daughter budding is a highly complex process that requires the coordinated assembly of numerous elements. We have recently demonstrated that the cytoskeleton that organizes the IMC is of even greater complexity than previously anticipated 
. We identified a family of 14 alveolin repeat intermediate filament-like proteins (IMC proteins) that showed a remarkably organized spatial and temporal assembly pattern for different family members. Our detailed analysis of the MORN1 null mutants strongly suggests that MORN1 is not required for initial bud formation. However, in the absence of MORN1 we do not detect any deposition of known components of the basal complex, which includes a subset of the IMC family residing in the basal complex (, , and S3
). The lack of the basal complex results in a failure to contract the basal end to taper the growing cytoskeleton as part of its maturation process. This provides direct evidence that the basal complex is involved in this constriction, with Centrin2 being the most likely candidate to provide the contractile force 
. As a result, daughter parasites are not separated from each other and form double-headed parasites with two apical ends. Remarkably, this does not prevent these Janus-headed parasites to initiate a next round of division, which is reflected in the duplication of their centrosomes () and the appearance of a population with a 3.5N nuclear content (). Such nuclear contents most likely represent parasites carrying two nuclei in the 1.8N arrest stage, which reflects a normal stage in the parasite's development cycle before undergoing cytokinesis 
. Interestingly, the parasites then can undergo additional rounds of budding, but again fail to constrict their basal complex. As a result, 4-headed parasites, and upon subsequent division rounds even 8- and 16-headed parasites appear ().
It is somewhat surprising there is no checkpoint for assembly or contraction of the basal complex before progressing toward disassembly of the mother's cytoskeleton, as this appears to be crucial for the formation of viable parasites. Therefore we questioned how this compares to well-studied model organisms. The late stages in Toxoplasma
division, i.e. maturation of the pellicle coordinated with mother disassembly is akin to the late stages in mammalian cells completing abscission. Although in mammalian cells the initial stages of cytokineses are mediated by an actin-myosin mediated cleavage furrow ingression, which compares with the basal complex constriction in Toxoplasma
, the final segregation is organized by a septin and membrane remodeling based machinery 
. Mutants in the abscission machinery do lead to conjoined mammalian cells 
. The lack of a checkpoint late in cytokinesis therefore is in line with the understanding of other systems. It is currently unclear whether T. gondii
harbors machinery analogous to the septin based machinery to complete cell division. An argument against such machinery is that even in normal parasite development abscission is not complete and parasites are often observed to share a cytoplasmic bridge at their basal end 
. Therefore it is more likely that Toxoplasma
breaks the small cytoplasmic bridge present after completion of cytokinesis either by mechanical stresses within the vacuole or alternatively, by pulling away in opposite directions upon egress. Such a mechanic abscission model has been shown for the related ciliates 
. However, in the multi-headed parasites the cytoplasmic bridge is too large to be disrupted by mechanical force.
While the effects of the loss of MORN1 on completion of cytokinesis are very strong, the formation and partitioning of other organelles and structures are largely intact with one prominent exception: the loss of apicoplast division (). The apicoplast is unique among plastids in that its division is linked to nuclear division by association with the centrosome 
. While this finding provided a strong model for segregation it did not explain organelle fission. Some authors argued for fission rings similar to those found in the chloroplast 
, however, no genes encoding homologs to the central component of that machinery FtsZ have been identified in Apicomplexa 
. Other authors suggested that cytokinesis alone and more specifically the MORN1 ring might be sufficient for apicoplast fission 
. Recent studies on the dynamin-related protein DrpA have demonstrated an absolute requirement for this protein for fission to occur. Detailed analysis of DrpA mutants showed that the MORN1 ring constricts the apicoplast during budding, but that this constriction is not sufficient for fission 
. Using the conditional MORN1 null mutant here we demonstrate that, while not sufficient, MORN1 is clearly required for apicoplast fission. Loss of MORN1 results in loss of apicoplast fission and this effect is manifest as early as 12 hours into ATc treatment. Why two constrictive machines? Biochemical work on the role of dynamin-like proteins involved in mitochondrial division suggests that these proteins preferentially assemble in regions of initial constriction 
. We propose a model in which the MORN1 ring provides this initial constriction and thus targets DrpA to the correct site. Activity of DrpA then completes fission. This model is fully consistent with the genetic requirement of both MORN1 and DrpA.
Upon complementation of the MORN1-KO with a full-length MORN1 encoding plasmid complete restoration of the wild-type phenotype is accomplished (). Efforts to complement with versions of MORN1 missing one or more MORN domains failed, indicating that all MORN domains are required for function. However, complementation could be established with a MORN1 mutant carrying a point mutation in a predicted palmitoylation site (). In this complementation MORN1.C4A was observed to polymerize into fibers from the centrocone. These data suggest that palmitoylation potentially plays a role in the centrocone, but not in basal complex assembly and function. Since formal proof of MORN1 palmitoylation is currently not available we cannot exclude that the amino acid change itself resulted in this effect. Furthermore, partial complementation was achieved with the construct missing the short linker between MORN domains 6 and 7 (). As shown in , contraction of the basal complex in this mutant was of lower efficiency. This suggests that this region could act as a potential flexible hinge required for remodeling of the MORN1 complex during basal complex contraction. Fitting a critical role for this linker region, expression of MORN1.DEL in a wild-type parasite background resulted in severe aggregate formation and was lethal, possibly through a mechanism trapping endogenous MORN1 in aggregates as observed with overexpression of full-length MORN1 (Fig. S4
A MORN1 null mutant recently isolated by Heaslip and colleagues shows a similar phenotype yet is viable in vitro 
. This contrasts with the loss of viability we observed for the conditional MORN1 null phenotype ( and S2
). In part this could be the result of the strategy employed by Heaslip et al to generate an unconditional MORN1 null mutant, which inherently selected for surviving parasites in the absence of a conditional allele. In this process there is a potential selective pressure for secondary mutations suppressing the most severe manifestation of the phenotype. This interpretation is strengthened by our observation in the complementation studies; when we performed phenotypic selection under ATc only, we obtained viable parasites for most deletion constructs. However, the inducible expression of the Myc2
-tagged MORN1 allele was lost and Myc2
-MORN was constitutively expressed. This indicates a genomic rearrangement or other mutations accrued in this phenotypic selection process.
In conclusion, we have shown that loss of MORN1 results in an early loss of apicoplast division and prevents the assembly of the basal complex. As a result parasites fail to fully separate while still re-initiating new rounds of division resulting in the formation of the striking multi-headed parasites. Finally, we showed that the linker region between MORN domains 6 and 7 potentially serves as hinge critical for flexibility in the MORN1 complex and is a critical region for correct functioning of MORN1. Future work will be needed to address a potential role of palmitoylation and/or other post-translational modifications in the function of MORN1.