Mutant analysis and genetic complementation are powerful strategies to link specific genes to biological pathways [46
]. Efforts to develop these methods in Toxoplasma
have resulted in an episome-based protocol [49
] and an insertion based approach [31
]. In the latter, phage recombination [50
] was used to mobilize cDNA fragments integrated into parasite transformants [30
]. While these earlier methods achieved some success, the redundancy inherent to cDNA libraries limited the complementation success (B. Striepen and M. White, unpublished data). The genomic-DNA based approach introduced here benefits from the large inserts carried by cosmid vectors to deliver genes in their natural chromosome organization. As a consequence of these improvements, our success rate for genetic rescue of ts
growth mutants has improved dramatically. Here we report on the complementation of a diverse selection of mutants (>20 mutants, with a failure rate <5%). The genes identified in these experiments represent a wide range of coding and genomic sizes. Consistent with a mostly regulatory role of their products, the level of transcript for these genes is generally modest based on microarray analysis of tachyzoite gene expression (see summary in Figure S2
One of the mutants complemented in this study (11C9) was the subject of an earlier cDNA-based complementation experiment that yielded a suppressor (TgXPMC2) of the genetic defect in this mutant [32
]. Multiple complementation experiments of 11C9 using the cosmid library have repeatedly identified a single Toxoplasma
gene, 50.m03077 (), which was likely underrepresented in the previously used cDNA libraries due to its large size and low expression level (<100 units average fluorescence intensity in tachyzoite microarrays for 50.m03077 versus ~15,000 units for GRA-1, which was the promoter source used to drive expression in the cDNA libraries [31
]). Likewise, we have not re-isolated TgXPMC2 by cosmid complementation. Because gene expression from cosmids relies on native regulatory regions, we believe there is a higher likelihood that the genes identified through this approach will represent the defective gene rather than a suppressor. In the three ts
mutants (109C6, VA-15, and FV-P6) where this question was examined so far sequencing and functional testing of the corresponding mutant alleles confirms this prediction. In summary, the cosmid system provides robust complementation for a broad range of genes. Furthermore, the identified loci can be readily biologically validated taking advantage of an extensive set of end-sequenced and tiled cosmids that provide essentially full genome coverage. The protocols and reagents developed in the course of this study should allow future forward genetic analysis of any essential aspect of parasite biology for which a mutant screen can be devised.
Tachyzoite growth rates differ dramatically among parasite strains and growth rate is a key virulence determinant in Toxoplasma
]. Despite the obvious importance of growth control, how the parasite regulates growth and cell division remains largely unknown. A series of defined biochemical controls and checkpoints regulating progression through one cell cycle phase to the next have been established for a variety of eukaryotic models [52
]. The catastrophic break down of cell cycle coordination observed in T. gondii
in the course of certain drug treatments has lead to the hypothesis that there might be significantly fewer cell cycle controls in this microrganism [29
]. By contrast, the phenotypic groups that have emerged from the collection of conditional growth mutants described in this study support the notion of specific mechanisms and checkpoints. For example, two ts
mutants were isolated that reversibly arrest in the G1 phase when shifted to 40°C (mutant 63H4 and 31F1). The presence of such a natural G1 checkpoint is further supported by the observation that end-stage differentiated parasite forms (sporozoite and bradyzoite) show a uniform haploid DNA content [28
], as do parasites that have been treated with the G1 phase inhibitor pyrrolidine dithiocarbamate [57
]. Tachyzoites released from this drug block, grow synchronously through at least two division cycles, indicating that pyrrolidine dithiocarbamate is likely acting on the same G1 checkpoint affected in our mutants.
Another important checkpoint controls entry into S phase, and in Saccharomyces
, this checkpoint (called START) also controls the initiation of spindle formation and budding [59
]. We have previously argued that the tachyzoite cell cycle likely has a similar checkpoint based on the observation that dNTP depletion arrests tachyzoite growth at the G1/S boundary (1N DNA content, centrosomes largely duplicated but not yet separated [22
]). We have isolated five ts
mutant 150B8) that display a very similar phenotype at the restrictive temperature, suggesting that the G1/S transition is an important restriction point in the tachyzoite cell cycle as it is in yeast.
Mitosis in Apicomplexa has several unique features: the nucleus remains intact, the intranuclear spindle(s) reside in a peculiar elaboration of the nuclear envelope the so called centrocone, and daughter cells are scaffolded as internal buds which develop in close proximity and likely under the control of the extranuclear centrosomes [1
]. There is significant evidence for the tight regulation of mitotic events in tachyzoites from three groups of mutants in our collection. These mutants arrest either in mitosis (11C9), display defects in chromosome segregation (5 mutants), or loose the coordination of karyokinesis with cytokinesis at various stages in the replication timeline (12 mutants).
Mitotic mutant V-A15 becomes both aneuploid and polyploid at the restrictive temperature and fails to initiate internal budding. The gene affected in this mutant is an NIMA-related serine/threonine kinases (Nek). This kinase family was first identified as essential for division in Aspergilus nidulans
] and its members have since been identified as cell cycle regulators throughout eukaryotes [62
], including protozoa [35
]. Neks have roles in microtubular dynamics in cilia, mitotic spindles and centrioles [63
]. Consistent with the known roles for NIMA-related kinases, the mutation in V-A15 leads to defects in the spindle apparatus (reflected in the loss of MORN1 organization) and this causes chromosome mis-segregation. Apicomplexa encode a family of related NEK proteins, and in P. falciparum
these genes have been shown to be expressed in a developmentally regulated fashion, making it likely that NEKs are critical to fine-tuning the cell cycle to different life-cycle stages and host cells.
At the non-permissive temperature mutants 42D6 and PO-B3 are promiscuous for nuclear reduplication leading to the formation of syncytial cells with multiple nuclei. Daughter budding is also abnormal in these mutants and uncoupled from the controls that ensure proper nuclear sorting into each daughter. Two distinct RCC1 domain proteins were found to rescue these mutants (25.m01896 and 72.m00409). In other eukaryotes, RCC1 domain proteins interact with Ran-GTPases to regulate spindle assembly as well as other mitotic progression controls through modulation of nuclear trafficking [67
]. Like mutant V-A15, mutant PO-B3 (and also uncoupling mutant 42D6) looses MORN1 organization at the restrictive temperature pointing to a potential spindle defect (data not shown). Overall the phenotype of this mutant, as well as several other members of the uncoupling class produced here (e.g.
42D6, 20C2, and 7A11), are similar to the abnormal daughter budding and the induction of unregulated nuclear replication associated with the disruption of the parasite spindle by pharmacological microtubule ablation [29
]. Collectively, these observations indicate that proper control over chromosome copy number and budding in Apicomplexa might critically rely on an intact intranuclear spindle or associated structures. In this context it is important to note that the unique centrocone structure that conducts the apicomplexan spindle into the nucleus appears to persist throughout the cell cycle at least in some Apicomplexa [20
]. This model is further supported by preliminary electron microscopy studies of mitotic mutant 11C9, which upon temperature arrest retains an intact spindle and early daughter scaffolds ([32
] and S. Halonen and M. White, unpublished data) and does not undergo nuclear reduplication as is seen in mutants 42D6 and PO-B3. Thus, these models predict that rather than the absence of cell cycle controls in the Apicomplexa, mitotic control mechanisms might require a strict physical context associated with the centrosome and/or the centrocone of the parasite nucleus. Breaking this physical context by drug treatment [29
], overexpression of a centrocone structural componenent [20
], or mutation (as in the uncoupling group of ts
mutants) results in catastrophic loss of regulation. There is considerable precedence for spatial control of cell cycle checkpoint proteins through compartmental exclusion as well as physical tethering of factors to the centrosomes and spindle structure [69
]. Strict compartmentalization of the cell cycle machinery could be a key to the cell cycle flexibility observed in these parasites. Further work is needed to validate this hypothesis. However, the large collection of mutants and genes identified in this screen provides an important pool of validated candidates for mechanistic dissection. Studies that link parasite cell cycle control to the adaptation to specific host cell niches and pathogenesis will be of particular interest.