Report

A Mite Species That Consists Entirely of Haploid Females

See allHide authors and affiliations

Science  29 Jun 2001:
Vol. 292, Issue 5526, pp. 2479-2482
DOI: 10.1126/science.1060411

Abstract

The dominance of the diploid state in higher organisms, with haploidy generally confined to the gametic phase, has led to the perception that diploidy is favored by selection. This view is highlighted by the fact that no known female organism within the Metazoa exists exclusively (or even for a prolonged period) in a haploid state. We used fluorescence microscopy and variation at nine microsatellite loci to show that the false spider mite,Brevipalpus phoenicis, consists of haploid female parthenogens. We show that this reproductive anomaly is caused by infection by an undescribed endosymbiotic bacterium, which results in feminization of haploid genetic males.

It is commonly thought that no female organism within the Metazoa exists exclusively in a haploid state, because selection favors diploidy in higher organisms (1, 2). There are several theories about the evolution of diploidy as the dominant state, and deleterious mutations (germline and/or somatic) are considered to be the driving force (3, 4). Yet experimental evidence is lacking, and currently no studies on life cycle evolution have been conducted in a higher organism.

Brevipalpus phoenicis Geijskes (Acari: Tenuipalpidae) is a minute phytophagous mite found throughout tropical and subtropical regions. It is polyphagous and is a major pest of many economically important crops such as citrus, coffee, tea, papaya, passion fruit, and palms (5). In citrus, it acts as a vector for citrus leprosis virus (Rhabdoviridae), a disease that causes millions of dollars of damage to the Brazilian citrus industry each year (6). Brevipalpus phoenicis, along with two closely related species, B. obovatus and B. californicus, is known to reproduce by thelytokous (obligate) parthenogenesis (7). Rare males are found in field populations; however, their function is not known (8). The closest sexual relative, B. russulus, is haplo-diploid, in which unfertilized eggs develop into haploid males (two chromosomes) whereas fertilized eggs develop into diploid females (four chromosomes) (9). Haplo-diploidy is characteristic of their superfamily, the Tetranychoidea. All three parthenogenetic species have only two chromosomes in somatic cells (Fig. 1A) (10); however, owing to the small size of these chromosomes and their apparent lack of any distinguishable morphological character, it is not known whether this represents the haploid or diploid state.

Figure 1

Chromosomes of B. phoenicisobtained from 2-day-old eggs and viewed under a fluorescence microscope. (A) Metaphase chromosomes stained with YOYO-1, showing the presence of two chromosomes in a mitotic division. (B and C) Prophase chromosomes stained with YOYO-1. Arrows indicate a single NOR. (D andE) FISH experiments. DAPI counterstained metaphase chromosomes are shown after hybridization to the biotinylated 18S rDNA probe detected with Cy3-conjugated streptavidin (Jackson ImmunoResearch Labs, West Grove, Pennsylvania). Arrows indicate the hybridization signal (red) on each sister chromatid. There is no corresponding hybridization signal on the second chromosome. Scale bar in (A) indicates 2 μm in (A) through (E).

Although it has been proposed that B. obovatus is a haploid parthenogen (7, 9), convincing cytological evidence has been lacking (11, 12). Cytological techniques using fluorescence microscopy provide an accurate way to determine whether these three species are haploid or diploid parthenogens. We used two such techniques, as well as genetic variation at nine microsatellite loci, to show that B. phoenicis is indeed a haploid female parthenogen.

Brevipalpus phoenicis females collected from a coffee plantation at the University of Sao Paulo, Piracicaba, Sao Paulo, Brazil, were used to initiate five isofemale lines (lines started with a single immature female) from which eggs were used for the following experiments. Using a fluorescent dye (YOYO-1, Molecular Probes) that stains both DNA and RNA, we visualized the nuclear organizing region (NOR), which is present during early prophase in mitotic divisions in eggs of B. phoenicis (13,14). Homologous pairs of chromosomes will have a NOR or NORs at exactly the same position. We found one NOR to be present during early prophase mitotic divisions in 2-day-old eggs (Fig. 1, B and C) from B. phoenicis. The NOR was found at the tip of one chromosome, with no corresponding NOR being present on the second chromosome.

To complement the finding of a single NOR, we also used fluorescent in situ hybridization (FISH) to locate the ribosomal DNA (rDNA) region(s) within metaphase chromosomes. Based on the finding of a single NOR, only one region of rDNA should be present. A probe was made from the small subunit 18S rDNA of B. phoenicis(15). The same FISH procedure was followed as that outlined in (16), with a slight modification (17). The 18S rDNA probe from B. phoenicis hybridizes at the tip of both sister chromatids of one chromosome, in a metaphase mitotic division (Fig. 1, D and E). No hybridization occurred on the second chromosome. This concurs with the single NOR found with YOYO-1 (Fig. 1, B and C).

We also screened 45 clonal lines of B. phoeniciscollected from Sao Paulo and Minas Gerais, Brazil, for variation at seven polymorphic microsatellite loci (18) and screened two of these clonal lines at two additional loci (Table 1). Ten individuals per clonal line were genotyped, with no differences found within a clonal line. No heterozygous individuals (in a total of 450 individuals), at any locus, were found. These data, along with the finding of a single NOR and only one rDNA region, clearly show B. phoenicis to be a haploid female parthenogen.

Table 1

Characteristics of nine polymorphic microsatellite loci (for n = 45 clonal lines, except loci Brev08 and 09, where n = 2 clonal lines) in B. phoenicis collected from two different host plant species (citrus and coffee) in Sao Paulo, Brazil. Ten individuals were assayed per clonal line, with no differences being found within a clonal line for each locus. No heterozygous loci were found in a total of 450 individuals. Microsatellites and their flanking sequences can be found in GenBank under accession numbers AF335574 through AF335582.

View this table:

During the above cytological examination of mitotic chromosomes in eggs, we found large numbers of an endosymbiotic bacterium (Fig. 2) and attempted to discover whether it was involved in female haploid parthenogenesis.

Figure 2

Bacteria stained with YOYO-1 from a single 2-day-old egg of B. phoenicis viewed under a fluorescence microscope. Scale bar, 5 μm.

We amplified and sequenced 1487 base pairs (bp) of the bacterial 16S rDNA from the isofemale lines used in the cytological experiments using general eubacterial primers (19). Only one bacterial 16S rDNA sequence was found. When compared with all known bacterial 16S rDNA sequences in GenBank, this bacterium's closest relative was an undescribed endosymbiotic bacterium found in the tick Ixodes scapularis(20) (98% sequence homology), belonging to the phylumCytophaga-Flavobacterium-Bacteroides. To test whether this bacterium is involved in causing parthenogenesis, we treated young adult females from one isofemale line of B. phoenicis with the antibiotic tetracycline hydrochloride (21). After 3 days of tetracycline treatment, females were allowed to lay eggs for 10 days. The resulting progeny were then scored at the adult stage for sex (Table 2). Young adult females from the same isofemale line were also allowed to lay eggs for 10 days without tetracycline treatment to act as a control. Significantly more males were produced after tetracycline treatment [one-way analysis of variance (ANOVA) after arcsine transformation,F 1,21 = 18.69, P < 0.001], with 51% of the progeny being male. This is expected under haplo-diploidy (which is likely to be the ancestral state in B. phoenicis), because the females are unfertilized, and curing of a bacterial infection that causes parthenogenesis will result in haploid males. The females that were produced would still be expected to carry the bacterial infection, because tetracycline curing of adults is not 100% effective. More males were produced in the control treatment (∼5%) than expected (10), and this may be due to the temperature at which the experiment took place (30°C). In some parthenogenetic Trichogramma species infected with the endosymbiotic bacterium Wolbachia (which causes the parthenogenesis), rearing at high temperatures (>28°C) can kill the bacteria, resulting in significantly greater male production (22).

Table 2

Sex ratio produced by B. phoenicis females treated with 0.2% tetracycline hydrochloride for 72 hours and left to lay eggs for 10 days.

View this table:

Using polymerase chain reaction (PCR) primers specific for the bacterium (23), all females tested, regardless of the treatment (50 from controls and 80 from tetracycline treatment) were infected with the bacterium, whereas all males tested (17 from controls and 72 from tetracycline treatment) were not infected.

Our results show that females are haploid and that the female haploid parthenogenesis is caused by an endosymbiotic bacterial infection, which results in feminization of genetic males. Once cured of the infection, an adult haploid female will lay haploid male offspring. How the bacterium induces feminization of genetic males is not known. This is the first time feminization by an extrachromosomal factor has been found outside of heterogametic reproductive systems. In the isopodArmadillidium vulgare, the bacterium Wolbachiaalso induces feminization (24) by blocking the formation of the androgenic gland, which produces the androgenic hormone responsible for male differentiation (25). A similar mechanism may occur here, resulting in haploid females instead of haploid males.

The data presented here illustrate that feminization can involve bacteria other than Wolbachia. The endosymbiotic bacteriumWolbachia has been known to induce a number of reproductive phenotypes, including cytoplasmic incompatibility, parthenogenesis, feminization, and male-killing in various arthropod species (22) and has been implicated in sex determination (26) and speciation (27). We have found another bacterium, unrelated to Wolbachia, that causes feminization. Recently, a similar bacterium (with 96% sequence homology for 16S rDNA) has been found to be associated with parthenogenesis in several species of parasitic wasps (28). This suggests that we should not be looking specifically forWolbachia to explain reproductive abnormalities, as some have done previously (29–32).

There remains a possibility that B. phoenicis is diploid and that one NOR (and corresponding rDNA) has been lost, as has been found with some thelytokous aphid species (33). However, the evidence does not support this. First, the only known cytogenetic mechanisms of parthenogenesis that could result in complete homozygosity at all microsatellite loci studied is gamete duplication or terminal fusion (with no crossing over) (34). A single NOR from a homologous pair cannot be lost through gamete duplication or terminal fusion, because loss would in the subsequent generation result in either no NOR at all (which presumably would be lethal) or the homologous pair would be restored (34). Second, if curing of the bacterium results in male progeny, then males must be haploid with n = 1. Yet the number of chromosomes in male sperm and developing embryos is two (8, 9).

Finally, because B. phoenicis appears to live entirely in a haploid state, there is a unique opportunity to study life cycle evolution and the advantages and disadvantages associated with haploidy in a higher organism, where diploidy dominates. Typically, these questions have been studied in bacteria, fungi, moss, algae, and ferns (1) where life cycles and ploidy levels are not always dominated by a diploid phase, and therefore their relevance may be questioned. If diploid sexual females of B. phoenicis exist or can be generated, or if generation of haploid female B. russulus is possible through the B. phoenicisbacterium, it may be possible to gain the first empirical data on the relative importance of somatic and germline deleterious mutations for the maintenance of diploidy in animals.

  • * To whom correspondence should be addressed at the Centre for Environmental Stress and Adaptation Research, Department of Biological Sciences, Monash University, Clayton 3168, Australia. E-mail: Andrew.Weeks{at}sci.monash.edu.au

REFERENCES AND NOTES

View Abstract

Navigate This Article