Structure and function of an atypical homodimeric actin capping protein from the malaria parasite

Apicomplexan parasites, such as Plasmodium spp., rely on an unusual actomyosin motor, termed glideosome, for motility and host cell invasion. The actin filaments are maintained by a small set of essential regulators, which provide control over actin dynamics in the different stages of the parasite life cycle. Actin filament capping proteins (CPs) are indispensable heterodimeric regulators of actin dynamics. CPs have been extensively characterized in higher eukaryotes, but their role and functional mechanism in Apicomplexa remain enigmatic. Here, we present the first crystal structure of a homodimeric CP from the malaria parasite and compare the homo- and heterodimeric CP structures in detail. Despite retaining several characteristics of a canonical CP, the homodimeric Plasmodium berghei (Pb)CP exhibits crucial differences to the canonical heterodimers. Both homo- and heterodimeric PbCPs regulate actin dynamics in an atypical manner, facilitating rapid turnover of parasite actin, without affecting its critical concentration. Homo- and heterodimeric PbCPs show partially redundant activities, possibly to rescue actin filament capping in life cycle stages where the β-subunit is downregulated. Our data suggest that the homodimeric PbCP also influences actin kinetics by recruiting lateral actin dimers. This unusual function could arise from the absence of a β-subunit, as the asymmetric PbCP homodimer lacks structural elements essential for canonical barbed end interactions suggesting a novel CP binding mode. These findings will facilitate further studies aimed at elucidating the precise actin filament capping mechanism in Plasmodium. Supplementary Information The online version contains supplementary material available at 10.1007/s00018-021-04032-0.

CPs are ubiquitous proteins, usually present in eukaryotes as multiple isoforms [24]. CPs play a crucial role in actin dynamics by binding to the fast-growing barbed end of filamentous actin (F-actin) with high affinity in a Ca 2+ -independent manner [25], limiting protomer exchange to the pointed end. Metazoan CPs characterized to date nucleate polymerization, decrease the elongation rate in preseeded systems, block dilution-induced depolymerization from barbed ends, and increase the critical concentration to the level of the pointed end [24]. CPs are present in various cytoskeletal structures, such as lamellipodial protrusions [26], dynactin [27], and in the sarcomere, linking microfilaments to the Z-disks [28]. CPs are essential for human and zebrafish morphogenesis [29] and belong to the core set of proteins needed to reconstitute actin-based motility in vitro [30]. The average cytosolic concentration of CP in eukaryotes is in the range of 0.5-1.5 µM [31,32], which, considering the high affinity and 1:1 stoichiometry of CP towards actin filaments, leads to a high number of constantly capped barbed ends in vivo [32]. However, formation of free barbed ends is essential for rapid actin network assembly [33,34], and control of CP expression levels is required for optimal actin-based cellular functions [30,31]. Thus, the mechanism of capping/uncapping is modulated by various external factors. Steric and allosteric regulators of CPs include polyphosphoinositides (PIPs), V-1/myotrophin, and CARMIL proteins, while indirect barbed end competitors include formins and ENA/VASP proteins [24].
CPs are comprised of CPα and CPβ subunits, each with a molecular weight of 32-36 kDa [24]. The individual subunits are conserved in most eukaryotes, but the sequence identity between the subunits is typically low [24]. Remarkably, CPα and CPβ share a strikingly similar fold, resulting in a quaternary structure where the two subunits take up a compact arrangement with a pseudo twofold symmetry [35]. Even though subunits and isoforms are expressed at different levels during different stages and cell types [36], the individual subunits are largely insoluble and non-functional when expressed alone in vitro, indicating that they only exist as heterodimers [37].
All currently available structures of heterodimeric CPs have a characteristic shape of a stipitate mushroom [27,35,38]. Three N-terminal antiparallel helices form the "stalk" domain flanked by a β-stranded "globule" domain of each monomer. The "cap" is comprised of a well-ordered arrangement of two 5-stranded antiparallel β-sheets, crested by a backbone of four helices formed conjunctly by the subunits. Emerging from the cap structure are the C-terminal "tentacles" of each subunit (termed α-and β-tentacle, respectively) containing a longer flexible loop region and an amphipathic helix [35]. The barbed-end binding of canonical CPs is largely reliant on an electrostatic interaction between a positively charged basic patch of the CP and a negatively charged acidic patch of the last two actin protomers of the filament [39][40][41]. The basic patch is mainly formed by a localized group of residues on the cap and the α-tentacle. The β-tentacle locks the complex by binding to the hydrophobic pocket of the last actin subunit. However, its main role is to exclude other regulators from binding to this pocket, thus regulating branched actin network assembly [41].
Contrary to the majority of eukaryotes, Plasmodium spp. encode only one isoform of each CP subunit [42]. Individual PbCP subunits have evolved distinct biological functions. The heterodimeric PbCPαβ is expressed in the motile extracellular zoite and ookinete stages and is able to reduce the average filament length of heterologous nonmuscle actin (β-actin) in vitro [22]. PbCPαβ also readily binds to homologous PfActI filaments reducing their pelletable form [43]. Plasmodium knowlesi CPαβ caps heterologous skeletal muscle actin (α-actin) in vitro in a Ca 2+ -independent, but PIP 2 -dependent manner, thus reducing filament length [44]. Disparate from the other subunit, PbCPβ is upregulated in the insect vector stages and is essential for the locomotion and salivary gland invasion of the highly motile sporozoite stages [22], possibly in its heterodimeric PbCPαβ form [23]. PbCPα, on the other hand, is moderately upregulated in the blood-stage schizonts [23], in which PbCPβ is phenotypically silent. PbCPα is essential for asexual blood stage replication, and loss of it could be rescued by the cognate PfCPα in blood stages, but not in the insect host or in the absence of the α-tentacle domain [23]. Interestingly, a mixed complementation of the CP subunits by P. berghei and P. falciparum genes was incapable of alleviating the defective mosquito stage. Only a double complementation of PfCPαβ was able to rescue the phenotype. In vitro, PbCPα forms homodimers (PbCPαα), capable of capping homologous PfActI [43] and heterologous β-actin filaments also in the absence of the α-tentacle [23]. These results suggest that the insect cell stages are regulated by the heterodimeric form, while the pathogenic blood stages are governed by PbCPα, perhaps in a homodimeric form [23,43].
Here, we present the first crystal structure of a homodimeric CP from the rodent malaria parasite P. berghei. While retaining some similarities, the structure demonstrates critical differences compared to canonical heterodimeric CPs. We complement our structural findings with biochemical characterization, showing that the homo-and heterodimers have distinct functions that differ from the canonical CP heterodimer (CapZαβ) and are specific to the Plasmodium actin filaments.

Crystal structure of the PbCPαα homodimer
To gain structural insight into the unusual homodimerization of PbCPαα, we determined the structure of a C-terminally truncated PbCPαα homodimer (PbCPαα ΔC20 ), which was the only version of the parasite CPs that produced well-diffracting crystals, to 2.2-Å resolution ( Fig. 1; Table S1). The two PbCPα ΔC20 subunits form a structure that is less compact and intertwined than the canonical heterodimer ( Fig. 1a, b), verifying earlier small-angle X-ray scattering (SAXS) and thermal stability analyses [43]. The quaternary structure closely resembles the mushroom shape of metazoan CPs, despite the low sequence conservation (Figs. S1, S2). Albeit being a homodimer, the structure is not completely symmetric. Thus, we will refer to chain A of the crystal structure as PbCPα 1 ΔC20 and chain B as PbCPα 2 ΔC20 . Superposition of the two subunits   (Fig. 1c). In canonical CP heterodimers, the stalk domain of the β-subunit pivots relative to the core structure, to allow a denser packing. In PbCPα ΔC20 , the presence of two identical stalk domains disrupts the formation of a compact structure (Fig. 1d). In both canonical heterodimers and the PbCPαα homodimer, the interactions between the stalk domains of the subunits are mainly hydrophobic. In CapZαβ, the helices in the β-subunit are shorter (13, 13, and 8 residues) than in the α-subunit (15,13, and 20 residues). In the PbCPαα ΔC20 homodimer, helices in PbCPα 1 ΔC20 are 14, 12, and 16 residues and in PbCPα 2 ΔC20 15, 15, and 16 residues long. Especially the length of H3, which leads to the globule domain, seems to affect the orientation of the entire stalk domain. In addition, there are other factors contributing to the symmetry breaking, discussed below.
PbCPα contains a sequence insertion, not present in metazoan CPs or PbCPβ (Figs. S1, S2). Similar insertions are found in several other apicomplexan ABPs [13]. Their functions are often unknown but could be involved in parasite-specific protein-protein interactions or immune evasion [14,45]. The 23-residue Plasmodium CPα-specific insert is located between the globule and cap β-sheets (Ser136-Ala159). In PbCPα 1 ΔC20 , this loop protrudes from the structure (Fig. 1a) making space for the long H5 helix (Ile243-Arg281), which would be followed by the α-tentacle in the full-length protein. The base of the insert is stabilized by a partial disulfide bond between Cys158 and Cys179 in the PbCPα 1 ΔC20 subunit but not in PbCPα 2

ΔC20
. PbCPα 2 ΔC20 is notably more disordered, due to the reduced number of intersubunit and crystal contacts and has a higher average B-factor than PbCPα 1 ΔC20 (Figs. S1, S3), similarly to the crystal structure of CapZαβ [38]. Consequently, the Plasmodium-specific insert is not visible in PbCPα 2 ΔC20 . The H5 helix of PbCPα 1 ΔC20 is broken at His267 and turns back towards the same subunit in a β-hairpin-like structure (Leu268-Leu286), instead of extending into PbCPα 2 ΔC20 , resulting in very different C termini in the subunits (Fig. 1a, c). Interestingly, the first short β-strand of this hairpin motif ends with Ala270, which is the previously identified C-terminal degradation point in full-length PbCPαα [43]. This implies that the C-terminal hairpin-fold of PbCPα 1 ΔC20 might be biologically relevant and not an artifact arising from the α-tentacle-truncated PbCPαα ΔC20 construct.

Homodimeric PbCP assembly lacks canonical interaction sites
The observed structural differences result in a non-canonical dimeric CP structure and an example of a rare asymmetric homodimer among proteins in general [46]. It is unclear, however, how these deviations would impact the possible barbed-end binding mode of PbCPαα ΔC20 . To understand whether a canonical binding mode is compatible with homodimeric CPs, we used molecular docking to generate a model of a PbCPαα ΔC20 -capped PfActI filament (Fig. 2a). The expression of Arp1 in P. berghei [47] could implicate the existence of PbCP-capped Arp1 filaments in Plasmodium, in a similar manner as the metazoan dynactin [27]. Therefore, we used the cryo-EM structure of CapZαβcapped Arp1 filament in Sus scrofa dynactin [48] as a basis for modeling. During the preparation of this manuscript, a high-resolution cryo-EM structure of mammalian β-actin capped with CapZ was published [41]. This has also been used for comparisons to our model (Figs. S4, S5).
The electrostatic potential surfaces at the actin filament barbed end (Fig. 2b) and CPs (Fig. 2c) reveal the absence of canonical charged interaction points in the Plasmodium proteins ( Fig. S4), even though residues involved in forming these interaction surfaces are conserved (Figs. S1, S2, S6). In PfActI, flanking residues mask the apparent charge of the patch (Figs. 2b, S4a). Differences in surface electrostatic potential [49] could also play a large role in the divergent nucleation and polymerization properties of PfActI [6]. Residues of the basic patch in PbCPαα ΔC20 are more buried in the structure, not in close proximity to each other, and the slight asymmetry of the homodimer does not compensate significantly for the loss of contributing residues of PbCPβ. The putative basic patch spans a larger area in the PbCPαβ model, only loosely resembling the arrangement in CapZαβ (Fig. S4b). Possibly due to the absence of CARMIL proteins in Plasmodium [12], the positively charged CARMILbinding pocket of canonical CPs [50] is less prevalent in the PbCPαβ model ( Fig. S4b) and absent in PbCPαα ΔC20 (Fig. 2c). Upon binding the barbed end, the β-tentacle of CapZαβ locks the complex by binding the hydrophobic pocket of the terminal actin subunit [39][40][41]. In our model, the Plasmodium-specific insert of PbCPα 1 ΔC20 is close to the expected position of the β-tentacle (Figs. 2d, S5). This insert shares many possible hydrophobic or electrostatic interaction points with the β-tentacle. These residues are mainly conserved across Plasmodium spp. (Fig. S1), suggesting a possible role for the insert in barbed-end binding in the absence of the β-tentacle. Mapping per-residue sequence conservation of Plasmodium CPα sequences on the structure of PbCPαα ΔC20 reveals that core residues, especially those involved in the intersubunit surface, are highly conserved, suggesting that CPαα homodimers likely exist in other Plasmodium spp. Canonical binding partner interaction points present in CapZαβ lack localized sequence conservation in PbCPαα ΔC20 (Figs. 2e, S1).
Comparing our barbed end model to the CapZαβ-capped β-actin filament [41] shows that the predicted relative orientation of PbCPαα ΔC20 is in good agreement with the Fig. 2 Atypical structural properties of PfActI and PbCPαα ΔC20 suggest a divergent barbed-end binding mode. a Model of PfActI filament (green to beige colored surfaces, PDB ID: 5OGW [7]) capped by PbCPαα ΔC20 (blue cartoon) in a canonical arrangement. b Electrostatic potential surface of the last (b) and penultimate (b −1 ) protomer of PfActI (left) and Arp1 (right, PDB ID: 6F1T [48]) filament barbed ends. c Electrostatic potential surface of PbCPαα ΔC20 (left) and CapZαβ (right, PDB ID: 6F1T). d Hydrophobic pockets of the last subunit of PfActI (left) and Arp1 (right) filament presented as green surfaces. The Plasmodium-specific insert, cap sheets, and cap heli-ces of PbCPαα ΔC20 (left) are colored purple, red, and teal, respectively. The CapZαβ (right) α-and β-subunits are shown in light and dark gray, respectively. Sequence alignment (bottom) of the Plasmodium-specific insert (PbCPa_ins) with the β-tentacles of P. berghei (PbCPb_C: A0A509AQN8) and S. scrofa CP (SsCPb_C: A9XFX6), grouped and colored respective to a Risler matrix, using the ESPript convention. e Structure of PbCPαα ΔC20 (left) and CapZαβ (right) with the surface colored by ConSurf residue conservation scores from low (cyan) to high (magenta) orientation of CapZ in the experimental structure (Fig. S5). The PbCPα 1 ΔC20 subunit is located 5-6 Å closer to the last actin subunits due to the absence of the tentacle domains and the compatible orientation of the Plasmodium-specific insert. The N-terminal stalk domains align surprisingly well, while the globule of PbCPα 2 ΔC20 is twisted compared to its CapZβ counterpart, as discussed previously. Interestingly, both dynactin-and β-actin-bound CapZαβ structures describe a more pronounced kink of the H5 helix, just before the α-tentacle, towards the central β-sheet (Fig.  S5c). The β-sheet in the β subunit is also slightly distorted to adjust for the H5 helix orientation. These differences stem possibly from the conformational changes required for barbed-end binding [51,52]. The unique β-hairpin structure in PbCPα 1 ΔC20 corresponds well to the location of this distortion (Fig. S5d) and could potentially serve a similar function.
The differences in the CARMIL-binding pocket instigated us to take a deeper look into other CP regulators in Plasmodium. While the majority of these are absent from Plasmodium [12], we cannot exclude the possibility of other, so far uncharacterized, regulators. Our search for V-1/myotrophin homologs and the CP-binding and uncapping motif of CAR-MIL proteins among known Plasmodium transcripts, or for the S100B interaction motif in Plasmodium CPs resulted in no clear hits. To date, PIP 2 is the only described direct CP regulator in P. knowlesi with HSC70 mentioned as a binding partner uninvolved in the capping activity [44]. Residues involved in canonical interactions with protein regulators are generally not conserved in Plasmodium CPs (Figs. S1, S2), with the exception of several basic residues (Lys278 and Lys282 in PbCPα, Arg233 in PbCPβ), which are involved in canonical F-actin interaction and PIP 2 binding [39,40,51].

Conservation of the CP domain structure
Canonical CP subunits share a strikingly similar fold, despite low sequence identity. Compared to the heterodimeric CPs, the homodimeric PbCPαα ΔC20 has an increased surface area but a significantly decreased dimer interface area (Table S2). Expectedly, the canonical CP heterodimers are more similar to each other than to the Plasmodium CP homodimer (Table S3). Despite the low sequence identity, both PbCPαα ΔC20 subunits are strikingly similar to canonical CPα isoforms (Table S4). This confirms our previous findings [43] and is further substantiated by the identifiable hits of CP-related CATH domains (1.20.1290.20 and 2.40.160.80) in the PbCPαα ΔC20 structure. The identified domains in PbCPαα ΔC20 superimpose well with canonical CP subunits (Table S5), suggesting the importance of conserved structural elements in the actin filament capping function among broad taxa, despite the lack of sequence conservation. Indeed, during the erythrocytic stages of P. berghei, where PbCPβ is phenotypically silent and homodimeric PbCPαα takes over functionally, PbCPα can be complemented by its PfCPα cognate despite the fact that they only share 52% sequence identity [23].

Plasmodium CPs form similar-shaped homoand heterodimers in solution
Our previous SAXS studies [43] agree well with the crystal structure suggesting that the homodimeric structure is a native conformation and not an artifact resulting from crystal packing or the absence of the α-tentacle domains. To gain further insight into the homo-and heterodimerization of PbCPs, we carried out further SAXS and homology modeling studies based on the crystal structure of PbCPαα ΔC20 .
PbCPαα ΔC20 , PbCPαα, and PbCPαβ all form similar pseudo-symmetric dimers in solution (Fig. 3). The SAXS data indicate folded, globular proteins containing flexible parts, with PbCPαβ being somewhat more compact than the homodimers (Fig. 3a, b). Interparticle distances are similar for all PbCPs, with the heterodimer showing slight bimodality (Fig. 3c). Modeling into the SAXS data suggests that PbCPαβ forms a more canonical tight structure, whereas both subunits of the homodimers adopt a loose arrangement, similar to the PbCPαα ΔC20 crystal structure (Fig. 3d). The orientation of the Plasmodium-specific insert and the tentacle domains of both subunits cannot be reliably deduced from the SAXS data, indicating that they are disordered in solution, which is characteristic of the tentacle domains in canonical CPs as well [35,38,53].

PbCPs control actin polymerization in a non-canonical manner
CPs typically block the barbed end, increasing its apparent critical concentration (Cc app ) to that of the pointed end [24]. PbCPs increase the supernatant fraction of polymerized PfActI in pelleting assays [43], which could be explained either by filament shortening or by limited depolymerization due to the increase of Cc app . However, in dilution series of pyrene-labeled PfActI filaments, none of the PbCPs affected the fluorescence signal of PfActI, even at high stoichiometric ratios (Fig. 4a). Similar behavior was observed with heterologous α-actin (Fig. 4b), despite PbCPs being able to modulate α-actin [44] or β-actin polymerization [22,23]. Furthermore, gelsolin, which is a major barbed end capper [54] absent from Plasmodium spp. [11], does not affect the Cc app of PfActI filaments (Fig. S7) or PfActI depolymerization dynamics [6]. The decreased amount of pelletable actin in the presence of PbCPs and the lack of effect on the fluorescence signal in the pyrene assay are likely due to a decrease in filament length to oligomers that do not sediment. Another possibility would be CP binding to the barbed end of actin in either monomeric or dimeric form, thus increasing its fluorescence signal similarly to polymerization.
In higher eukaryotes, CPs nucleate filaments, abolishing the lag phase, but block subunit exchange at the barbed end [24], reducing both the initial elongation velocity and steady-state filament mass. Contrary to expectations, PbCPs increased PfActI elongation rates and considerably raised the steady-state fluorescence levels, indicating elevated amounts of actin in non-monomeric forms (Fig. 5). The effect of PbCPs on nucleation was ambiguous, due to the nature of PfActI polymerization curves, which lack a pronounced lag phase [6]. The heterodimeric CPs did not modulate the steady-state fluorescence levels of PfActI to the extent seen with the PbCP homodimers, which could also be observed as slope differences in the Cc assays (Fig. 4). Interestingly, all PbCPs seem to block the barbed end of α-actin, similarly to CapZαβ (Fig. S8). However, they do not display any notable nucleation activity, and unlike heterodimeric CPs, the PbCP homodimers increase the initial fluorescence levels. PbCPαα ΔC20 displays an only slightly diminished capping effect, despite the absence of the tentacle domains [39,40]. Despite its in vivo indispensability [23] the reduced importance of the Fig. 3 PbCP homo-and heterodimers have a canonical shape in solution. a Experimental scattering curves of PbCPαβ, PbCPαα, and PbCPαα ΔC20 (red, green, and blue open circles, respectively) and the respective fits of DAMMIN (white dashed lines) and SREFLEX (black line) models to the data. The χ 2 values of the different models to the data are indicated. Weighted residuals of the SREFLEX model fits are denoted in the lower graph (lines colored respectively), where Δ∕ = I exp (q) − cI mod (q) ∕ (q) . b Dimensionless Kratky plot (colored similar to a). c Real-space distance distribution plot (colored similar to panel A). d Cartoon models of PbCPs (colored as in a, with chain A in lighter color) and the corresponding SAXS envelopes as surface representations. The tentacle domains of each structure are colored black. Experimental data for PbCPαα ΔC20 for comparison were used from a previous publication [43] Plasmodium α-tentacle domain in PbCPs has been suggested before [43].
Major differences between the homo-and heterodimers are seen in the steady-state PfActI fluorescence levels (Fig. 5a). We also measured length distributions of PfActI filaments polymerized with and without homo-and heterodimeric CPs from negative-stained electron micrographs (Fig. 6). PfActI alone forms mainly short, non-helical structures with a sporadic presence of long helical filaments [8]. Upon incubation with CPs, the long filaments  b Polymerization of PfActI on capped, preformed homologous filaments in the absence and presence of increasing concentrations of PbCPαβ, PbCPαα, PbCPαα ΔC20 , and CapZαβ. The PfActI-only polymerization curve is shared between preseeded measurements and the fluorescence levels are comparable. RFU relative fluorescent unit were completely removed (Fig. 6a) and the shorter species became more abundant. The average filament length is reduced by half (Fig. 6b, c) as also seen for canonical actins [22,23,44]. In line with the polymerization assays, the reduction in filament length is more pronounced with the homodimeric PbCPs, and the tentacle domain is at least partly dispensable for the PbCP function.
PfActI filaments are highly dynamic and unstable [5][6][7] with high disassociation rates at both ends [9]. Upon dilution below their Cc app , filaments decompose rapidly, albeit slower than α-actin if filament end concentration is taken into account [6]. To investigate the barbed-end blocking efficiency of PbCPs, we followed the disassembly of fluorescently labeled actin filaments. In contrast to canonical CPs [24], both homo-and heterodimeric PbCPs facilitate the depolymerization of both PfActI and α-actin filaments (Figs. 7, S9). The increased velocity can be partly attributed to an increased filament end concentration [6], caused representation of the filament distributions for PfActI alone or capped with CPs. PfActI filaments longer than 300 nm are displayed separately. Boxes encompass 80% of the distribution around the median (horizontal bars) and whiskers extend to the 1st and 99th percentiles.
The mean values are shown as squares. Asterisks represent significant difference (***p < 0.001) between the medians when compared against the smaller PfActI species (98.5% of measured PfActI filaments) using Mann-Whitney tests by shortening of the filaments (Fig. 6). However, this may not be the only explanation. Interestingly, CapZαβ also increased the rate of PfActI depolymerization (Fig. S9), suggesting that the inherent instability of PfActI filaments [8] also modulates the activity of actin regulators. This is in line with CapZαβ increasing the initial elongation rate of PfActI in the polymerization assays (Fig. 5a).

Discussion
Owing to the intertwined and complementary nature of the canonical CPα and CPβ subunits, both are usually required to reconstitute a functional CP in vitro [37,55]. Thus, the presence of a homodimeric CP in Plasmodium is surprising [23]. Sequence conservation mapped on Plasmodium CPα subunits reveals highly conserved residues located mainly at intersubunit interfaces in the PbCPαα ΔC20 structure. This suggests that CPαα homodimers likely are conserved across the different Plasmodium species, as underlined by previous PfCP complementation experiments [23]. So far, we have not been able to detect any evidence of CPββ homodimers, although the presence of such or CPα or CPβ monomers in the parasite cannot be ruled out. It is notable that Drosophila melanogaster CPαβ subunits are also encoded by single genes with independent in vivo functions implicated, but the protein still remains an obligate functional heterodimer [56,57]. PbCPαα is an example of an asymmetric homodimer, not very common among proteins [46]. The twofold rotational symmetry of the CP dimer is broken by the disrupted interdimer interface due to the steric occlusion of the stalk domains. The lack of similar structural complementarity in PbCPαα ΔC20 as in canonical CP heterodimers must stem from the limitations set by the identical amino acid sequences of the subunits. While the protruding loop of the Plasmodium-specific insert does not impact the conserved fold of the cap structure, the helical backbone of the dimer is distorted by the presence of a C-terminal β-hairpin motif of PbCPα 1 ΔC20 . Interestingly, the previously identified C-terminal Ala270 degradation point of PbCPαα [43] is located on the first β-sheet of this hairpin motif. This further supports our hypothesis that this cleavage may be biologically relevant by alleviating any residual steric incompatibilities upon homodimerization. The deviations from the homodimeric symmetry perturb the canonical binding sites for actin [39,41], PIP 2 [58], and other CP regulators [51,52,59]. In line with this, we could not identify any regulatory proteins for CP in the Plasmodium genome. CPs characterized to date cap filament barbed ends through a highly conserved mechanism [27,39,41]. They play critical roles in regulating actin dynamics by increasing Cc, nucleating filament formation, and decreasing F-actin elongation and disassembly rates [24]. The actin binding of homodimeric PbCP in the absence of canonical barbedend binding motifs, however, challenges our current understanding of parasite actin filament capping. Despite large structural differences, homo-and heterodimeric PbCPs affect filament kinetics of both parasite and canonical actins. Also PfActI differs greatly from α-actin in terms of filament structure [5,7], dynamics [6,8] and barbed-end electrostatics. Furthermore, C-terminally truncated PbCPαα retains its ability to bind both β-actin [23] and PfActI filaments [43]. This implies that actin filament capping by PbCPαα is independent of the α-tentacle and the basic patch. The Plasmodium-specific insert contains several highly conserved residues and could compensate for the missing conventional interaction points. The presence of two of such inserts and a non-canonical cap structure indicates an atypical barbed-end binding mode and maybe a distinct functional role for the CP homodimers, as compared to the canonical or parasite CP heterodimers [22,23,43].
Both homo-and heterodimeric PbCPs increase the amount of PfActI present in the soluble fraction after pelleting the filaments [43]. Here, we show that this is due to filament shortening and not due to an increase in the steadystate G-actin concentration, thus explaining the inability of PbCPs to raise the Cc app . While homo-and heterodimeric PbCPs regulate actin kinetics in vitro in a similar manner, it seems that the homodimers may also be able to stabilize or sequester short, non-nucleating structures, perhaps lateral dimers, which are a significant species in PfActI at least in vitro [6]. Such a function could be facilitated by the presence of two Plasmodium-specific inserts in the homodimer. The sequestration of dimers would explain the increased steady-state and initial fluorescence levels in polymerization assays and the disappearance of filaments. It is possible that PbCPs cannot completely block the barbed end or that the Cc for PfActI is equal at both filament ends, as proposed by a similar k off at each end [9]. Subunit exchange at the CP-bound barbed end would suggest a "wobbly" capping behavior, found in processive barbed-end cappers [24], such as formins [54]. The existence of flexible domains in PbCPs and their positive effect on filament assembly correspond well with the properties of a wobbly capper.
PbCPs appear to lack prominent nucleating capabilities. Thus, the acceleration of actin depolymerization by PbCPs could be explained by severing. However, such mode of action has been disproven for CapZαβ [50,60,61]. Another explanation could stem from the innate fragility of PfActI filaments and their propensity to fragmentation [8]. Collision of PbCPs with the barbed end could promote severing events and depolymerization instead of forming a stable cap structure. After such an interaction they could dissociate or remain weakly bound to the filaments. This hypothesis is supported by the inability of CapZαβ to co-sediment with PfActI filaments, yet still increasing the soluble actin fraction [43] and reducing the average filament length.
Intracellular eukaryotic parasites undergo reductive genome evolution towards a streamlined set of proteins [62]. A significantly reduced repertoire of ABPs in Plasmodium has given rise to atypical functions for many of the proteins [13]. One outcome of this adaptation may be the emergence of a homodimeric form of CP [23,43]. Motility and host cell invasion of Plasmodium parasites rely on rapid turnover of short, unbranched filaments [63]. As PfActI is less polymerization competent and less stable than canonical filaments [5], stringent F-actin capping might be too restrictive for the parasite motility requirements. Wobbly capping by PbCPs that promotes turnover and reduces the length of PfActI filaments may be suited for the parasitic lifestyle. Despite the functional redundancy in vitro, PbCPαα and PbCPαβ cannot complement each other in vivo [23]. Thus, the adaptation of the homo-and heterodimeric forms to the cellular niches of their respective insect and vertebrate hosts stems from their inherent structural differences.
Heterodimeric PbCP has two different interfaces with the last two actin protomers at the barbed end, which could enable it to selectively bind to filaments but not to symmetric dimers. PbCPαβ is required for sporozoite motility and salivary gland invasion [22]. The quick utilization of actin monomers during these processes is possibly enabled by the reduced dimer-binding capability of PbCPαβ as compared to PbCPαα. The β-tentacle of CapZαβ is crucial for the controlled growth of a branched actin network [41]. In the absence of the Arp2/3 branching complex and nucleation promoting factors [12], Plasmodium only has short linear actin filaments. Thus, without the need for a such control element, PbCPβ has evolved a divergent C-terminal domain. While the conserved hydrophobic residues implied in a canonical barbed-end interaction are still present, it also contains additional stretches of residues, possibly enabling looser capping of the filaments.
Homodimeric PbCP contains two nearly identical interaction surfaces and might thus be more suited for binding two actin monomers, like the active form of gelsolin [64]. As gelsolin is missing from the Plasmodium genome [42], a similar activity could have evolved in PbCPαα. The proposed dimer binding by PbCPαα could play a role in restricting the number of filaments during the non-motile erythrocytic stages, during which the homodimers are indispensable [23]. In the absence of the Arp2/3 complex or gelsolin, formin remains the only characterized actin nucleator in Plasmodium [17]. Formins are localized at the moving junction (MJ) during erythrocytic invasion by non-motile merozoites [65]. As the correct formation and stringent cytoskeletal control of the MJ is critical [66], the wobbly capper formins could partially compensate for the phenotypic lack of PbCPβ [22]. The existence of an actin:CP:formin ternary complex [67] supports this hypothesis.
Based on our kinetic data, PbCPs have at least an order of magnitude lower affinity towards PfActI filaments than CapZαβ to α-actin. Left unregulated, canonical CPs would constantly cap actin filaments, completely arresting cell motility [32]. Due to the reduced set of CP regulators, PbCPs might have evolved a low-affinity wobbly capping behavior, allowing filament turnover without the need for external modulating agents. Without such regulators in the genome, canonical regulator interaction sites are also missing. The selfsufficiency of PbCP regulation is further underlined by the reduced intersubunit interface area of PbCPαα ΔC20 together with the low thermal stability of the PbCPs, and a broader unfolding peak of PbCPαβ [43]. These may reflect subtle temperature-induced structural changes, possibly relevant for the transitioning of the parasite between the cold-blooded insect host where PbCPβ is upregulated [68], and the warm-blooded vertebrate host where the PbCPαα form is relevant [23].

Concluding remarks
CPs are essential regulators of the cytoskeleton and conserved among metazoans as heterodimeric proteins, which are important, e.g., for lamellipodial protrusions at the leading edge of crawling cells [26]. Apicomplexan parasites display fast gliding motility without changes in cell shape [69]. It seems that due to different cytoskeletal requirements in different lifecycle stages, these parasites have adopted a homodimeric form of CP [43] as part of their limited repertoire of ABPs [13]. In contrast to canonical CPs [24], homoand heterodimeric PbCPs facilitate the rapid turnover of the dynamic Plasmodium actin filaments supporting the unusual actomyosin machinery of the parasite.

Structural bioinformatics
Model of PbCPαα ΔC20 capped PfActI filament was prepared by aligning 5 protomers of PfActI filament (PDB ID: 5OGW [7]) and PbCPαα ΔC20 to the CapZαβ capped barbed end of the Arp1 filament in dynactin (PDB ID: 6F1T [48]) using TM-align [76]. Gaps in the structure of PbCPαα ΔC20 were modeled in using SWISS-MODEL. The relative orientation of PbCPαα ΔC20 was refined with RosettaDock [77], while residues involved in the interface were energy minimized with UCSF Chimera. Electrostatic potential surfaces were calculated using APBS [78]. EMBOSS Matcher [79] was used for local sequence alignments and ESPript [80] for their visualization. 21 non-redundant Plasmodium CPα sequences from PlasmoDB were analyzed, with conservation score subsequently mapped on the structure of PbCPαα ΔC20 using ConSurf [81]. CapZαβ was assessed similarly, using automatically retrieved homolog sequences. Details of the surface area and r.m.s.d. calculations are found in Supporting Information. Database searches against homologs and motifs were carried out using the PlasmoDB database [82], Scan-Prosite [83], MyHits [84], and PATTINPROT [85].

Small-angle X-ray scattering
SEC-SAXS data of PbCPαα and PbCPαβ, at respective concentrations of 12 and 5.8 mg/mL, were collected on the B21 beamline at Diamond Light Source (Didcot, UK). Data processing, ab initio DAMMIN [86] modeling, and visualization were carried out as previously described [43]. Initial models of the homodimers were prepared by extending the crystal structure of PbCPαα ΔC20 using SWISS-MODEL [87]. The tentacle domain of PbCPα 1 was modeled using EOM [88]. A model of PbCPαβ was assembled from the modeled PbCPα 2 subunit and I-TASSER [89] modeled PbCPβ monomer using the structure of dynactin-bound S. scrofa CapZαβ (PDB ID: 6F1T [48]) as a template. The models were corrected against major steric clashes using Coot [74] and energy minimized using UCSF Chimera [90]. The PbCP models were split into domains and refined with normal mode analysis using SREFLEX [91].

Actin-CP interaction assays
Fluorescence-based polymerization assays were carried out in triplicate and analyzed as previously described [6] with minor modifications. Cc app was determined using a dilution series prepared from 10 µM actin polymerized together with 10 µM CP for PfActI, and 0.4 µM CP for α-actin. 4 µM total actin concentration was kept in the polymerization assays. 0.5 µM polymerized unlabeled PfActI was used as nuclei in the preseeded polymerization assays. In the depolymerization assays 4 µM PfActI was polymerized together with the CPs for 16 h at 20 °C before dilution in F-buffer below Cc app . The polymerization data were despiked before averaging. The depolymerization data were shifted to match the baseline of the actin control. No pointed end cappers were used in the assays to limit subunit exchange to the barbed end. 20 µM PfActI was polymerized with or without 25 µM CPs in F-buffer at 20 °C for 16 h. Prior sample application, the carbon-coated 300-mesh copper grids (Gilder, US) were glow-discharged for 1 min. After diluting PfActI to 1 µM with F-buffer, grids were immediately prepared using a previously reported protocol [8]. Negative stained grids were imaged using a Tecnai G2 Spirit microscope (FEI, US) operated at 120 kV with a pixel size of 0.592 nm. Micrographs were analyzed using the Ridge Detection plugin [92] of Fiji [93]. Particles smaller than 12 nm were excluded to omit bias caused by any unbound CP. Mann-Whitney tests were carried out using OriginPro (OriginLab, US).