Genome Segregation and Packaging Machinery in Acanthamoeba polyphaga Mimivirus Is Reminiscent of Bacterial Apparatus

ABSTRACT

Genome packaging is a critical step in the virion assembly process. The putative ATP-driven genome packaging motor of Acanthamoeba polyphaga mimivirus (APMV) and other nucleocytoplasmic large DNA viruses (NCLDVs) is a distant ortholog of prokaryotic chromosome segregation motors, such as FtsK and HerA, rather than other viral packaging motors, such as large terminase. Intriguingly, APMV also encodes other components, i.e., three putative serine recombinases and a putative type II topoisomerase, all of which are essential for chromosome segregation in prokaryotes. Based on our analyses of these components and taking the limited available literature into account, here we propose for the first time a model for genome segregation and packaging in APMV that can possibly be extended to NCLDV subfamilies, except perhaps Poxviridae and Ascoviridae. This model might represent a unique variation of the prokaryotic system acquired and contrived by the large DNA viruses of eukaryotes. It is also consistent with previous observations that unicellular eukaryotes, such as amoebae, are melting pots for the advent of chimeric organisms with novel mechanisms.

IMPORTANCE Extremely large viruses with DNA genomes infect a wide range of eukaryotes, from human beings to amoebae and from crocodiles to algae. These large DNA viruses, unlike their much smaller cousins, have the capability of making most of the protein components required for their multiplication. Once they infect the cell, these viruses set up viral replication centers, known as viral factories, to carry out their multiplication with very little help from the host. Our sequence analyses show that there is remarkable similarity between prokaryotes (bacteria and archaea) and large DNA viruses, such as mimivirus, vaccinia virus, and pandoravirus, in the way that they process their newly synthesized genetic material to make sure that only one copy of the complete genome is generated and is meticulously placed inside the newly synthesized viral particle. These findings have important evolutionary implications about the origin and evolution of large viruses.

INTRODUCTION

Genome delivery and packaging are fundamental processes in the life cycle of a virus. A large genome size (1) coupled with complex replication and assembly (2, 3) processes makes genome packaging in Acanthamoeba polyphaga mimivirus (APMV) an intriguing phenomenon. The molecular mechanisms governing genome packaging have yet to be explored in APMV as well as other well-studied nucleocytoplasmic large DNA viruses (NCLDVs), such as vaccinia virus. Recent tomographic and cryo-electron microscopic studies indicate that APMV has evolved a unique two-portal system for genome packaging and delivery (3, 4). The ATPase hypothesized to be responsible for genome translocation in APMV (and all the other NCLDVs and virophages) is related to prokaryotic chromosome segregation and packaging motors, such as FtsK/SPOIIIE/HerA (5). However, unlike prokaryotes, which usually make two copies of the chromosome during replication, APMV makes hundreds of copies of its genome within the inner confines of the viral factory (2), and these are likely to be catenated. The replicated genomes thus need to be disentangled and resolved from one another before packaging. Interestingly, along with packaging ATPase, APMV also encodes three putative recombinases and a putative type II topoisomerase, which are integral parts of the prokaryotic chromosome segregation and translocation machinery (6–8). Sequence and phylogenetic analyses of the three components of the packaging machinery (packaging ATPase, recombinase, and topoisomerase II) present in APMV suggest that mimiviral genome packaging has more parallels with the bacterial chromosome segregation mechanism than with other viral systems. In this report, we suggest that genome segregation and packaging in APMV are coupled. We also propose a model and compare it to prokaryotic genome segregation and packaging mechanisms.

MATERIALS AND METHODS

Data set retrieval.

Comparative genomic studies of FtsK-HerA superfamily proteins by Iyer et al. showed that NCLDV packaging ATPases share a common ancestry with the FtsK-HerA superfamily (5). Searches initiated with position-specific scoring matrices (PSSMs) PSSMs for orthologs of the bacterial FtsK and archaeal HerA superfamily retrieved packaging ATPases of NCLDVs with E-values in the range of 10⁻⁴ to 10⁻⁵ (5). On the basis of this and our own analysis, FtsK/SPOIIIE/HerA packaging ATPase protein sequences from representative organisms belonging to each of these groups, along with packaging ATPases from NCLDVs, were used to prepare the data set for phylogenetic analysis. Type II topoisomerase (L480) and recombinase (L103) of APMV were used for BLAST searches against the NCBI nonredundant (nr) database, and this retrieved bacterial and archaeal homologs. Data sets were prepared for L480 and L103 and their orthologs from bacteria and archaea. In the case of type II topoisomerase, a BLAST search also retrieved eukaryotic orthologs. A separate tree was also constructed to include these sequences.

Phylogenetic reconstruction.

The sequence data sets were subjected to multiple-sequence alignment with ClustalW in MEGA5, using default parameters, except that the multiple-alignment gap-opening penalty was changed to 3, and the gap extension penalty to 1.8, as recommended previously (9). All the alignments were analyzed manually for conserved domain architecture. Neighbor-joining method-based trees were generated in MEGA5 (9), using the P distance as a substitution model, with complete deletion of gaps/missing data (10). The bootstrap method was used as a test for phylogeny, applying 500 replications (11). The trees were drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree. The evolutionary distances were computed using the P distance method (12) and were reported as numbers of amino acid differences per site. The trees so obtained were manually analyzed for outliers, and sequences with bootstrap values of >50 were selected to reconstruct the phylogeny by using the same parameters as those mentioned earlier.

RESULTS AND DISCUSSION

Critical components of APMV genome packaging machinery.

Prokaryotic chromosome segregation and genome translocation machineries essentially have three critical components: an ATP-driven translocase motor, a recombinase, and a type II topoisomerase. Our analysis of NCLDV and virophage genomes revealed that while an FtsK-type packaging ATPase is invariably present in all NCLDVs, the other two components, recombinase(s) and topoisomerase II, are present in some but not all NCLDVs, suggesting the evolution of new strategies for the same theme (Table 1).

TABLE 1.

Components of genome packaging machinery in NCLDVs and virophages^a

NCLDV or virophage family	Example	Putative recombinase(s)	Putative topoisomerase type II	Putative packaging ATPase(s)	Genome topology
Ascoviridae	Spodoptera frugiperda ascovirus	ND	ND	ORF110*	Circular/linear
Asfarviridae	African swine fever virus	D345L	P1192R	B354L	Linear
Iridioviridae	Invertebrate iridiovirus 22	101L	141R	131R	Linear/circular
Marseilliviridae	Marseillievirus	ORF216	ORF173	ORF157	Circular
Mimiviridae	APMV	R80, L103, R771	R480	L437	Linear
Phycodnaviridae	PBCV NY2A	B083L, B381L, B715L	B781L	A392R	Linear
Poxviridae	Vaccinia virus	A22R	ND	A32L	Linear
Pandoraviridae	Pandoravirus salinus	ORF545	ND	18, 32, 527-528, 2214	Linear
Virophages	Sputnik	ORF10	ND	ORF3	Linear

^aND, not detected. *, unannotated, identified by homology.

Packaging ATPase.

The bacterial FtsK protein is a directional translocase motor required for faithful segregation of the newly replicated chromosome copies during cell division, while SPOIIIE is involved in transfer of chromosomes into spores in spore-forming bacteria such as Bacillus subtilis (13). The archaeal counterpart of FtsK could be the HerA motor (5). The well-studied Escherichia coli FtsK motor consists of an N-terminal transmembrane domain that anchors the protein at the septum, in a cell cycle-specific manner (14), and a C-terminal motor part (consisting of ATPase and DNA binding domains) (5), with the two parts connected by a long linker region (Fig. 1A, panel I). The motor segment has been classified further into three domains: α, β, and γ. The α and β domains are part of the ATPase domain, whereas the γ region is the DNA binding domain and is hypothesized to provide directionality for the movement of the motor (Fig. 1A, panel I) (15). Deletion of the ATPase domain in E. coli leads to defective chromosome segregation (16). Only the C-terminal motor domain of FtsK is conserved in HerA and NCLDV packaging ATPases (Fig. 1A, panel I) (5, 17). Repression of the putative FtsK-type packaging ATPase (A32L) in vaccinia virus by use of an inducible system resulted in the formation of DNA-deficient, noninfectious virus particles (18). The membrane-tethering function in HerA is carried out by a separate small transmembrane protein called MJ1617 (5). Although APMV and other NCLDV packaging ATPases lack the obvious transmembrane region, they all show the presence of a pore-lining region (Fig. 1A, panel I) (19). Studies have shown that the FtsK motor operates as a hexamer during genome segregation (20), and our preliminary experimental data also suggest that a hexamer is the functionally active form of the APMV packaging ATPase (T. Ranjan and K. Kondabagil, unpublished data).

FIG 1.

(A) Organization of functional regions on the polypeptide chains of FtsK-type ATPase (I), topoisomerase type II (II), and recombinase (III) in bacteria, archaea, and NCLDVs. The schematics represent approximate consensus organizations of functional regions of a few representative homologs from each domain. Schematics are drawn approximately to scale. (B) Phylogenetic trees for homologs of FtsK-type ATPase (I), topoisomerase II (II), and recombinase (III) in bacteria, archaea, and NCLDVs. Only sequences with bootstrap values of >50 were selected. Detailed trees and accession numbers for sequences are given in the supplemental material.

Type II topoisomerase.

During bacterial chromosome segregation, type II topoisomerase cuts both strands of the DNA, allowing the entangled DNA to pass through this transient break before resealing it in an ATP-dependent manner (21). Two-hybrid system and immunoblot studies have shown that the C-terminal domain of E. coli FtsK (the ATPase motor part) interacts specifically with the ParC subunit of topoisomerase IV (a subfamily of type II), but not with other topoisomerases, to activate chromosome decatenation (22). Furthermore, the two subunits of topoisomerase (ParC and ParE) are found in different parts of the cell during most stages of the life cycle and become active and concentrated at the cell center only after completion of genome replication, indicating their role in genome packaging (22, 23). It was suggested that concentration of topoisomerase IV at the cell center might have an additional advantage in efficient removal of any knots formed during chromosome dimer resolution by XerC/D late in the cell cycle (24). In the case of Bacillus subtilis, a spore-forming bacterium, it was observed that topoisomerase IV becomes concentrated strategically at the poles to participate in chromosome segregation (25). Interestingly, the putative topoisomerase II identified in APMV is a single long polypeptide chain comprising both subunits (Fig. 1A, panel II). The catalytic and DNA binding domains of type II topoisomerase are conserved in all bacteria, archaea, and NCLDVs (Fig. 1A, panel II). Topoisomerase II in APMV is an intermediate to late protein; its transcript is present in considerable amounts at later stages during virus assembly, implying multiple roles for this protein (see Fig. S1 in the supplemental material) (26). APMV topoisomerase II may be required for genome decatenation. Even though topoisomerase II is not encoded by vaccinia virus (Table 1), the selective recruitment of cellular topoisomerases IIα and IIβ to sites of virus assembly has been reported. More importantly, it was shown that vaccinia virus replication was inhibited by topoisomerase II inhibitors, indicating its importance in the viral assembly process (27). The Ascoviridae family members also do not seem to code for both topoisomerase II and recombinases (Table 1); they might also recruit these components from the host machinery.

Recombinase.

In bacteria, replication of the circular chromosome leads to the formation of dimeric circles that are resolved by site-specific recombinases through homologous recombination. These recombinases act in concert with FtsK ATPases at sites known as dif sites, comprising a 28-nucleotide palindromic sequence of two inverted repeats with a hexanucleotide stretch in the middle, located near the genome terminus (8). Recombinases are broadly classified into two families: tyrosine and serine recombinases. Bacteria possess both types of recombinases, and they act through two different mechanisms. The nucleophilic tyrosine residue (present in the active site) of tyrosine recombinase (XerC/D) cleaves and joins one strand of each duplex by creating a covalent DNA-protein phosphotyrosine linkage at the 3′ end of the DNA resolving Holliday junction before chromosome separation. On the other hand, serine recombinase acts on both strands of each duplex DNA by nucleophilic attack by serine residues, forming covalent DNA-protein phosphoserine linkages (28). APMV and other NCLDVs characteristically possess only serine recombinase(s). The DNA binding and catalytic domains are conserved in all bacteria, archaea, and NCLDVs (Fig. 1A, panel III). Of the three recombinases encoded by APMV (R80, L103, and R771), only the L103 gene product is a late protein (see Fig. S1 in the supplemental material) (26), and preliminary experimental data obtained using a yeast two-hybrid system suggest that the L437 packaging ATPase interacts only with the L103 serine recombinase (A. Shukla and K. Kondabagil, unpublished data).

Evolutionary significance of genome packaging components in APMV.

Phylogenetic trees were constructed for FtsK ATPases (L437), topoisomerase II (R480), and serine recombinase (L103) by using the neighbor-joining method (MEGA5). Concise trees are shown in Fig. 1B, and detailed trees are shown in Fig. S2 in the supplemental material. In the case of FtsK-type packaging ATPases, homologs for the APMV query sequence were found in all NCLDVs, including virophages. However, virophage packaging ATPases are not shown in the tree, because they are too diverged to fit into the correct tree topology with bootstrap values of >50 (see Fig. S2A). We could not find an apparent topoisomerase II and recombinase(s) in all NCLDVs. In particular, Poxviridae and Ascoviridae family members seem to lack both recombinase/resolvase and topoisomerase II (Table 1). It appears that these are predominantly encoded by viruses infecting lower eukaryotes, viz, Acanthamoeba. Interestingly, most of the homologs of APMV serine recombinase in bacteria were from organisms belonging to the phylum Cyanobacteria. However, no cyanobacterial homologs were retrieved during a BLAST search for APMV FtsK-type packaging ATPases. Iyer et al. reported that the FtsK system is missing in cyanobacteria and is replaced by the archaeal HerA-NurA system (5). A BLAST search against the complete sequence of APMV type II topoisomerase (APMV topoisomerase II is a fusion protein) (Fig. 1A, panel II) retrieved topoisomerase IV (subtype of type II topoisomerase) subunit B homologs aligning at the N-terminal region of the query. To identify homology for the C-terminal region of the query, a BLAST search was performed using the previously nonaligned C-terminal sequence, and this sequence was found to be homologous to topoisomerase IV subunit A of bacteria. Phylogenetic trees were constructed using full-length NCLDV type II topoisomerase and subunits A and B of type II topoisomerase individually from archaea and bacteria. Since archaeal subunit A is highly divergent and has a completely different domain architecture (Fig. 1A, panel II), we could not construct a meaningful phylogenetic tree. Hence, we show a tree with only subunit B of type II topoisomerase (Fig. 1B, panel II; see Fig. S2B).

Phylogenetic and sequence analyses of packaging ATPases, recombinases, and type II topoisomerases suggest that the NCLDV packaging ATPase (see Fig. S2A in the supplemental material) and recombinase are related to bacterial counterparts (see Fig. S2C), whereas topoisomerase II diverged very early (see Fig. S2B). Interestingly, unlike those of bacteria and archaea, the topoisomerase II of NCLDVs is a fusion protein, as is the case for eukaryotic type II topoisomerases. Since a BLAST search against the nr database for type II topoisomerase of NCLDVs also fetched eukaryotic type II topoisomerase, a phylogenetic tree was constructed by including eukaryotic orthologs (see Fig. S3). This tree suggests that NCLDV type II topoisomerases are probably ancestral to eukaryotic orthologs and that the fusion of subunits A and B might have occurred in ancestors of NCLDVs.

The significant evolutionary relationships of the components of the NCLDV packaging machinery with those of the prokaryotic genome segregation machinery shed light on the probable mechanism for genome packaging in NCLDVs. The early divergence observed in the components of the genome segregation and packaging machinery suggests a distinct mechanism in APMV that also needs to be considered with its DNA replication mechanism.

Genome segregation in prokaryotes.

The bacterial FtsK translocase motor is activated near the septum after late cell division, when it encounters the KOPS (FtsK-orienting polar sequence; 5′-GGGNAGGG-3′) at the end of the chromosome. Here topoisomerase II interacts with FtsK and the KOPS sequence directs this complex to the recombinase, already bound at the dif site. The complex so formed (Fig. 2A, panel I) resolves the catenated chromosome (29, 30). In archaea, clustering of the HerA gene with rad50, mre11, and nurA indicates their functional correlation during chromosome segregation at the septum (31). NurA is functionally similar to bacterial recombinases, whereas Rad50/Mre11, which bridges the double-strand breaks in DNA and helps in end processing, is orthologous to topoisomerase II (5). HerA, NurA, and Mre11/Rad50 form the complex at the cell septum and resolve the concatenated chromosome into two daughter cells (Fig. 2A, panel II).

FIG 2.

(A) Mechanism of chromosome segregation in bacteria and archaea. Chromosome decatenation by FtsK/SPOIIE (I) and HerA (II) at the septum (gray) is shown. Translocation of FtsK and HerA over the genome toward the dif site (at the termination region) helps in positioning of the catenated chromosomes. Chromosome segregation is carried out with the help of topoisomerase IV (red) in bacteria and Rad50/Mer11 (red) in archaea, with the helper protein binding to the C domain of the packaging motor (dark blue). The archaeal HerA motor is tethered to the membrane with the help of the MJ1617 protein (light blue). This complex then forms recombination synapses at dif sites by interacting with XerC/XerD (yellow) in bacteria and NurA (yellow) in archaea. DNA exchange finally completes the recombination by restoring two daughter chromosomes. The direction of DNA translocation across the septum is indicated by arrows. (B) Speculative model for chromosome segregation and packaging in AMPV. (I) Pinching of small vesicles from the host Golgi apparatus (blue) surrounding the viral factory and their fusion, leading to the formation of a membrane scaffold, over which the capsid is assembled. Simultaneously, genome replication takes place inside the viral factory. (II) Genome segregation orchestrated through the packaging ATPase (green), directed by KOPS (purple) toward topoisomerase II (yellow) and recombinase (red) at a dif site (pink star), leaving the resolved daughter genome bound to the packaging ATPase. (III) Completely resolved daughter genome with packaging ATPase bound at the end. (IV) After trimming of excess membrane, capsid assembly is completed, leaving a nonvertex opening for packaging ATPase docking and translocation of the segregated genome. (V) Finally, a single copy of the genome is docked at the opening by packaging ATPase, and the genome is translocated into the capsid.

Model for genome packaging in APMV.

Based on the presence of intergenic short inverted terminal repeats, APMV was hypothesized to assume a unique Q shape during replication (32, 33). The implications of this shape of the replicative genome for the end product of replication need to be understood. In the case of lambdoid phages, it was proposed that initial replication takes place via the theta mode, which then switches to the sigma mode, enabling it to produce long, linear, end-to-end genome polymers for packaging (34). Although APMV makes use of a prokaryotic-like machinery for chromosome segregation, its DNA replication machinery produces hundreds of copies of its genome that need to be segregated faithfully into progeny capsids. Since APMV encodes a homolog of the archaeal Mre11/Rad50 fusion protein, recombination-dependent DNA replication (RDR) was suggested (35). From all the cues, it appears that the AMPV segro-packasome machinery is probably geared more toward resolving individual genomes from concatenated dimers, as in the case of prokaryotes. We speculate that APMV might use either the rolling circle mode of replication, recombination-dependent DNA replication, or a combination of both (36) to produce several hundred copies of the genome. This is followed by circularization of the genome and switching to theta mode to produce concatenated dimers. The plasmids that undergo the theta mode of replication are known to have much more segregational stability than those that undergo the rolling circle mode of replication (37). Bacteriophage P1, an Escherichia coli phage, uses its well-known Cre-loxP recombinase system to circularize the injected genome and resolve the chromosome dimers formed as a result of DNA replication or homologous recombination (38, 39). In the case of APMV, the concatenated genome so formed, dimeric or in some other unknown form, is resolved and packaged by the segro-packasome complex, consisting of packaging ATPase, topoisomerase II, recombinase, and other, as yet unidentified components.

A recent study of APMV infection by use of cryo-electron tomography, by Mutsafi et al., revealed many finer details of the initiation stages of capsid and membrane assembly. Small membrane vesicles budding out from host Golgi cisternae are recruited to the periphery of a viral factory after about 7.5 h of APMV infection (4, 40). These small vesicles fuse among themselves to form multivesicular bodies and rupture, leading to the formation of a membrane sheet. Recruitment of additional small vesicles helps in expanding this membrane sheet. A stargate is assembled on top of this membrane sheet, and the icosahedral capsid is generated with the help of a scaffolding protein (Fig. 2B, panel I) (3). Overhanging of the inner membrane sheet might help in prevention of premature closing of the icosahedral capsid and provides a window for the packaging ATPase at the opposite face of the gate (Fig. 2B, panel I) (3).

We hypothesize that the APMV packaging ATPase oligomerizes around a strand of genome and translocates on it until it encounters the entangled region (Fig. 2B, panel I). In the case of bacteria, the directionality for FtsK motor movement is provided by the interaction of the γ domain with a short, 8-bp DNA sequence known as the KOPS (5′-GGGNAGGG-3′) (29). The γ domain also has a KRKA amino acid loop that is critically required for interaction with XerD recombinase (41). The APMV packaging ATPase possesses a KRKA motif from residues 227 to 230, toward the C terminus, which could be the potential recombinase interaction site. However, the presence of the KRKA motif is not a conserved feature in NCLDVs. We also found potential KOPS-like as well as dif-like sequences (see Table S1A and B in the supplemental material) in the APMV genome. Packaging ATPase likely is activated when it encounters a KOPS-like sequence and might recruit topoisomerase II, and this complex is directed to the recombinase already bound at the dif site (Fig. 2B, panel II). The complex so formed resolves the concatenated genome and generates the individual unit length of a genome (Fig. 2B, panel III), which could still be circular or nearly circular. The topoisomerase II and recombinase might leave the complex, as these proteins could hinder efficient translocation of the viral genome by the packaging ATPase. The capsid assembly and membrane acquisition process leads to the formation of the empty capsid on the periphery of the viral factory (Fig. 2B, panel IV). The packaging ATPase, still bound to a copy of the resolved genome, docks at the membrane-capsid protein interface of the transient opening, and ATP-dependent translocation of DNA ensues (Fig. 2B, panel V). Finally, after encapsidation of the whole viral genome, we propose that the packaging ATPase leaves the nonvertex packaging site, as it has not been identified in viral proteomic analyses (42). The packaging ATPase can be utilized for another round of packasome complex formation and genome segregation. After production of a large number of mature viral particles, Golgi cisternae start to leave the membrane assembly zone (3).

It appears that the packaging process can largely be divided into two parts: genome segregation and packaging. Both processes are most likely coupled and highly regulated by the packaging ATPase and certain features of the DNA that orchestrate the recruitment of other components in a spatiotemporal manner. Here we suggest that two strands of DNA are packaged at the same time, with the help of two packaging ATPase hexamers tethered to the membrane capsid interface (Fig. 2B, panel V). Unlike the portal of bacteriophages, the transient opening in APMV is unusually large (20 nm) and, in principle, allows the passage of two strands of DNA at the same time (3). Although the APMV packaging ATPase lacks obvious membrane-binding regions, there are two predicted pore-lining regions that probably form the membrane-spanning channel for DNA translocation. None of the available prediction programs predicted the presence of a transmembrane region in APMV packaging ATPase. An unknown factor is probably required for anchoring the packaging ATPase at the portal. After completion of packaging, disengagement from DNA might help in release of the motor from the transient opening, leading to sealing of the membrane and completion of the packaging assembly process.

Surprisingly, the vaccinia virus packaging ATPase forms a separate clade from the rest of the NCLDV packaging ATPases (see Fig. S2A in the supplemental material) and has a unique linear duplex genome with a covalently closed hairpin loop-like structure at the end. Replication in vaccinia virus, using a nicked genome terminus as a primer, leads to the formation of a dimeric tail-to-tail cruciform-like structure of concatemers at the distal end (43, 44). We propose that the packaging ATPase complex assembles at this junction and resolves the cruciform-like structure by nicking and strand exchange by recruiting host factors as discussed earlier. It was shown that the A22R protein (Table 1) is critically required for the generation of single genome copies from the concatemer (45). The packaging ATPase assembly then docks a copy of the genome to the capsid vertex for packaging and leaves the capsid after genome encapsidation. This indicates that a slightly different mechanism for genome packaging might be operational in Poxviridae.

The model presented here deals with the overall features of the possible mechanism of genome packaging in NCLDVs. Experimentation will help in testing, refining, and modifying the proposed model. The mechanism of genome segregation and packaging in APMV and other NCLDVs might represent a unique variation of an ancient prokaryotic system acquired and adapted by viral parasites of amoebae and is consistent with previous observations of amoebae as a melting pot for the advent of chimeric organisms with novel mechanisms. It might also represent a novel variation and divergent evolution of an ancient genome segregation and packaging system that was probably present in the last universal common ancestor (LUCA) or in a different life form that coexisted with the LUCA (46).