in SA pipiens and quinquefasciatus.

January 2003

CORNEL ET AL.: Cx. pipiens COMPLEX IN CALIFORNIA AND SOUTH AFRICA

Pgi. The two alleles observed at this locus were Pgi90 and Pgi100, which, based on the similarity in latitudinal cline in frequencies in California, are most likely equivalent to UrbanelliÕs et al. (1997) alleles of Gpi95 and Gpi100, respectively. Pgi100 was the most common allele in both pipiens and quinquefasciatus in SA. Pgi90 was present at low frequencies in Johannesburg pipiens and quinquefasciatus, but was not detected in Boane quinquefasciatus. Pgm. Three alleles were observed at this locus in California and SA, which is one less than that found in California by Urbanelli et al. (1997) and in African quinquefasciatus populations by Urbanelli et al. (1985). Pgm90 was uncommon in central and southern California (0.02) and uncommon in both pipiens and quinquefasciatus from SA (0.05). Frequencies of alleles Pgm100 and Pgm110 were similar to published (Tabachnik and Powell 1983, Urbanelli et al. 1997) records for the Californian Cx. pipiens complex. Both alleles showed a clinal transition from high frequency of Pgm110 in Cx. pipiens in the north to high frequency of Pgm100 in Cx. quinquefasciatus in southern California and intermediate frequencies in the hybrid zone. Pgm100 was the most common allele in both pipiens and quinquefasciatus in Johannesburg and Boane quinquefasciatus. Pgm110 was present at low frequencies in Johannesburg pipiens and quinquefasciatus and Boane quinquefasciatus. Among mosquitoes from California, there were no instances when a population deviated from Hardy Weinberg expectations (Table 5). Three enzyme loci (Hk, Pgm, and Pgi [Fig. 3]) showed a cline in allele frequency from northern to southern California, and hence showed some degree of differentiation between pipiens and quinquefasciatus. Only one enzyme locus (Ao) was not in Hardy Weinberg equilibrium in the Boane population, which, based on DV/D ratio and the other morphological characters, was quinquefasciatus. Allele frequencies for many of the enzyme loci (Ao, 6-Pgdh, Mdhp-1, Mdh-2, and Pgm) deviated from Hardy Weinberg expectations in the Johannesburg population. In all instances, too few heterozygotes (positive FIS values) were observed, suggesting the Johannesburg sample is not a single panmictic population. In most instances, Cx. pipiens complex members from California and SA possessed alleles of the same mobility. The only exceptions were alleles Aat-195, Aat-1110, Ao95, Hk90, Hk110, Idh-195, and Mdhp-290. Aat-1110, Hk110, and Ao95 were present at low frequency in California only, Aat-195 at low frequency in Johannesburg quinquefasciatus, Idh-195 at low frequency in Johannesburg pipiens and quinquefasciatus, and the remainder in low frequency in Johannesburg pipiens only. The three loci, Pgm, Pgi, and Hk, that indicated a latitudinal cline and allelic distinction between pipiens and quinquefasciatus populations in California showed no differentiation between SA pipiens and quinquefasciatus populations. Interestingly, the one enzyme locus (Mdhp-1) that proved to be diagnostic between SA pipiens and

45

quinquefasciatus was monomorphic in all populations of California. Wolbachia infection status. In all three Californian Cx. pipiens populations, all male and female mosquitoes tested were determined to be Wolbachia positive by PCR ampliÞcation of both the Wolbachia 16S rDNA and the Wolbachia wsp gene (Fig. 4). These data indicate that Wolbachia is widespread throughout California and occurs in pipiens, quinquefasciatus, and intermediate phenotypes. Conversely, only 22% (5 of 23) of the families reared from females collected in Johannesburg were Wolbachia positive. All positive individuals originated from quinquefasciatus families, and all negative mosquitoes originated from families identiÞed as pipiens based upon DV/D ratios. Individuals tested from each family were either all negative or all positive. All individuals tested from four quinquefasciatus families from Boane (Mozambique) were positive, and all individuals from four quinquefasciatus families reared from females collected in Albertsnek (100 km west of Boane just within the borders of SA) were also positive in both PCR assays. Although our sample sizes were relatively small, the data provide preliminary evidence that Wolbachia is present in SA quinquefasciatus, but is absent or occurs at a lower frequency in SA pipiens. Discussion Results from our study illustrate the complexity of taxonomic relationships among global populations of mosquitoes in the Cx. pipiens complex, and we are left with no clear resolution for the nomenclature of members of the Cx. pipiens complex if we follow the biological species deÞnition (Mayr 1970). In SA, the existence of Þve morphological characters that can reliably distinguish pipiens from quinquefasciatus, deviations from Hardy Weinberg expectations for several isoenzyme loci, presence of a diagnostic isoenzyme locus (Mdhp-1), and presence of Wolbachia in quinquefasciatus and absence in pipiens all provide compelling evidence that pipiens and quinquefasciatus remain reproductively isolated in an area in which they are sympatric. They behave as separate species in this part of the world, according to the biological species concept. However, such as in the Far East, these two taxa interbreed (Ishii 1980). In California, sympatric panmixes result, which prevents the use of the above-mentioned characters for separation of the taxa. This dichotomy of regional differences in reproductive isolation and extent of gene ßow among members of the complex has important implications. For instance, in California, we should expect to Þnd identical insecticide-resistance genes and mechanisms spreading in quinquefasciatus and in pipiens, whereas in SA, each of the two phenotypes could evolve distinctly different insecticide-resistance mechanisms that may require different resistance management strategies. If insecticide-susceptible or disease-blocking genes were purposefully introduced using either transposable elements or Wolbachia endosymbionts as carriers, two different scenarios should be expected

46

JOURNAL OF MEDICAL ENTOMOLOGY

Vol. 40, no. 1

Fig. 3. Plots of frequencies of alleles Hk100, Pgi100, and Pgm100 obtained from four Cx. pipiens populations from north to south in California. Best-Þt curve is also given with its equation and R2 value.

among members of the complex, depending on whether transgenetic individuals were released in California or in SA. In California, the introduced genes should spread from quinquefasciatus to pipiens or vice versa, but in SA, the genes should remain conÞned within whatever members of the complex was originally molecularly manipulated. Because of genetic and behavioral differences (host feeding preferences [Jupp 1973]) between quinquefasciatus and pipiens in SA, one should expect to observe associated differences in vectorial capacity for various diseases such as WN, Sindbis, and Rift Valley fever viruses. Hence in SA, quinquefasciatus has never been implicated as vectors of any of the above three viruses, whereas pipiens

has been implicated as playing a minor role in zoonotic cycles of WN and Sindbis viruses (Jupp and McIntosh 1967, Jupp 1976) and as an important epizootic vector of Rift Valley fever virus (Meegan 1979). Morphological characters whose measurements were nearly always different between pipiens or quinquefasciatus DV/D phenotypes in SA were overlapping in California, indicating that none of the morphological features we investigated, other than DV/D ratio, can be used with great conÞdence when distinguishing California populations of pipiens from quinquefasciatus. DV/D ratio measurements obtained from mosquitoes collected at all six Californian locations fell within expected ranges based on previous

January 2003

CORNEL ET AL.: Cx. pipiens COMPLEX IN CALIFORNIA AND SOUTH AFRICA

47

Fig. 4. Wolbachia and mosquito PCR-ampliÞed products run through an ethidium bromide-stained 1% agarose gel. Lane 1, SA quinquefasciatus; lane 2, SA pipiens; lane 3, individual from Anderson; lane 4, individual from Fresno; lane 5, individual from Los Angeles; lane 6, individual from tetracycline-treated colony of pipiens; lane 7, control (no DNA); and lane M, DNA size ladder. Expected size fragments ⫽ 400 bp for insect mitochondrial 12rDNA, 550 bp for Wolbachia surface protein gene (wsp), and 900 bp for Wolbachia 16S rDNA.

studies (Tabachnik and Powell 1983, Urbanelli et al. 1997). Cross matings between Johannesburg pipiens and quinquefasciatus, which were performed by Jupp (1978), conÞrmed that DV/D ratio, cross vein, maxillary palp and siphonal indexes, and subventral tuft branching of the F1 and F2 hybrids overlapped the ranges typical for pipiens and quinquefasciatus. Hence, mosquitoes found in the wild that had morphological values overlapping those that typically separate pipiens from quinquefasciatus would most likely represent hybrids. For comparative purposes, the ranges of data measured by Jupp in 1978 for southern African pipiens and quinquefasciatus are included in Table 4. All specimens reared from females collected in SA in 1989 and 2000 had DV/D ratios falling within either the range typical for pipiens (⬍0.2) or quinquefasciatus (⬎0.4), with none resembling ratios indicative of hybrids (0.2Ð 0.4). This is identical to that observed over 20 yr ago by Jupp (1978). In all instances, the male maxillary palp index, siphonal index, and branching of the siphonal setae 1aS and 1bS with few exceptions corroborated with JuppÕs (1978) data, and their ranges followed a priori family identiÞcation of either pipiens or quinquefasciatus based on their DV/D ratio. The morphological data provide evidence that pipiens and quinquefasciatus are still not interbreeding, and support the recommendation that they be considered separate species in SA. Jupp (1978) reported that of the Þve morphological features that showed the most distinction between Johannesburg populations of pipiens and quinquefasciatus, the cross vein index was the least reliable, with 10% of the males and 3% of the females sampled overlapping. Our data on male cross vein index concur, and in fact indicate an even greater overlap; thus, we

caution reliance on the use of this character to distinguish between male SA quinquefasciatus and pipiens. Skukuza village occurs in the low-lying region of SA that is warmer than Johannesburg and is considered subtropical, where according to convention only highly anthropophilic and endophilic quinquefasciatus populations occur. This is not the case, because only pipiens were collected in Skukuza. Even more unexpected was that in Skukuza, the pipiens were not collected in bird-baited traps, but by human biting and in CO2 and goat-baited net traps. This suggests that the pipiens phenotype is not uniform in its behavior throughout SA, and requires further investigation if we wish to clearly deÞne the distribution and rates that pipiens have in disease transmission in SA and elsewhere in Africa. Our allozyme data on Californian mosquitoes are in agreement with previous results by Cheng and Hacker (1976), Cheng et al. (1982), and Tabachnik and Powell (1983), and in partial agreement with Urbanelli et al. (1997). Urbanelli et al. (1997) identiÞed six enzyme loci that characterized California populations of pipiens and quinquefasciatus when electrophoresed through starch gels. We used acrylamide gels, and our data corroborate Urbanelli et al.Õs (1997) data, which depict a latitudinal cline of only three of the loci (Pgm, Pgi, and Hk). Our data at the Aat-2 and Mdhp-2 loci do not show the latitudinal cline reported by Urbanelli et al. (1997). We can offer no reason for those differences other than that what we observed were different enzymes or loci. Polymorphic, but highly diffuse banding that could not be conÞdently scored was observed when we stained for Mdhp using a continuous Triscitrate buffer system. Perhaps this diffuse system produces better resolution in the starch gels used by

48

JOURNAL OF MEDICAL ENTOMOLOGY

Urbanelli et al. (1997). We did not stain for the enzyme hydroxybutyrate dehydrogenase (E.C. 1.1.1.30), thus precluding that comparisons could be made. The high frequency of Aat-1100 in Johannesburg pipiens is consistent with the observation made by Urbanelli et al. (1985) of its predominance in European pipiens and molestus, although this is not the case in Madagascar, where alleles 90 and 100 occurred in similar frequencies in both pipiens and quinquefasciatus (Urbanelli et al. 1995). Recent evidence indicates that a gene that most likely codes for Ao is tightly linked to and coampliÞed with a cluster of insecticide resistance-associated carboxy-esterases in quinquefasciatus (Hemingway et al. 2000). Southern Africa has a long history of chemical indoor malaria vector control programs, and it is possible that endophilic and highly anthropophilic quinquefasciatus mosquitoes have been exposed to insecticides as nontarget organisms. The Boane quinquefasciatus population is resistant to numerous organophosphates and pyrethroids and is polymorphic for elevated ␣ and ␤ carboxy-esterases (Cornel, unpublished data). Hence, deviations from Hardy Weinberg equilibrium for Ao in a malaria-endemic location such as Boane may be because of selection to insecticide-resistance esterases to which Ao is linked. The fact that all other enzyme loci did not show deviations from Hardy Weinberg equilibrium expectations in Boane suggests that the Boane population represents a panmictic population of quinquefasciatus. Ao allele frequencies were also not in Hardy Weinberg equilibrium at the 95% conÞdence interval for the Johannesburg Cx. pipiens complex populations. While assortative matings between pipiens and quinquefasciatus alone would account for this disequilibrium, unique allele mobilities and staining intensities of nonspeciÞc esterases were also evident between Johannesburg pipiens and quinquefasciatus families (Cornel, unpublished data). Hence, linkage disequilibrium between Ao and esterase loci may also be occurring in Johannesburg in response to household use of insecticides. We hypothesize that clinal transitions of DV/D ratios and enzymes in California and the rest of North America are not a reßection of interrupted gene ßow, but a consequence of environmental temperature selection. If there was interrupted gene ßow between pipiens and quinquefasciatus in California, in areas of sympatry we would expect to see discrete morphological indexes and allelic patterns like those that identify separation between SA pipiens and quinquefasciatus. The only morphological character that shows differences between northern and southern Californian populations is the male DV/D ratio. However, even the neutrality of DV/D ratios is in question, as breeding experiments by Wilton and Jakob (1985) have shown that DV/D ratio values of Cx. pipiens complex hybrids are inßuenced by temperatures. Exposure to cooler temperate temperatures over several generations selected smaller DV/D ratios (pipiens phenotype) and exposure to warmer subtropical to tropical temperatures selected larger DV/D ratios

Vol. 40, no. 1

(quinquefasciatus phenotype). In climatically temperate Johannesburg, DV/D ratios typical for quinquefasciatus are still observed, and in climatically subtropical Skukuza (SA), DV/D ratios within the pipiens range still exist. The reason for this, supporting Wilton and JakobÕs (1985) observations, is that the SA Cx. pipiens complex members are not hybrids. Enzyme electrophoresis as a genetic tool to distinguish between quinquefasciatus and pipiens should be used with caution. Latitudinal clines observed in the enzymes Hk, Pgm, and Pgi in this study and in hydroxybutyrate dehydrogenase and Mdhp-2 by Urbanelli et al. (1997) could be because of selection for alleles that perform optimally at speciÞc temperatures. Thermal stability differences of the partially puriÞed major alleles of the enzymes Hk, 6-phosphogluconate dehydrogenase (E.C. 1.1.1.44), and ␣-glycerophosphate dehydrogenase (E.C. 1.1.1.8) have been found by comparative relative activity studies at different temperatures (Pryor et al. 1980a, b, Pryor and Ferrell 1981). Alleles at the Aat-1 locus among members of the Cx. pipiens complex in Europe are also suspected to be inßuenced by selection for preference to breeding in hypogeous versus epigeous habitats (Chevillon et al. 1998). There are other allozymes such as Lactate dehydrogenase (E.C. 1.1.1.27) and Alcohol dehydrogenase (E.C. 1.1.1.1) that exhibit correlations between biochemical properties and geographical location in other organisms for which there are no questions about their taxonomic status (Pryor et al. 1980a). In California, Wolbachia infection is widespread in Cx. pipiens complex populations throughout the state. Mosquitoes from distant localities are reproductively compatible with one another (Rasgon, unpublished), consistent with the idea that the same strain of Wolbachia infects these populations. Similarity in Wolbachia strains has also been conÞrmed by sequencing data (Rasgon, unpublished). Results from our Wolbachia analyses are consistent with extensive gene ßow throughout California pipiens complex populations. If gene ßow between members of the Cx. pipiens complex occurred in Johannesburg, we would expect Wolbachia to spread through and be detected in both pipiens and quinquefasciatus. We detected Wolbachia only in quinquefasciatus. Members of the Cx. pipiens complex examined worldwide have mostly been shown to be infected with Wolbachia (Laven 1967, Irving-Bell 1983, OÕNeill and Paterson 1992, Kittayapong et al. 2000). Irving-Bell (cited in Miles and Paterson 1979) reported absence of rickettsial endosymbionts in southern African pipiens and presence in quinquefasciatus. Over 20 yr later, data from our study indicate that this situation still continues today in SA. Wolbachia-induced cytoplasmic incompatibility can cause matings between different cytotypes to be sterile (Barr 1982). Although Wolbachia-induced cytoplasmic incompatibility has been envisioned as a possible mechanism for reproductive isolation and, therefore, an agent for speciation (Werren 1997), others suggest that Wolbachia alone is insufÞcient to act

January 2003

CORNEL ET AL.: Cx. pipiens COMPLEX IN CALIFORNIA AND SOUTH AFRICA

as a reproductive isolating mechanism, and thus drive speciation events (Turelli 1994). In SA, there is no reason to suspect reduced opportunity for contact between pipiens and quinquefasciatus, as they share similar above ground breeding niches (Hopkins 1952 and unpublished data by Cornel), blood hosts, and resting sites, and are not temporally isolated. This is unlike the situation in London, where there is reduced opportunity for contact between members of the complex (Byrne and Nichols 1998). Acknowledgments We thank L. Braack for providing permission and assistance in collection of mosquitoes in the Kruger National Park; Prof. R. H. Hunt (University of the Witwatersrand) for collecting mosquitoes from Boane (Mozambique) and for providing comments on the manuscript; J. Beehler (West Valley Mosquito and Vector Control District) for collecting the sample from Chino (CA); and John Albright (Shasta Mosquito and Vector Control District) for providing Þeld assistance in Redding (CA). This work was supported in part from funds allocated from the University of California Mosquito Research Program.

References Cited Ayala, F. J., and J. R. Powell. 1972. Allozymes as diagnostic characters of sibling species of Drosophila. Proc. Natl. Acad. Sci. USA 69: 1094 Ð1096. Anderson, J. F., T. G. Andreadis, C. R. Vossbrink, S. Tirrell, E. M. Wakem, R. A. French, A. E. Garmendia, and H. J. van Kruiningen. 2000. Isolation of West Nile from mosquitoes, crows, and a CooperÕs hawk in Connecticut. Science 286: 2331. Atkinson, C. T., K. L. Woods, R. J. Dusek, L. S. Sileo, and W. M. Iko. 1995. Wildlife disease and conservation in Hawaii: pathogenicity of avian malaria (Plasmodium relictum) in experimentally infected Iiwi (Vestiaria coccinea). Parasitology 111 (suppl.): S59 ÐS69. Barr, R. A. 1982. The Culex pipiens complex, pp. 551Ð572. In W.W.M. Steiner, W. J. Tabachnik, K. S. Rai, and S. Narang [eds.], Recent developments in the genetics of insect disease vectors. Stipes, Urbana, IL. Belkin, J. N. 1962. The mosquitoes of the south PaciÞc (Diptera: Culicidae), vols. 1 and 2. University of California Press, Berkeley. Ben Cheikh, H., Z. Ben Ali-Haouas, M. Marquine, and N. Pasteur. 1998. Resistance to organophosphorus and pyrethroid insecticides in Culex pipiens (Diptera: Culicidae) from Tunisia. J. Med. Entomol. 35: 251Ð260. Bisset, J. A., M. M. Rodriguez, J. H. Hemingway, C. Diaz, A. J. Small, and E. Ortiz. 1991. Malathion and pyrethroid resistance in Culex quinquefasciatus from Cuba: efÞcacy of pirimiphos-methyl in the presence of at least three resistance mechanisms. Med. Vet. Entomol. 5: 223Ð228. Black, W. C. IV, and L. E. Munstermann. 1996. Molecular taxonomy and systematics of arthropod vectors, pp. 438 Ð 470. In B. J. Beaty and W. C. Marquardt [eds.], The biology of disease vectors. University Press of Colorado, Niwot, CO. Black, W. C., and N. M. DuTeau. 1997. RAPD-PCR and SSCP analysis for insect population genetics studies, pp. 361Ð373. In J. M. Crampton, C. B. Beard, and C. Louis

49

[eds.], The molecular biology of insect disease vectors. Chapman & Hall, London, UK. Bourguet, D., D. Fonseca, G. Vourch, M.-P. Dubois, F. Chandre, C. Severini, and M. Raymond. 1998. The acetylcholinesterase gene ACE: a diagnostic marker for the pipiens and quinquefasciatus forms of the Culex pipiens complex. J. Am. Mosq. Control Assoc. 14: 390 Ð396. Brewer, G. J., and C. F. Singh. 1970. An introduction to isoenzyme tenchniques. Academic, New York. Byrne, K., and R. A. Nichols. 1998. Culex pipiens in London Underground tunnels: differentiation between surface and subterranean populations. Heredity 82: 7Ð15. Callaghan, A., T. Guillemaud, N. Makate, and M. Raymond. 1998. Polymorphisms and ßuctuation in copy number of ampliÞed esterase genes in Culex pipiens mosquitoes. Insect Mol. Biol. 7: 295Ð300. Chandre, F., F. Darriet, M. Darder, A. Cuany, J.M.C. Doannio, N. Pasteur, and P. Guillet. 1998. Pyrethroid resistance in Culex quinquefasciatus from West Africa. Med. Vet. Entomol. 12: 359 Ð366. Cheng, M. L., and C. S. Hacker. 1976. Inheritance of 6-phosphogluconate dehydrogenase variance in Culex pipiens quinquefasciatus Say. J. Hered. 67: 215Ð219. Cheng, M. L., and C. S. Hacker. 1979. The genetics of hexokinase in a mosquito, Culex pipiens. Genetics 92: 903Ð913. Cheng, M. L., C. S. Hacker, S. C. Pryor, R. E. Ferrel, and G. B. Kitto. 1982. The ecological genetics of the Culex pipiens complex in North America, pp. 581Ð 627. In W.W.M. Steiner, W. J. Tabachnik, K. S. Rai, and S. Narang [eds.], Recent developments in the genetics of insect disease vectors. Stipes, Urbana, IL. Chevillon, C., Y. Rivet, M. Raymond, F. Rousset, P. E. Smouse, and N. Pasteur. 1998. Migration/Selection balance and ecotypic differentiation in the mosquito Culex pipiens. Mol. Ecol. 7: 197Ð208. Crabtree, M. B., H. M. Savage, and B. R. Miller. 1997. Development of a polymerase chain reaction assay for differentiation between Culex pipiens pipiens and Cx. p. quinquefasciatus (Diptera: Culicidae) in North America based on genomic differences identiÞed by subtractive hybridization. J. Med. Entomol. 34: 532Ð537. Dobrotworsky, N. V. 1967. The problem of the Culex pipiens complex in the south PaciÞc (including Australia). Bull. W.H.O. 37: 251Ð255. Farid, H. A., R. E. Hammad, M. M. Hassan, Z. S. Morsy, I. H. Kamal, G. J. Weil, and R.M.R. Ramzy. 2001. Detection of Wuchereria bancrofti in mosquitoes by the polymerase chain reaction: a potentially useful tool for large-scale control programmes. Trans. R. Soc. Trop. Med. Hyg. 95: 29 Ð32. Fowler, J., L. Cohen, and P. Jarvis. 1998. Practical statistics for Þeld biology. John Wiley, New York. Guillemaud, T., N. Pasteur, and F. Rousett. 1997. Contrasting levels of variability between cytoplasmic genomes and incompatibility types in the mosquito Culex pipiens. Proc. R. Soc. London Ser. B 264: 245Ð251. Handler, A. M., and A. A. James. 2000. Insect transgenesis: methods and applications. CRC LLC, Boca Raton, FL. Harbach, R. E., B. A. Harrison, and A. M. Gad. 1984. Culex (Culex) molestus Forskal (Diptera: Culicidae): neotype designation, description, variation, and taxonomic status. Proc. Entomol. Soc. Wash. 86: 521Ð542. Harbach, R. E., C. Dahl, and G. B. White. 1985. Culex (Culex) pipiens Linnaeus (Diptera: Culicidae): concepts, type designations, and description. Proc. Entomol. Soc. Wash. 87: 1Ð24. Hemingway, J., and H.P.P. Karunaratne. 1998. Mosquito carboxylesterases: a review of the molecular biology and

50

JOURNAL OF MEDICAL ENTOMOLOGY

biochemistry of a major insecticide resistance mechanism. Med. Vet. Entomol. 12: 1Ð12. Hemingway, J., M. Coleman, M. Paton, L. McCarrol, A. Vaughan, and D. DeSilva. 2000. Aldehyde oxidase is coampliÞed with the worldÕs most common Culex mosquito insecticide resistance-associated esterases. Insect Mol. Biol. 9: 93Ð99. Hertig, M., and S. B. Wolbach. 1924. Studies on rickettsialike micro-organisms in insects. J. Med. Res. 44: 329 Ð374. Hopkins, G.H.E. 1952. Mosquitoes of the Ethiopian region. I. Larval bionomics of mosquitoes and taxonomy of culicine larvae. British Museum (Natural History), London, UK. Hubalek, Z., and J. Halouzka. 1999. West Nile fever: a reemerging mosquito borne viral disease in Europe. Emerg. Infect. Dis. 5: 643Ð 650. Irving-Bell, R. J. 1983. Cytoplasmic incompatibility within and between Culex molestus and Culex quinquefasciatus. J. Med. Entomol. 20: 44 Ð 48. Ishii, T. 1980. On the Culex pipiens group in Japan. III. A historical review of its research. IV. Review of the adult character (3). J. Sci. Coll. Gen. Ed. Univ. Tokushima 13: 26 Ð 62. Jupp, P. G. 1973. Field studies on the feeding habits of mosquitoes in the highveld region of South Africa. S. Afr. J. Med. Sci. 38: 69 Ð 83. Jupp, P. G. 1976. The susceptibility of four South African species of Culex to West Nile and Sindbis viruses by two different infecting methods. Mosq. News 36: 166 Ð173. Jupp, P. G. 1978. Culex (Culex) pipiens pipiens Linnaeus and Culex (Culex) pipiens quinquefasciatus Say in South Africa: morphological and reproductive evidence in favour of their status as two species. Mosq. Systematics 10: 461Ð 473. Jupp, P. G., and B. M. McIntosh. 1967. Ecological studies on Sindbis and West Nile viruses in South Africa. II. Mosquito bionomics. S. Afr. J. Med. Sci. 32: 15Ð33. Jupp, P. G., and B. M. McIntosh. 1970. Quantitative experiments on the vector capability of Culex (Culex) pipiens fatigans Weidemann with West Nile and Sindbis viruses. J. Med. Entomol. 7: 353Ð356. Kasai, S., I. S. Weerashinghe, and T. Shono. 1998. P450 monooxygenase in Culex quinquefasciatus Say larvae. Arch. Insect Biochem. Physiol. 37: 47Ð56. Kittayapong, P., K. J. Baisley, V. Baimai, and S. L. O’Neill. 2000. Distribution and diversity of Wolbachia infections in southeast Asian mosquitoes (Diptera: Culicidae). J. Med. Entomol. 37: 340 Ð345. Lai, C. H., K. C. Tung, H. K. Ooi, and J. S. Wang. 2000. Competence of Aedes albopictus and Culex quinquefasciatus as a vector of Dirofilaria immitis after blood meal with different microÞlarial density. Vet. Parasitol. 90: 231Ð 237. Lanciotti, R. S., J. T. Roehrig, V. Deubel, J. Smith, M. Parker, K. Steele, B. Crise, K. E. Volpe, M. B. Crabtree, J. H. Scherret, R. A. Hall, J. S. MacKenzie, C. B. Cropp, B. Panigrahy, E. Ostlund, B. Schmitt, M. Malkinson, C. Banet, J. Weissman, N. Komar, H. M. Savage, W. Stone, T. McNamara, and D. J. Gubler. 1999. Origin of the West Nile virus responsible for an outbreak of encephalitis in the northeastern United States. Science 286: 2333Ð2337. Lanciotti, R. S., A. J. Kerst, R. S. Nasci, M. S. Godsey, C. J. Mitchell, H. M. Savage, N. Komar, N. A. Panella, B. C. Allen, K. E. Volpe, B. S. Davis, and J. T. Roehrig. 2000. Rapid detection of West Nile virus from human clinical specimens, Þeld-collected mosquitoes, and avian samples by a TaqMan reverse transcriptase-PCR assay. J. Clin. Microbiol. 38: 4066 Ð 4071.

Vol. 40, no. 1

Laven, H. 1967. Speciation and evolution in Culex pipiens, pp. 251Ð275. In J. W. Wright and R. Pal [eds.], Genetics of insect vectors of disease. Elsevier Publishing Company, Amsterdam, The Netherlands. Manchenko, G. P. 1994. Handbook of detection of enzymes on electrophoretic gels. CRC, Boca Raton, FL. Mattingly, P. F. 1967. The systematics of the Culex pipiens complex. Bull. W.H.O. 37: 257Ð261. Mayr, E. 1970. Populations, species and evolution. Harvard University Press, Cambridge, MA. Meegan, J. M. 1979. The Rift Valley fever epizootic in Egypt 1977Ð1978. Trans. R. Soc. Trop. Med. Hyg. 73: 618 Ð 623. Miles, S. J., and H. E. Paterson. 1979. Protein variation and systematics in the Culex pipiens group of species. Mosq. Systematics 11: 187Ð202. Nielsen-Leroux, C., F. Pasquier, J. F. Charles, G. Sinegre, B. Gaven, and N. Pasteur. 1997. Resistance to Bacillus sphaericus involves different mechanisms in Culex pipiens (Diptera: Culicidae) larvae. J. Med. Entomol. 34: 321Ð327. O’Neill, S. L., R. Giordano, A.M.E. Colbert, T. L. Karr, and H. M. Robertson. 1992. 16S rRNA phylogenetic analysis of the bacterial endosymbionts associated with cytoplasmic incompatibility in insects. Proc. Natl. Acad. Sci. USA 89: 2699 Ð2702. O’Neill, S. L., and H.E.H. Paterson. 1992. Crossing type variability associated with cytoplasmic incompatibility in Australian populations of the mosquito Culex quinquefasciatus Say. Med. Vet. Entomol. 6: 209 Ð216. O’Neill, S. L., R. H. Gooding, and S. Aksoy. 1993. Phylogenetically distant symbiotic microorganisms reside in Glossina midgut and ovary tissues. Med. Vet. Entomol. 7: 377Ð383. Pryor, S. C., M. L. Cheng, R. E. Ferrell, and C. S. Hacker. 1980a. Biochemical genetics of the Culex pipiens complex. I. 6-Phosphogluconate dehydrogenase. Comp. Biochem. Physiol. 65B: 663Ð 668. Pryor, S. C., M. L. Cheng, R. E. Ferrell, and C. S. Hacker. 1980b. Biochemical genetics of the Culex pipiens complex. II. Hexokinase. Comp. Biochem. Physiol. 67B: 705Ð 710. Pryor, S. C., and R. E. Ferrell. 1981. Biochemical genetics of the Culex pipiens complex. III. ␣-Glycerophosphate dehydrogenase. Comp. Biochem. Physiol. 69B: 23Ð28. Raymond, M., A. Callaghan, P. Fort, and N. Pasteur. 1991. Worldwide migration of ampliÞed insecticide resistance genes in mosquitoes. Nature 350: 151Ð153. Raymond, M., and F. Rousset. 1995. Genepop (version 1.2): population genetics software for exact tests and ecumenicism. J. Hered. 86: 248 Ð249. SAS Institute. 1998. StatView reference. SAS Institute, Cary, NC. Shaw, C. R., and R. Prasad. 1970. Starch gel electrophoresis of enzymes: a compilation of recipes. Biochem. Genet. 4: 297Ð320. Sinkins, S. P., and S. L. O’Neill. 2000. Wolbachia as a vehicle to modify insect populations, pp. 271Ð287. In A. M. Handler and A. A. James [eds.], Insect transgenesis: methods and applications. CRC LLC, Boca Raton, FL. Sirivanakarn, S., and G. B. White. 1978. Neotype designation of Culex quinquefasciatus Say (Diptera: Culicidae). Proc. Entomol. Soc. Wash. 30: 360 Ð372. Small, G.J.S., H.P.P. Karunaratne, D. D. Chadee, and J. Hemingway. 1999. Molecular and kinetic evidence for allelic variants of esterase Est␤1 in the mosquito Culex quinquefasciatus. Med. Vet. Entomol. 13: 274 Ð281. Sokal, R. R., and F. J. Rohlf. 1981. Biometry, 2nd ed. Freeman, San Francisco, CA.

January 2003

CORNEL ET AL.: Cx. pipiens COMPLEX IN CALIFORNIA AND SOUTH AFRICA

Steiner, W.W.M., and D. J. Joslyn. 1979. Electrophoretic techniques for the genetic study of mosquitoes. Mosq. News 39: 35Ð54. Tabachnik, W. J., and J. R. Powell. 1983. Genetic analysis of Culex pipiens populations in the Central Valley of California. Ann. Entomol. Soc. Am. 76: 715Ð720. Tsai, T. F., and C. J. Mitchell. 1989. St. Louis encephalitis, pp. 113Ð143. In T. P. Monath [ed.], The arboviruses: epidemiology and ecology, vol. IV. CRC, Boca Raton, FL. Turelli, M. 1994. Evolution of incompatibility-inducing microbes and their hosts. Evolution 48: 1500 Ð1513. Urbanelli, S., L. Bullini, and F. Villani. 1985. Electrophoretic studies on Culex quinquefasciatus Say from Africa: genetic variability and divergence from Culex pipiens L. (Diptera: Culicidae). Bull. Entomol. Res. 75: 291Ð304. Urbanelli, S., F. Silvestrini, G. Sabatinelli, F. Raveloarifera, V. Petrarca, and L. Bullini. 1995. Characterization of the Culex pipiens complex (Diptera: Culicidae) in Madagascar. J. Med. Entomol. 32: 778 Ð786. Urbanelli, S., F. Silvestrini, W. K. Reisen, E. De Vito, and L. Bullini. 1997. California hybrid zone between Culex

51

pipiens pipiens and Cx. p. quinquefasciatus revisited (Diptera: Culicidae). J. Med. Entomol. 34: 116 Ð127. Varma, M.G.R. 1960. Preliminary studies on the infection of culicine mosquitoes with the Tamiland strain of West Nile virus. Indian J. Med. Res. 48: 537Ð548. Werren, J. H. 1997. Biology of Wolbachia. Annu. Rev. Entomol. 42: 587Ð 609. Wilton, D. P., and W. L. Jakob. 1985. Temperature-induced morphological change in Culex pipiens. J. Am. Mosq. Control Assoc. 1: 174 Ð177. Yuan, Z., Y. Zhang, Q. Cai, and E.-Y. Liu. 2000. High-level resistance to Bacillus sphaericus C3-41 in Culex quinquefasciatus from Southern China. Biocontrol Science Technology 10: 41Ð 49. Zhou, W., F. Rousset, and S. O’Neill. 1998. Phylogeny and PCR-based classiÞcation of Wolbachia strains using wsp gene sequences. Proc. R. Soc. Lond. B. Biol. Sci. 265: 509 Ð515. Received for publication 29 October 2001; accepted 1 July 2002.