ORIGIN OF VIRUSES REVEALED BY THE GENOMIC STUDY OF

PROTEIN DOMAIN STRUCTURES

BY

ARSHAN NASIR

THESIS

Submitted in partial fulfillment of the requirements

for the degree of Master of Science in Bioinformatics
with a concentration in Crop Science
in the Graduate College of the
University of Illinois at Urbana-Champaign, 2012

Urbana, Illinois

Adviser:

Professor Gustavo Caetano-Anolles

 ii
 

ABSTRACT

The discovery of giant viruses with complex proteomes, remnants of translation

machinery and virus-specific parasites have raised important questions about their origin.

Evidence advocates for their inclusion into global phylogenomic studies and their consideration

as a distinct and ancient form of life. Here we reconstruct phylogenies describing the evolution

of proteomes and protein domain structures of viruses and cells that define viruses as a ‘fourth

supergroup’ along with cellular superkingdoms Archaea, Bacteria, and Eukarya. Universal trees

of life (uToLs) place viruses at their root and trees of domains indicate they have evolved via

massive reductive evolutionary processes. Since viral domains are widespread among cellular

proteomes we propose that viruses mediate gene transfer between cellular species and crucially

enhance biodiversity. Results call for a change in the way viruses are perceived. They likely

represent a distinct and most ancient form of life and a very crucial part of our planet’s

biosphere. Additionally, we assign functions to molecular structures in nearly a thousand

proteomes and highlight the conserved nature of cellular proteomes.

 iii
 

ACKNOWLEDGEMENTS

I am grateful to many people for their continuous support, understanding and endless

love. The first and foremost, I am grateful to my late father Muhammad Nasir Qureshi and my

mother Shaheen Nasir. My parents, brother and sister have given me their unequivocal support

throughout for which my mere expression of thanks does not suffice. It is only due to their

efforts and prayers that I am able to accomplish this much. This thesis would not have been

possible without the input of my principal adviser, Professor Gustavo Caetano-Anolles. I wish to

express my gratitude for his kindness, invaluable assistance, support and guidance throughout the

duration of my degree. I consider myself fortunate to start my research career under his

supervision. The good advice and friendship of my second supervisor, Dr. Kyung Mo Kim has

been invaluable both on academic and personal level, for which I am extremely grateful. I also

wish to thank Drs. Sandra Rodriguez-Zas and James Whitfield for serving on my graduate

committee and for valuable comments. Thanks are also extended to Liudmila Yafremava and

Kyung Mo Kim for information regarding the organism lifestyles, which proved useful in my

analysis. I specially thank the COMSATS Institute of Information Technology (CIIT) of

Pakistan for providing fellowship to support my education. Last, but by no means least, I

acknowledge my friends Zohaib, Afraz, Yasir, Masood, Waqar, Nadeem, Ahmed, Tahir, Abbas,

Zeeshan, Tayyab, and Rafia for their encouragement. In the end, I am happy to have worked with

the most amazing group of young scientists at the Department of Crop Science, UIUC and the

members of Evolutionary bioinformatics laboratory.

 iv
 

TABLE OF CONTENTS

CHAPTER 1: BACKGROUND ..................................................................................................... 1

CHAPTER 2: ORIGIN OF VIRUSES REVEALED BY THE GENOMIC STUDY OF

PROTEIN DOMAIN STRUCTURES.......................................................................................... 16

CHAPTER 3: FUNCTIONAL MAKE UP OF PROTEOMES IS CONSERVED1 ..................... 50

CHAPTER 4: STRUCTURE-FUNCTION PARTNERSHIP EMPOWERS PHYLOGENETIC

ANALYSIS AND DESCRIBES THE PROTEOMIC MAKE UP ............................................... 78

REFERENCES ............................................................................................................................. 81

APPENDIX A ............................................................................................................................... 90

APPENDIX B ............................................................................................................................... 92
 1
 

CHAPTER 1: BACKGROUND

Viruses are tiny infectious agents that depend upon a cellular host to reproduce. They

have a very simple architecture and consist of genetic material (either DNA or RNA) that is

enclosed inside a protein coat called capsid (1). Some viruses (e.g., influenza viruses)

additionally have lipid membranes that cover these capsids. These outer membranes are

recognized as viral envelopes (2). Envelopes mediate entry of viral DNA into the host cells by

means of interaction between the surface glycoproteins present on viral envelopes and receptors

on the host cell membranes (3).

Viruses are universal parasites

Viruses infect all cellular (Archaea, Bacteria and Eukarya) (4) and acellular life forms

(virophages i.e., parasites of viruses) (5). Viruses that infect Bacteria are known as

bacteriophages. They have contributed considerably to our understanding of modern biology by

serving as model systems (6). Viruses isolated from Archaea are highly diverse and

morphologically complex (7). Many have been isolated from extreme thermophilic environments

raising intriguing questions about their origins (7). Eukaryal viruses are responsible for

numerous diseases such as Hepatitis, AIDS, and flu. Additionally, there is now evidence for the

existence of virophages, i.e., viruses that infect other viruses (5, 8, 9). This ubiquitous nature of

viruses suggests they are biologically active and able to affect cellular life (8).

The abundance and diversity of the virosphere

Collectively, the group of all viruses is referred to as the ‘virosphere’ (10). The

virosphere is the most abundant and diverse group of biological entities on planet Earth (10).

Their abundance is noteworthy especially in marine environments where they are highly

concentrated (11). In seawater, there are 5-10 virus-like particles (mostly bacteriophages) per
 2
 
bacterium (12) and nearly 10 million bacteriophages per milliliter (13). More than 5,000

genotypes of viruses can be found in 100L of seawater (14) and nearly 94% of the oceanic

nucleic acid material is represented by the virus-like particles (15).

In addition, the virosphere exhibits remarkable levels of diversity in terms of physical and

genome size, morphology and other characteristics. Viruses range from being very small (~200

nm in diameter) (1) to giant viruses, with some viruses being as large as numerous bacteria (e.g.,

mimiviruses and megaviruses) (16, 17). The number of genes encoded by the viral genome

ranges from 2-3 genes (e.g., retroviruses) to nearly 1,000 for mimiviruses and megaviruses (16,

17). Morphologically, they form various structures (e.g., helical, polyhedral, complex etc.).

Genetic material can either be DNA (e.g., adenoviruses) or RNA (e.g., human immunodeficiency

virus) that in turn can either be single stranded or double stranded (1). As such, the remarkable

abundance and diversity of the virosphere has led to its acceptance as an integral component of

the biosphere (10, 15).

Theories about the origin of viruses

The question about the origin of viruses is fundamental to our understanding of life.

Unfortunately, studying viral origins is challenging because first, unlike cellular organisms, they

do not have any fossil record. We are therefore left to deduce their evolutionary trajectories by

studying modern day viruses. Second, the virosphere is highly diverse (as explained above).

Such high levels of diversity, and the knowledge that no single gene unites all the viruses,

suggest they had multiple evolutionary origins (i.e., their origin was polyphyletic) (18). In

contrast, living cellular organisms are united by the presence of the ribosome and a set of genes

that are shared by all the organisms (19). There are however groups of viruses that share

conserved sets of genes and appear to have a common evolutionary origin (i.e., monophyletic)
 3
 
(read below). However, the virosphere as a whole has been suggested to be not monophyletic

(20). Given these realizations, three theories have been proposed to explain the origins of

viruses.

1. The virus-first hypothesis

Proponents of this theory argue that viruses are more ancient than the cellular organisms

and predate the ‘Last Universal Common Ancestor’ (LUCA). This theory implies that ancient

viruses had the ability to self-sustain and replicate on their own (18, 21, 22). This theory has

received a fair amount of criticism because all modern day viruses need a cellular host to

replicate and follow a parasitic lifecycle (23).

2. The reduction hypothesis

This theory states that viruses are degenerates (reduced forms) of modern day parasitic

organisms and evolved by reductive evolutionary processes. In doing so, they lost all the genes

required for self-sustenance (e.g., metabolic genes) and maintained only those required to

parasitize the host cells (24).

The reduction hypothesis has been contested on grounds of lack of intermediates between

viruses and parasitic organisms (23). Recently, giant viruses (e.g., mimiviruses and megaviruses)

have been shown to cross the barriers between viruses and small parasitic organisms (16),

providing support to the reduction hypothesis.

3. The escape hypothesis

The escape hypothesis considers viruses to have evolved from genetic material of the

host cell that escaped from its control and developed autonomous evolving abilities (25). The

support of this theory comes from the observation that modern day viruses can act as mediators

of gene transfer and pick-pocket genes from their cellular hosts, a mechanism known as
 4
 
horizontal gene transfer (HGT). Some scientists believe that HGT has been a dominant force in

shaping the viral proteomes (26, 27). This theory, however, fails to explain the presence of

structures and genetic material in viruses that have no homologs in the cellular world (10, 28).

Which theory is true?

The theories presented above have been highly debated and questioned (23). Some

scientists believe that a composite or new explanation may be more accurate and sound. Raoult

and Forterre (2008) proposed the division of life into two forms where cellular organisms are

united by the presence of ribosomes (ribosome encoding organisms, REOs) and viruses united by

the widespread presence of capsid (capsid encoding organisms, CEOs) (19).

Status of viruses in the living world

Viruses are generally regarded as neither living nor non-living (29). They are entirely

dependent upon the host cells for reproduction and for this reason are considered to be obligate

intracellular parasites (29). Viruses however follow both the biologically active and inactive

lifecycles. Once inside the cells, they either follow the ‘lytic life cycle’ during which they lyse

the cell and release viral progeny or the ‘lysogenic life cycle’ during which they integrate their

genome into the host DNA and remain dormant (30). One of the striking findings to come out of

the sequencing of human genome project was the realization that endogenous retroviruses

account for nearly 8% of the human genome (31). These endogenous retroviruses resemble

modern day retroviruses and are considered to be the remnants of ancient viral infections (31).

Outside the host cells, viruses are metabolically inactive and survive long periods of time as

crystals (chemical objects) (32). Based on these realizations and dual life cycles (both biological

and chemical), viruses are regarded as intermediates between living and non-living forms (29)

and considered to be at the border of living and non-living organisms.

 5
 
Viruses with large genomes

One important realization is that not all viruses have minute genomes. In particular, a

large group of dsDNA viruses that infect Eukarya have large genomes with hundreds of genes

(33, 34). This group is known as Nucleo-Cytoplasmic Large DNA Viruses (NCLDV) and is

apparently a monophyletic group that includes five major families of viruses including

Poxviridae, Asfarviridae, Phycodnaviridae, Ascoviridae/Iridoviridae and Mimiviridae (33, 34).

Unlike other viruses that are strictly dependent on the host cells for replication, NCLDV

are not exclusively dependent (35). They encode several key proteins involved in the processes

of DNA replication, translation, and DNA repair (20). Additionally, all the families of NCLDV

share a conserved set of 9 genes that are present in all the families and 22 genes present in at

least 3 out of 4 families (34). Presence of this conserved core suggests that NCLDV originated

from a common ancestor that had a complex gene repertoire (20).

Mimiviridae

The most exciting addition to the NCLDV was the introduction of Mimiviridae. The first

member of Mimiviridae, Acanthameoba polyphaga mimivirus (APMV) was identified in 2003

by researchers in France (36). APMV (simply mimivirus) is one of the largest and most complex

viruses known to date and for that reason, it was initially mistakenly identified as gram-positive

bacterium until its viral nature was described in 2003 (16, 36). Mimivirus infects amoeba and

indirect evidence (presence of antibodies) suggests it may be responsible for human pneumonia

(37). Mimivirus is a ‘Gulliver among Lilliputians’ (38) as it’s sheer physical (750 nm diameter)

and genome size (>1Mb) outstrips almost all the known viruses and numerous parasitic bacteria

(16). Mimivirus has many other additional ‘unviral’ features as well, such as the presence of

genes related to DNA repair and protein translation. Mimivirus enters the host cell by a unique
 6
 
entry mechanism called ‘Stargate’ (39) and encodes promoter-like elements that share structural

resemblance with the eukaryotic promoters (40). Mimivirus relatives are highly abundant in

marine environments (41) suggesting it is not the only known giant virus. Recently, the

discovery of Megavirus chilensis (megavirus) and description of its genome identified megavirus

as the largest virus known to date (17). Both the particle and genome size of megavirus exceed

that of mimivirus and is considered to be a phylogenetic relative of mimivirus (17).

Implications of these findings

Our knowledge about the viral world has increased dramatically since the discovery of

viruses harboring large genomes (16, 17, 36). These viruses defy the definition of virus,

overlapping numerous bacteria both in genomic and physical features. Their existence has raised

interesting questions about the status of viruses in the living world and called for a redefinition

(19).

Some scientists believe that the ability to release viral progeny and the overlap between

giant viruses and cellular organisms supports their identification as a distinct form of life (42).

Others believe that because most viruses have a simpler organization and lack ribosomes

(essential components of translation machinery) they cannot and should not be compared to the

cellular organisms (27, 29). The question remains one of the most debated ones in the field of

evolutionary biology (9, 28, 29, 42, 43).

The advancements in molecular biology and comparative genomics have enabled us to

study the evolutionary relationships between viruses and cellular organisms on firmer grounds.

Sequencing of viral genomes with genome sizes comparable to bacteria gives us the opportunity

to reconstruct their phylogenetic history by comparing them to other viruses and/or cellular

organisms.
 7
 
The concept of a Tree of Life (ToL)

The unusual genomic and physical features of giant viruses argue for the inclusion of

viruses, at least the ones with large genomes, into the universal tree of life (ToL) together with

the cellular proteomes of Archaea, Bacteria and Eukarya. The ToL is a model that explains

diversification of living organisms into three distinct superkingdoms i.e., Archaea, Bacteria and

Eukarya. It was first described by Carl Woese who identified Archaea as a distinct domain of life

and proposed the three-domain classification system (44, 45). Prior to Woese’s work, the

organisms of the ToL were divided into prokaryotic and eukaryotic groups.

Viruses in the phylogenomic picture

Phylogenetics is the study of evolutionary relationships between organisms. Phylogenetic

trees are built from data that are either morphological or molecular features (characters),

expressing different values or properties (character states), and describe the history of a

biological system (taxa) (46). These trees are network representations with branches, nodes

connecting branches and leaves (taxa). Prior to the advancements in molecular biology,

phylogenies were restricted to the use of morphological data and fossil records. These

phylogenies were limited due to the lack of comparable characters between the many lineages of

microorganisms and higher organisms (46).

The advances in molecular biology enabled the use of molecular characters for

phylogenetic reconstruction, such as gene sequence, gene order, gene content and structural

information (47). This led to many studies that grouped organisms into kingdoms and

superkingdoms and produced phylogenies that corroborated morphological data. Viruses

however were not included in these analyses (44, 45, 48, 49).
 8
 
Discovery of giant viruses encouraged scientists to place viruses on the global

evolutionary scale. Initial phylogenomic analysis based on the concatenation of seven universally

conserved proteins (Arg-tRNA synthetase, Met-tRNA synthetase, Tyr-tRNA synthetase, RNA

polymerase II largest subunit, RNA polymerase II second largest subunit, Proliferating cell

nuclear antigen, and 5’-3’ exonuclease) in the proteomes of mimivirus, Archaea, Bacteria and

Eukarya placed mimivirus at the base of Eukarya and identified mimivirus as a distinct

superkingdom in the ToL (16). This study was challenged and future analysis involving

individual proteins, rather than the concatenated set nested mimivirus consistently within

Eukarya and close to amoeba (the cellular host of mimivirus) (50). Authors proposed that

previous analysis was flawed due to the use of concatenated protein sequences that could

produce misleading trees if proteins evolved by different evolutionary mechanisms (50). Nesting

of mimivirus within Eukarya suggested that the gene repertoire of mimivirus was acquired from

amoeba by HGT (50). These experiments highlight the differences in the methods used to build

phylogenetic trees (e.g., concatenated set VS individual sequences) and suggest that HGT has

tailored the proteome of mimiviruses. However, both these analyses used only a limited number

of species and a very small set of proteins. Increased sampling of species from Archaea,

Bacteria, Eukarya and a large number of giant viruses coupled with a phylogenetic model that is

robust against the artifacts described above appears to be the most likely solution to study the

origins of viruses and life itself.

Identifying the right approach for ToL reconstruction is important

In general, there are two main approaches, sequence-based approaches that utilize

nucleic acid or protein sequence information and structure-based approaches that focus on

molecular structural information.

 9
 
1. Sequence-based approaches

Sequence-based approaches deal with gene sequence(s) and use an alignment file to

reconstruct phylogenetic trees. An alignment file is a representation of arranged DNA, RNA or

protein sequences such that homologus regions are aligned and identified easily. The aligned

regions are of particular interest and reflect structural and evolutionary importance whereas gaps

in alignment represent past insertion or deletion events (i.e., indels) (51). Multiple genes can be

aligned simultaneously (52) and large sampling of taxa is advised to increase the power of

phylogenetic reconstruction and decrease the incidence of phylogenetic artifacts (46).

Sequence-based approaches are not restricted to the use of alignment to infer

phylogenies. Other features such as gene content and gene order (whole-genome features) have

been utilized in past that do not require alignment computation (53, 54). These features compare

genomes simultaneously and are considered superior over the use of sequence alignment because

changes in gene content and gene order in genomes result in billions of possible states as

opposed to only four states for nucleotide alignment and twenty for protein alignment (46). As

such, they are relatively robust against the affects of HGT and problems resulting from the use of

sequence alignment (read below) (46). Both approaches are computationally intensive and

depend upon the phylogenetic model, data and assumptions that can be influential when

examining deep and ancient relationships.

2. Structure-based approaches

Structure-based approaches embed structure information into phylogenetic analysis and

reap the benefit of processes occurring at higher and more conserved levels of the structural

hierarchy (read below). These processes are responsible for the redundant appearance and

accumulation of modules in the structure of living organisms [e.g., how many times particular
 10
 
protein domains are present in proteomes (i.e., genomic abundance)]. Domain structure is

maintained in proteins and remains conserved as apposed to the amino acid sequence that is

dynamic and variable (48, 55-58). For this reason, protein domains are also considered

evolutionary units (48, 59, 60) and make useful ‘phylogenetic characters’ when recovering deep

evolutionary relationships (10, 48, 57, 61).

Domains are grouped into different hierarchical classifications based on their

evolutionary relatedness (59, 62, 63) such as in the well-established ‘Structural Classification of

Proteins’ (SCOP) (59, 63). In SCOP, domains having high sequence conservation (>30%) are

grouped into fold families (FFs), FFs with evidence of common ancestry, based on structural and

functional relatedness, into FSFs, FSFs with common secondary structure in similar

arrangements into folds (Fs), and Fs sharing a common secondary structure into classes (58, 59,

63). Domains are identified using concise classification strings (css) (e.g., d.211.1.1, where d

represents the protein class, 211 the F, 1 the FSF and 1 the FF). The 110,800 domains indexed in

SCOP 1.75 (corresponding to 38,221 PDB entries) are classified into 1,195 F, 1,962 FSFs, and

3,902 FFs. Compared to the number of protein entries in UniProt (534,695 total entries as of

February 22, 2012) the number of domain structural designs is quite smaller and suggests that the

fold space is finite and highly conserved (57, 61).

Because domains defined at higher levels of SCOP classification (i.e., FSFs) are more

conserved than domains defined at lower levels (i.e., FFs and sequences) and less likely to be

affected by HGT (61, 64-67), the genomic abundance of FSFs in proteomes has been

successfully utilized as the phylogenomic characters to reconstruct the trees of proteomes (ToPs)

and trees of protein domains (ToDs) (48, 49). This focus on structure as general evolutionary

principle of biology offers several advantages over standard phylogenetic methods and
 11
 
overcomes important limitations imposed by the violation of assumptions that occur when

attempting to extract deep phylogenetic signal present in molecular sequence data. The top ten of

these advantages are listed below and in Table A.1.

1. ToDs and ToPs are derived from non-parametric models of genomic abundance that

are free from problems of homology in the alignments of sequences (68, 69). Once structural and

functional considerations assign a protein sequence F, FSF or FF level of structural abstraction of

SCOP (59, 63), homology is established. In contrast, sequence alignment remains problematic

because there is still not an objective function in bioinformatics that can describe homology in

sequence (especially, remote homology) (68).

2. Domain structure is more conserved than amino acid sequence and valuable when

studying deep evolutionary relationships (57).

3. ToDs and ToPs are not affected by the serious problem raised by Maddison (70) of

characters that are not applicable to all taxa in a data set [e.g., insertion/deletion (indel) sites in

sequence alignment]. This problem plagues phylogenetic analysis of sequences (71).

4. Mutation saturation destroys phylogenetic signal in sequences, a serious problem

affecting the validity of deep phylogenetic inference (72). This problem does not apply to

domain abundance, which is evolutionarily highly conserved, increases with time, and in doing

so enhances deep phylogenetic signal.

5. ToDs and ToPs are appropriately based on a historical analysis of molecular units of

evolution, function and structure, the proteins domains (59, 73). In contrast, trees of genes

(ToGs) generally consider genes are evolutionary units. However, a substantial number of genes

code for proteins that have multiple domains [55% in Archaea, 72% in Bacteria, and 84% in

Eukarya], each of which contribute confounding histories to phylogenetic reconstruction (74).

 12
 
6. Macromolecules are assemblies of substructural components with different

evolutionary histories (75). Sequence analysis fails to take into consideration the historical

relationships and evolutionary heterogeneities that exist between subsets of sequence sites. In

contrast, the study of molecular domains is impervious to the history of domain make up. It

involves the entire molecular module.

7. The troublesome task of solving the problem of orthology (genes separated by a

speciation event) and paralogy (duplicated genes) in sequence analysis (76) is inapplicable to

domains at any level of structural abstraction, which by definition include all domain sequence

variants (59). According to SCOP definitions, protein domains that belong for example to an FF

can be orthologous to each other. Consequently, each FF is an orthologous evolutionary unit.

8. Evolutionary processes such as convergent evolution and horizontal gene transfer can

confound phylogenetic analysis leading to erratic interpretations when dealing with molecular

sequence data. However, the effect of these processes on ToLs built from domain structures

defined at conserved and higher levels of SCOP hierarchy appears to be very limited (49, 64-67,

73). Even the lower Pfam hierarchical level of structural organization showed limited influence

of HGT (<10%) (65). Phylogenetic and statistical analyses have revealed that indeed the

convergent evolution of domain structures is rare (49, 64-67, 73) and the diversity of protein

world is largely due to enhancements of functions utilizing domain rearrangements processes

(74). Additionally, the hypothesis suggestion multiple origins for proteins is not well supported

(77).

9. Taxon sampling represents a problem in phylogenetic analysis that impacts accuracy

and phylogenetic error (78). ToDs are refractory to the problem since they sample the set of all

known domains (i.e., they portray history of an operationally finite set of taxa).
 13
 
10. Finally, the most fundamental principle of phylogenetic analysis, character

independence, states that each character must serve as an independent hypothesis of evolution

(79). Violation of character independence is serious and results in phylogenies that do not reflect

true evolutionary history (80). Molecular structure is defined by interactions between nucleotide

sites in a protein sequence. This mere fact violates character independence of sequence analysis,

especially when ToGs include sequences with structures that are divergent. In contrast, ToDs and

ToPs are free from violation of character independence as long as individual domains

(orthologous according to SCOP definition) or proteomes, respectively, do not co-evolve with

each other [e.g., in cases of symbiosis or parasitism] (48).

The phylogenomic approach utilizing structural information as phylogenetic characters

has been used previously to reconstruct the ToLs proposing an early divergence of Archaea

relative to Bacteria and Eukarya (67, 81, 82), reveal the reductive trends in the proteomes of

cellular organisms (48, 83), used to reconstruct the structural make up of the last universal

common ancestor (LUCA) of life (49), to link molecules and planetary events (inferred from

geology and paleobiology) with a clock of folds (84) among others.

For these reasons, global evolutionary statements considering viruses should be tested

with a phylogenetic model that is robust against the artifacts that result from the use of sequence-

based approaches. The genomic abundance counts of protein domains (defined into FSFs) are

useful phylogenetic character states that have been successfully utilized in the past to describe

the evolution of protein domains and organisms (see chapter 2 for description of methodology).

Assigning FSF structures to proteomes

SUPERFAMILY hosts a library of hidden Markov models (HMMs) that are profiles

based on multiple sequence alignments, representing a protein superfamily (or family). These
 14
 
models represent all the proteins with known structure and are generated by the iterative

Sequence Alignment and Modeling (SAM) algorithm (85). SAM takes as input a protein

sequence with known 3D structure and searches for its homologs in the sequence databases.

Once the homologs have been retrieved, a multiple sequence alignment file (MSA) is generated

that is used by SAM to generate a SUPERFAMILY model which best describes the features of

this alignment. SAM is considered to be one of the best algorithms for the detection of remote

protein homologies (86, 87). Once the models have been generated, any protein sequence with an

unknown structure can be searched against the entire library of these models and the model with

the best hit is returned. This model represents a FSF and that FSF is assigned to the query protein

sequence with high accuracy (error rate <1%) (85). SUPERFAMILY currently provides FSF

structural assignments for a total of 1,258 model organisms including 96 Archaea, 862 Bacteria

and 300 Eukarya.

Assigning functional categories to protein domains

While structural information can be helpful in building phylogenies, functional

annotation contributes to the description of the functional make up of proteomes (58). Assigning

molecular functions to FSFs is a difficult task since majority of the FSFs (~80%) perform

multiple functions and are quite diverse (49). For example, most of the ancient FSFs, such as the

P-loop-containing NTP hydrolase FSF (c.37.1), are highly abundant in nature and include many

FFs (20 in case of c.37.1) (58). Each of those families may participate in multiple pathways and

perform related but different functions. The SUPERFAMILY functional annotation scheme

introduced by Vogel and Chothia is a one-to-one mapping between FSFs and molecular

functions and is based on information from various resources, including the Cluster of

Orthologus Groups (COG) and Gene Ontology (GO) databases and manual surveys (88-91).
 15
 
When a FSF is involved in multiple functions, the most predominant function is assigned to that

multi-functional FSF under the assumption that it is the most ancient and predominantly present

in all proteomes. The error rate in assignments is estimated to be <10% for large FSFs and <20%

for all FSFs (58).

The SUPERFAMILY functional classification maps 7 major functional categories to 50

minor functional categories. The 7 major categories include Metabolism, Information,

Intracellular processes (ICP), Extracellular processes (ECP), Regulation, General, and Other.

The major categories are subdivided into 50 minor categories and the mapping between FSFs

(defined in SCOP 1.73) and molecular functions corresponding to the major and minor is

described in chapter 3 (58).

In this study, we take advantage of both the SUPERFAMILY structural and functional

annotation schemes for domains defined at FSF level and make evolutionary statements about

the origin of viruses. For this exercise, we consider viruses with medium-to-very-large genomes

(mainly NCLDV) since their lineages apportion considerable diversity to the virosphere. We also

explore the functional make up of cellular proteomes and discover that it is remarkably

conserved with the exception of parasitic organisms. Results yield significant insights into the

evolution and origins of viruses and describe the mapping between structures and functions for

nearly a thousand organisms.

 16
 

CHAPTER 2: ORIGIN OF VIRUSES REVEALED BY THE

GENOMIC STUDY OF PROTEIN DOMAIN STRUCTURES

Introduction

Ever since the discovery of viruses, scientists have wondered about their origins and roles

in cellular evolution. There is considerable abundance and diversity in the virosphere (the group

of all viruses) for it to be regarded as an integral component of the biosphere (15). The last few

decades have seen an increase in our knowledge about the viruses, particularly boosted by the

discovery of viruses with large genomes (17, 36). These viruses question the way we have

perceived viruses in the past and call for a redefinition (9, 19, 92) In this study, we take

advantage of the advancements in the fields of genomics and bioinformatics to study the

evolution of viruses on a scale comparable to the cellular organisms. We sample viruses with

medium-to-very-large proteomes (giant viruses) since their lineages apportion considerable

virospheric diversity (42). This study offers valuable insights into the origin and evolution of

viruses and highlights their crucial contribution to the biosphere.

 17
 
Materials and Methods

Cellular dataset

The SUPERFAMILY database provides genomic assignments of SCOP domains defined

at FSF level for publically available sequenced genomes (85, 93). We downloaded the FSF

assignments for a total of 981 organisms (70 Archaea, 652 Bacteria, and 259 Eukarya) from the

SUPERFAMILY web server (date: August 29th, 2010). These 981 proteomes constitute the

cellular dataset.

Viral dataset

We downloaded the protein sequences encoded by 56 viral genomes including 51

NCLDV and 5 viruses from Archaea, Bacteria and Eukarya from NCBI viral genome resource

homepage (link: http://www.ncbi.nlm.nih.gov/genomes/GenomesHome.cgi?taxid=10239). These

56 proteomes constitute the viral dataset.

Assignment of FSFs to viral dataset

We assigned structural domains corresponding to 1,830 FSFs (out of 2,223 defined by

SCOP 1.75) to the viral dataset using HMMs of structural recognition in SUPERFAMILY at

probability cutoff E value of 10-4. The cellular and viral datasets define a total dataset of 1,037

proteomes (56 Viruses, 70 Archaea, 652 Bacteria, and 259 Eukarya) with a total FSF repertoire

of 1,739 FSFs (91 FSFs had no representation in our datasets and were excluded from the

analysis).

Phylogenomic analysis

We generated phylogenomic trees that describe the evolution of protein domains (ToDs)

and proteomes (ToPs) using the genomic abundance of FSFs as phylogenomic characters. We

began by counting the number of times each FSF was represented in each proteome for the total
 18
 
dataset. We define this count as the genomic abundance value (g) and presented these values in a

1037 * 1733 ([total number of proteomes] x [total number of FSFs]) matrix. Because large

genomes are expected to have higher counts of FSFs compared to smaller genomes, g values can

range from 0 (absent) to thousands (82, 87). In order to account for the unequal genome sizes

and large variances, we normalized the g values to a 0-23 scale in an alphanumeric format (0-9

and A-N) using the following formula (82, 87).

gab_norm = Round [ln(gab + 1)/(gmax + 1) * 23 ]

In this formula, a and b represent a FSF and a proteome respectively; gab describes the g

value of the FSF a in the proteome b; gmax is the maximum g value in the matrix and the Round

function normalizes the gab value taking into account the gmax and standardizes the values to a 0-

23 scale (gab_norm) (82, 87). The normalization results in 24 transformed values that are

compatible with PAUP* ver. 4.0b10 (phylogenomic reconstruction software) (94) and represent

linearly ordered multistate phylogenetic character states (82, 87).

For the trees describing evolution of FSF domains (i.e., ToDs), we declared N (maximum

value) as the ancestral character state under the assumption that the most abundant FSF

(character N) appeared first in evolution. For this reconstruction, FSFs were treated as taxa and

proteomes as characters. For the trees describing evolution of proteomes (i.e., ToPs), we declared

0 (minimal value) as the ancestral character state under the assumption that the ancestral

proteome had a fairly simpler architecture and there was a progressive trend towards organismal

complexity. For this reconstruction, data matrix was transposed to represent proteomes as taxa

and FSFs as characters. Ancestral states were declared using the ANCSTATE command and

trees were rooted using the Lundberg method that does not require to specify the outgroup taxa.

Maximum Parsimony (MP) was used to search for the best possible tree. To evaluate the
 19
 
reliability of phylogenetic trees, we carried out a bootstrap analysis with 1,000 replicates. From

the ToDs, we calculated the relative age of each FSF defined as the node distance (nd) using a

PERL script that counts the number of nodes from a hypothetical ancestral FSF at the base of the

tree to each leaf and provides it in a relative 0-1 scale (48). In order to evaluate homoplasy (i.e.,

conflict between data and tree) affecting the ToPs, retention indexes (ri) were calculated for

individual FSF characters using the ‘DIAG’ option in PAUP* (87, 95).

Distribution of FSF domains in organismal groups

In order to evaluate the spread/popularity of FSFs across the three cellular

superkingdoms, Archaea (A), Bacteria (B) and Eukarya (E), and viruses (V) (which we

collectively refer to as supergroups) we assigned FSFs appearing in all superkingdoms to the

ABEV category, those present in all but one superkingdoms to the ABE, AEV, BEV, and ABV

categories, those present in two superkingdoms to the AB, AE, AV, BE, BV, and EV categories,

and those present in only one to the A, B, E, and V categories.

We used a previously reported distribution index (f) to describe the popularity of FSFs

across all the proteomes. This index ranges from 0 to 1 and represents the fraction of proteomes

harboring a particular architecture (48). An f value of 1 indicates that a particular FSF is present

in all the proteomes and a value close to zero indicates that it present in only very few

proteomes.

Estimating the bias in the spread of viral FSFs among cellular proteomes

The total repertoire of FSFs present in each of the cellular superkingdom was split into

two components: (i) the set of FSFs shared with viruses, and (ii) the set of cellular FSFs of that

superkingdom (shared or not shared with other superkingdoms). The f index for both components
 20
 
was represented in a boxplot generated using R programming language (link: http://www.r-

project.org/).

Enrichment of taxonomic groups with viral FSFs

In order to estimate which taxonomic groups are enriched with viral FSFs, we compared

the counts of FSFs in cellular taxonomic groups (background) against the viral groups (sample).

The probability of enrichment of a particular taxonomic group was calculated using the

hypergeometric distribution and the equation (49, 66):

In this equation, M indicates the number of FSFs in the background; k indicates the

number of FSFs in the sample (i.e., viral taxonomic group); N is the total number of FSFs of the

two sets; n represents the total number of FSFs in sample and P(X = k) is the probability that

implies a chance that a random variable X has k FSFs for a given taxonomic group (49, 66).

Referring to the equation above and previous literature (49, 66), we calculated P values for the

individual viral taxonomic groups having k/n larger (i.e., overrepresented) or smaller (i.e.,

underrepresented) than M/N, and evaluated statistical significance with 95% confidence level (P

< 0.05).

Functional annotation of viral FSFs

We used the functional annotation scheme described by Vogel and Chothia (88-90) to

assign molecular functions to viral dataset. Mapping between the categories is described in

chapter 3 (58).
 21
 
Results

A significant number of FSFs exist in the viral dataset

A large number of domains (4,679) corresponding to a rich repertoire of FSFs (304

distinct FSFs) were detected by the SUPERFAMILY HMMs in the sampled viral proteomes

(Table 2.1). Six out of these FSFs were unique to viruses (had no cellular representation) and

were part of a small subset of viral specific FSFs that are responsible for functions unique to

viruses, such as attachment to the host cell receptors and DNA (b.126.1, b.21.1, a.54.1, g.51.1),

inhibiting caspases to trigger anti-apoptosis (b.28.1), and acting as major capsid proteins

(b.121.2). On average, ~34% (total FSFs/total proteins) of proteins in viral proteomes received

SCOP structural assignments (Table 2.1). This was higher than expected, given the simplistic

and reduced nature of viral proteomes. As expected, mimivirus had the most number of

assignments (530 out of its 911 proteins; 58%).

Mimivirus is comparable to parasitic riboorganisms

Despite considerable proteomic coverage (~58%), mimiviral assignments were restricted

to only 163 distinct FSFs, a rather poor repertoire when compared for example with FSFs present

in the proteomes of riboorganisms exhibiting free-living (FL) lifestyle (ranging from 407 FSFs

in Staphylothermus marinus to 1,084 in Capitella sp.). FSF number was however comparable to

small riboorganisms with reduced genomes or parasitic/symbiotic lifestyles [e.g., Guillardia

theta (189 distinct FSFs), NanoArchaeam equitans (211 distinct FSFs) and Candidatus

Hodgkinia (115 distinct FSFs)]. The average reuse level of mimiviral FSFs was quite low as well

(total FSFs/distinct FSFs = 3.25) but still comparable to that of organisms with similar genome

size or lifestyles (e.g., 3.03 in Staphylothermus marinus, 1.42 in Candidatus Hodgkinia, 2.01 in

NanoArchaeam equitans, and 2.48 in Guillardia theta). Thus while mimiviruses have a genome
 22
 
size comparable to numerous small bacteria they also seem to share with them a very simple

proteome. Both the genomic and proteomic features of mimiviruses overlap significantly with

the obligate parasitic unicellular organisms and exhibit a level of complexity that was never

associated with viral proteomes before.

The sharing of FSF domains between cellular and viral taxonomic groups

FSFs are not equally shared by proteomes of the three cellular superkingdoms, Archaea

(A), Bacteria (B) and Eukarya (E), and viruses (V), which we here refer collectively as

supergroups. In turn, FSFs exist that are uniquely present (groups A, B, E or V) or are shared by

two (AB, AE, AV, BE, BV, EV), three (ABE, ABV, AEV, BEV) or all superkingdoms (ABEV).

We will refer to these distinct groups of FSFs as ‘taxonomic groups’.

A Venn diagram (Figure 2.1) describes FSF distributions in supergroups and highlights

the differential enrichment of viral FSFs within the cellular taxonomic groups. All cellular

taxonomic groups share FSFs with the viral superkingdom. ABE is the most populated group

with 557 FSFs, BE is the second largest with 291 FSFs and ABEV makes the third largest group

with 229 FSFs. Eukaryotes have the highest number of supergroup-specific FSFs with 335

(~19% of the total FSFs present only in eukaryotes). Bacteria have the second highest number

with 163 (9.37%) bacteria-only FSFs, followed by Archaea with 22 (1.26%) and viruses with 6

(0.345%) supergroup-specific FSFs, respectively. The relatively low number of viral specific

FSFs in our dataset can be explained by the fact that the current sequencing trend is biased

towards the sequencing of viruses with medical or economic importance (10). Sequencing of

more viral genomes, especially the large genomes of giant viruses (like mimiviruses and

megaviruses) will lead to the identification of more viral-specific FSFs. However, we expect that

patterns of FSF sharing between viral and cellular taxonomic groups will remain the same.
 23
 
Primordial reductive evolutionary processes explain viral make up

We generated ToDs using information in the total dataset of 1,037 proteomes that

included free-living and parasitic riboorganisms and viruses and embodied 1,739 FSFs (Figure

2.2). These trees describe the appearance of protein domains in proteomes on an evolutionary

scale and represent structural chronologies that unfold directly from the trees by counting the

number of nodes along a lineage from the root to an individual leaf (i.e., an FSF). This node

distance (nd) defines the age of each FSF on a relative 0-1 scale, with nd = 0 (48) representing

the origin of proteins and nd = 1 the present, and is linearly proportional to geological time and

thus can be used to date domains defined at FSF level accurately (48, 84, 87). Previous studies

show that these chronologies uncovered unprecedented details of protein and proteome

diversification at fold, FSF and FF levels of structural complexity (48, 74). Here, trees of FSFs

highlight the distribution of FSFs present in viruses at an evolutionary time scale when nd values

are plotted against the relative number of riboorganisms and viruses using each FSF (distribution

index; f) (Figure 2.3). We find that most viral FSFs originated either very early or very late,

showing a clear bimodal pattern of proteome diversification. For example, both the oldest

(c.37.1) and youngest (d.211.1) FSF are present in viruses (Figure 2.2)

The distribution of FSFs in the total dataset revealed that the most ancient FSF, the P-

loop hydrolase fold FSF (c.37.1), was omnipresent in all proteomes (f = 1), including the viral

proteomes (Figure 2.3:Total). In total, 28 ancient FSFs had f >0.947 and were present in almost

all cellular and viral proteomes. However, the representation of FSFs decreased in the timeline

with increasing nd until f approaches 0 at about nd = 0.58 (Figure 2.3:Total). From this point

onwards, an opposite but interesting trend takes place and the FSFs increase their representation

in proteomes increasing nd. Distribution plots for the individual superkingdoms (Figure 2.3)
 24
 
confirmed that, in general, the most ancient FSFs (nd = 0-0.4) are shared by most proteomes.

However, the representation of ancient FSFs decreases in time, first in the viral proteomes

(Figure 2.3:Viruses), and then in the cellular proteomes, starting with Archaea (Figure

2.3:Archaea), then Bacteria (Figure 2.3:Bacteria) and finally Eukarya (Figure 2.3:Eukarya).

However, FSF distributions approached a minimum first in viruses, and then in Archaea,

Eukarya and Bacteria, in that order, while in general, viruses, Archaea, and Bacteria maintained

small representation of younger FSFs in proteomes while FSF representation increased in the

eukaryal proteomes as f reached 1 again at nd = 1 (Figure 2.3:Eukarya).

Mimivirus represents a living viral fossil

In general, loss of ancient architectures was abrupt and massive for viruses. It started very

early in evolution but substantially dropped in the nd = 0.4-0.6 range (Figure 2.3:Viruses). The

class II aminoacyl-tRNA synthetases (aaRS) and biotin synthetases FSF (d.104.1) was the first

FSF to be completely lost in our viral proteomic set nd = 0.0516 (Figure 2.3:Viruses and boxplot

for ABE taxonomic group in Figure 2.4). This FSF includes the class II aaRS enzymes that

alongwith class I aaRS enzymes are responsible for charging tRNAs with correct amino acids

and make the central components of the translational machinery (96, 97). Megavirus has been

shown to encode for class II aaRS for AsnRS in addition to 6 class I enzymes (TyrRS, MetRS,

ArgRS, CysRS, TrpRS, and IleRS) and this makes it the only virus known to possess both class I

and class II asRS enzymes (17). While d.104.1 is absent in the viral dataset, we find that

mimiviruses maintain a small representation of class I aaRS (TyrRS, MetRS, ArgRS, CysRS;

c.26.1, f = 0.035). This indicates that loss of class II enzymes from the mimiviral lineage

occurred very early in evolution. These enzymes make the central components of the

translational apparatus (98). The mimivirus genome encodes 4 functionally active class I aaRSs
 25
 
and phylogenetic and statistical evidence (read below) suggests they were not transferred from

its cellular host (amoeba) via HGT (17). In addition, c.26.1 has a higher character retention index

(ri) value (0.84) relative to the mean ri values of FSFs in Archaea (0.72), Bacteria (0.75), and

Eukarya (0.76) (Figure 2.5). The ri measures levels of homoplasy (conflict in how data matches

the reconstructed tree) that portray processes other than the vertical inheritance of characters

(FSFs) on a relative scale of 0-1 and is sensitive against the large number of taxa (95). Thus, the

existence of a partial translational apparatus in the mimiviral genome does not owe its presence

to HGT but rather to reductive evolution of the proteome of an ancestral more complicated virus,

which included functional translational machinery (17). In conclusion, the substantial loss of

domain structures that materializes very early in our timeline supports the concept that viruses

are reductive evolutionary variants of a primordial form that coevolved with the ancestor of

diversified cellular life (Figure 2.3). Mimivirus is the least reduced form of that ancestral virus

and thus represents a living virus fossil.

The distributions of FSFs in the proteomes of Archaea (Figure 2.3:Archaea) confirm the

first reductive tendencies in the emerging world of cellular organisms (48, 83). Our results show

this occurred after viral reductive tendencies were already in place (Figure 2.3:Viruses).

Archaeal f values started to drop at nd = 0.185 and the first domain structure completely lost in

Archaea was the lysozyme-like FSF (d.2.1, nd = 0.185). In contrast, Bacteria and Eukarya

maintained higher representation of FSFs in their proteomes and the total loss of an FSF is seen

at nd = 0.587 in Bacteria (b.37.1) and nd = 0.413 in Eukarya (d.244.1) (Figure 2.3:Bacteria-

Eukarya). The first superkingdom-specific FSF appeared in Bacteria at nd = 0.413 (d.244.1;

f.16.1) and congruently in viruses (b.28.1; b.126.1), Archaea (h.1.19) and Eukarya (g.71.1) at nd

= 0.586 (Figure 2.3).

 26
 
The timeline of domain discovery supports the evolution of viruses and prokaryotes by

reductive evolutionary processes and of eukaryotes by genomic expansion (48, 83). Both viruses

and Archaea, the most ancient organismal groups, suffered an extended history of genomic

reduction and diversified into superkingdoms much later and concurrently with Eukarya (nd =

0.5867). These reductive processes and diversification patterns define the appearance of

taxonomic groups of FSFs and their relative numbers (Figure 2.4). The ABEV is the most

ancient group with nd ranging between 0-1 and a median nd of 0.22 with a total of 229 FSFs (the

third largest after ABE and BE) that are common between the cellular domains and viruses. Its

appearance was followed by the ABE, BEV, and BE taxonomic groups, in that order.

Superkingdom-specific groups appear much later. Interestingly, the appearance of BV, EV and

AV FSFs (Figure 2.4) occurs soon after the appearance of the respective superkingdom-specific

FSFs or after the diversification of the respective superkingdoms. We hypothesize that these are

the FSFs that were discovered when viruses began to infect their hosts and adapted to a parasitic

life style. This occurred when lineages of diversified riboorganisms were already in existence

(read below).

uToLs show viruses predate cellular superkingdoms

We generated uToLs that describe the evolution of proteomes in our total dataset of 1037

riboorganisms and viruses. Intrinsically rooted phylogenomic trees were built using FSFs (total

set: 1,739 FSFs) as distinguishable linearly ordered multistate phylogenomic characters. The

most parsimonious tree reconstruction showed that organisms in Archaea, Bacteria, Eukarya, and

Viruses formed 4 distinct groups, some monophyletic, and that viruses occupied the most basal

position of the uToL (Figure B.1). Inclusion of parasitic (P) and obligate parasitic (OP)

riboorganisms induced a topology of the cellular superkingdoms that favored the canonical
 27
 
rooting of the ToL (44, 45) with the early divergence of Bacteria relative to Archaea and

Eukarya (Figure B.1). Exclusion of P an OP organisms in the total dataset and in equally

sampled sets of riboorganisms (50 Archaea, 50 Bacteria, and 50 Eukarya) and capsid-encoding

organisms (50 Viruses) from superkingdoms produced trees that retained viruses as the most

ancient monophyletic group but placed Archaea as the second oldest (Figure 2.6). The viral

monophyletic group has very little diversity (small branches) compared to the cellular

proteomes. We explain the lack of diversity in viral proteomes due to their prolonged history of

reductive evolution and the fact that the set of FSFs that distinguish viruses are unevenly

distributed in the virosphere. For example, the translation related enzymes are only present in

mimiviruses (and megaviruses). Similarly, other viral specific FSFs are unevenly distributed.

Thus, the virosphere as a group makes the most diverse group on the planet but sets of single

viral proteomes lack that diversity.

In order to evaluate the relative contribution of vertical and horizontal processes of

inheritance in the evolution of proteomes, we calculated ri for each phylogenetic character (ri for

each FSF) used to build the uToL as a relative measure of homoplasy. It has often been proposed

that viruses pickpocket genes from their cellular hosts via HGT (27, 29). Therefore, we

compared the distribution of FSFs in viruses with the distribution of cellular FSFs by plotting

them against their ri values. Both the viruses and cellular FSFs follow a similar bimodal

distribution and do not appear significantly different from each other (Figure 2.5). In contrast,

viral characters (FSFs) are distributed with relatively higher ri values, supporting a better fit of

viral characters to the phylogeny. We thus conclude that better fit of viral characters to the

phylogeny indicate FSFs in viruses are not acquired horizontally from their hosts and both the

viral and cellular FSFs are subject to the similar levels of HGT.
 28
 
A plot that describes the diversity (use) and abundance (reuse) of FSFs (total number of

distinct FSFs versus the total number of FSF domains that are encoded in a proteome) show

viruses have the simplest proteomes, followed progressively by Archaea, Bacteria and Eukarya,

in that order (Figure 2.7). Remarkably, organisms follow a congruent trend towards structural

diversity and organismal complexity. This trend confirms our initial evolutionary model of

proteome growth and again reveals the ancestrality of viruses and Archaea (48, 84).

Viruses enhance planetary biodiversity

The spread (f) of viral FSFs relative to cellular FSFs in individual proteomes of Archaea,

Bacteria and Eukarya appeared significantly biased (Figure 2.8). When compared to cell-specific

FSFs, FSFs shared by viruses and cells were significantly widespread in the proteomes of a

superkingdom. Viruses hold 294, 265, and 239 FSFs in common with Eukarya, Bacteria and

Archaea, respectively. Median f values of these FSFs were considerably higher than those of

corresponding cellular FSFs in Eukarya (0.978 vs. 0.416), Bacteria (0.8826 vs. 0.329), and

Archaea (0.742 vs. 0.514). This bias is remarkable in the case of Eukarya where ~98% of

proteomes are enriched with viral FSFs (Figure 2.8). Archaeal and bacterial proteomes are also

enriched with viral FSFs but at lower levels. Remarkably, patters of enrichment follow patterns

of reductive evolution in the superkingdoms.

The popularity and abundance of viral FSFs in cellular proteomes suggests that viruses

have been a very active and crucial factor in mediating domain transfer between the individual

proteomes and enhancing biodiversity. These domains are present in a wide array of hosts with

remarkable diversity, ranging from very small to very complex organisms, and provide further

support to the ancient and primordial nature of viruses (18). The transfer of domains from virus

to host is a relatively new concept and needs to be explored further. A significant proportion of
 29
 
human and all vertebrate genomes harbor endogenous retroviruses (31, 99). The capsid proteins

and RNA-dependent-RNA-polymerase from viral families, Totiviruses and Partitiviruses have

widespread homologs in the eukaryal superkingdom (100). Together these observations and our

findings support the view that viral-to-host gene transfers are common events and viruses play an

active role in the evolution of cellular organisms.

Functional annotation suggests a late adaptation to the parasitic lifestyle

We studied the molecular functions of 293 (out of 304) FSFs that are present in the viral

dataset using the functional annotation scheme described by Vogel and Chothia (88-90). For the

rest of 11 FSFs, functional annotation was not available. When plotted against time (nd), we note

that the majority of the viral FSFs either appeared very early (nd<0.4) or very late (0.6<nd<1.0)

(Figure 2.9) supporting previous results. We find that most of the viral FSFs perform metabolic

functions, followed by informational FSFs, Intracellular processes, Regulation, General, Other

and Extracellular processes, in that order. This order matches the functional distribution

described previously for the cellular superkingdoms (58) and provides supports to the hypothesis

that viruses coevolved with the cellular ancestors. A significant drop in the number of

FSFs/functions is seen in the nd range 0.4-0.6 which is the period marked by massive gene loss

in both viruses and cellular organisms (Figure 2.3). In contrast, a relatively even distribution of

functions is seen in the nd range 0.6-1.0 which is the period marked by superkingdom

diversification and genome expansion in Eukarya (Figure 2.3) (48). The functions acquired by

viruses during this late period include those related to Extracellular processes (toxin/defense,

immune response, cell adhesion), General (protein interaction, general, ion binding, small

molecule binding) and Other (viral proteins, and proteins with unknown functions) functional

categories (Figure 2.10). We hypothesize that viruses acquired these functions to adapt to the
 30
 
parasitic lifestyle after suffering massive gene loss between nd 0.4-0.6. This is also evident by

the appearance of superkingdom specific taxonomic groups (AV, BV, and EV) after the

appearance of the respective superkingdoms (Figure 2.4). In contrast, the number of FSFs

corresponding to Metabolism, Information, Intracellular processes, and Regulation is lower

compared to nd<0.4 (Figure 2.10).

Effect of HGT

Since HGT-derived domain structures are expected to be over-represented in proteomes

(49), significant enrichment (over-representation) of FSFs in viruses (in the ABEV, BEV, AEV,

ABV, AV, BV, and EV taxonomic groups) are taken as indication that they have acquired FSFs

from their cellular hosts through lateral transfer. We calculated the probability of enrichment of a

particular taxonomic group using the hypergeometric distribution and found that only the ABEV

FSF group was significantly over-represented (P<0.05). The AEV FSF group was also over-

represented but it was not statistically significant (P = 0.29). In contrast, all the other taxonomic

groups were under-represented (Table 2.2). Results suggest viral FSFs did not originate in the

cellular superkingdoms and then transfer to viruses. Significant overrepresentation of the ABEV

group, which (as shown above) represents the most ancient FSF group, and the under-

representation of viral FSFs in all the other (more derived) taxonomic groups suggests common

FSFs have spread mostly by vertical descent and not by HGT.

 31
 
Discussion

Viruses have a negative connotation linked to them. They are ubiquitous parasites (20)

and infect both cellular and acellular life forms. On a planetary scale, they make up the

virosphere, the most abundant and diverse part of our biosphere (13, 15, 101). However, as our

knowledge continues to grow and evidence is presented for the existence of giant viruses (17,

36), the overlap between capsid-encoding and ribosome-encoding organisms grows. Giant

viruses now defy the word ‘virus’ by definition and call for a change in the way we have

perceived viruses in the past.

Consequently, there is an ongoing debate whether to redefine viruses, give them the

status of living beings, and include them in uToL (19, 27, 29, 42, 102). Proponents for the

inclusion argue viruses are a distinct form of life based on their resemblance with numerous

parasitic bacteria. Adversaries of the theory do not find the unique nature of mimiviral genomes

compelling enough but instead argue that viruses have been ‘pick pocketing’ genes from cellular

species and thus are gene robbers (27, 29). In short, they contend that the set of genes of

mimiviruses could be explained by HGT from host cell to viral form because they share

sequence homology with the cellular species. However, the majority of the mimiviral genes lack

clear cellular homologs and only ~10% of the genes are shared between mimiviruses and the

cellular world (42). If viruses are considered gene robbers, then so are numerous parasitic

bacteria with reduced proteomes (42). More importantly, the origin of viral specific genes/FSFs

(our results) remains unanswered if viruses are gene robbers and not worthy of a living status. In

contrast, the popularity of viral FSFs in hosts ranging from very simple to complex organisms

(our results) together with the abundance of VHGs in very diverse set of viruses support the

existence of an ancient viral world (18). In summary, a significant number of FSFs exist in the
 32
 
virosphere, including a lower than expected count for viral specific FSFs. Sequencing of more

viral genomes, especially larger genomes, will probably lead to the identification of more viral

specific FSFs. However, we expect that the proportion and distribution of FSFs in the rest of the

taxonomic groups will not be affected.

The presence of a partial protein translational apparatus alongside promoter elements

structurally equivalent to the eukaryotic TATA-box is consistent with the reductive mode of

evolution for viral proteomes. Reductive tendencies in proteomes have been explained in detail

previously and numerous obligate intracellular parasitic bacteria are known to have gone this

route (48, 83). Our results reveal a massive loosing trend of FSFs in viruses that started very

early in evolution. This trend supports the reductive evolutionary model. Archaea followed the

route soon afterwards and then later on were joined by Bacteria and Eukarya. In light of our

results, mimiviruses should be viewed as the least reduced form of an ancestral virus that

coevolved with the universal common cellular ancestor. Mimivirus thus represents a living fossil.

We have recently shown that the phylogenomic method used for the reconstruction of

proteome trees is robust against unequal sampling and thus the selection of a significantly large

number of bacterial proteomes (in our case 652 bacterial lineages) does not affect the overall

topology of the tree, as long as organisms included exhibit a free-living lifestyle (49). However,

the presence of highly reduced proteomes, such as small obligate intracellular parasitic bacteria

and viruses adds a bias to the analysis of trees because the species with simpler repertoires of

FSFs (i.e., absence of domains) generally occupy the most basal branches of the tree (48, 49).

In our analyses, we used FSFs as characters to distinguish between proteomes. Domains

defined at the FSF level represent a much higher level of structural organization and the

evolutionary impact of HGT is very limited at such higher levels of structural organization (64-
 33
 
67, 87). Furthermore, the extent of homoplasy in our tree measured with ri values (95) for all

FSFs suggest domain structures present in viruses were not acquired from their hosts via HGT.

Moreover, FSFs acquired via HGT are expected to be over-represented in genomes but our

results show that taxonomic groups of FSFs that include viruses are highly under-represented

(49).

While FSFs were not acquired from their hosts, our data suggests virus-to-host HGT has

been pervasive since very ancient times. We show that viral FSFs (common to viruses and

cellular superkingdoms) are significantly more widespread across proteomes than cellular FSFs,

suggesting viruses mediated transfer of these FSFs into cellular hosts. Furthermore, the fact that

the viral FSFs are also universally present in the proteomes of organisms that range from very

simple to very complex, advocates for the primordial and ancient nature of viruses and their

continued impact in evolution of cellular life. Remarkably, one of the most surprising findings to

come out of the human genome project was the discovery that a significant proportion of the

human genome represents remnants of ancient viral infections (103). Viruses are thus a

previously disregarded factor in enhancing biodiversity. They are capable of leaving imprints

even in the genetic make up of higher order organisms such as humans.

 34
 
Conclusions

We show that viruses harbor a significant number of FSFs and suggest that they have

evolved via massive reductive evolutionary processes. In addition, viruses occupy the most basal

branches of the uToL, as a monophyletic and distinct group along with the three other generally

accepted superkingdoms: Archaea, Bacteria, and Eukarya. This placement and the unique and

complex taxonomic distribution of FSFs suggest viruses truly represent a genuine superkingdom.

Based on these findings, we propose that mimiviruses are living fossils of ancestral viruses that

coevolved with primordial cells. We finally highlight the crucial contribution of the virosphere to

biodiversity by mediating HGT between cellular species. We contend that the virosphere has

been playing a pivotal role in the evolution of cellular superkingdoms since the origins of

cellular life. Our analyses provide evidence for the existence of viruses as a distinct form of life

that lacks ribosomes. We propose that viruses should be regarded as symbionts and not parasites

of cellular species, and that parasitic relationships with their hosts are late adaptations. Our

results call for a change in the way we perceive viruses in our world, highlighting the central role

of the virosphere in planetary history.

 35
 
Figures

Figure 2.1 The Venn diagram highlights the distribution of FSFs in the taxonomic groups.
 36
 
Figure 2.2 A phylogenomic tree of protein domain structures describing the evolution of 1,739 FSFs in 1,037

proteomes (4,63,915 steps; CI = 0.051; RI = 0.795; g1 = -0.127). Taxa are FSFs and characters are proteomes.

Terminal leaves of viruses and cellular FSFs were labeled in red and blue respectively.
 37
 
Figure 2.3 Distribution index (f, the number of species using an FSF/total number of species) of each FSF plotted

against relative age (nd, number of nodes from the root/total number of nodes) for the four supergroups (Total) and

individually for Viruses, Archaea, Bacteria, and Eukarya.

 38
 
Figure 2.4 Boxplots displaying distribution of FSFs in viral and non-viral taxonomic groups with respect to age

(nd). Vertical lines within each distribution represent medians. Dotted vertical lines represent important evolutionary

events in protein domain innovation.

 39
 
Figure 2.5 Retention index (ri) of each FSF plotted against relative age (nd). Viral FSFs are colored red and cellular

FSFs are represented in blue. Both groups of FSFs follow a similar distribution and generally the viral FSFs are

distributed with higher ri values supporting a better fit of viral characters to the phylogeny.
 40
 
Figure 2.6 One optimal (P<0.01) most parsimonious phylogenomic tree describing the evolution of 150 FL

riboorganisms (50 each from Archaea, Bacteria, and Eukarya) and 50 viruses generated using the census of

abundance of 1,739 FSFs (1,517 parsimoniously informative sites; 62,061 steps; CI = 0.156; RI = 0.804; g1 = -

0.325). Terminal leaves of Viruses (V), Archaea (A), Eukarya (E) and Bacteria (B) were labeled in red, blue, black

and green respectively.

 41
 
Figure 2.7 FSF diversity (number of distinct FSFs in a proteome) plotted against FSF abundance (total number of

FSF domains that are encoded) for riboorganisms and viruses (50 for each supergroup). Both axes are in logarithmic

scale.
 42
 
Figure 2.8 Viral FSF enrichment. Boxplots comparing the distribution index (f) of viral FSFs with cellular-only

FSFs (shared or not shared with other cellular superkingdom) for each cellular superkingdom. Pie-charts above each

superkingdom represent distribution of FSFs within each superkingdom.

 43
 
Figure 2.9 Functional distribution of viral FSFs in the major functional categories. Histogram comparing the

number of viral FSFs corresponding to major functional categories plotted against time (nd).
 44
 
Figure 2.10 Functional distributions of viral FSFs in minor functional categories. Histograms comparing the number

of viral FSFs corresponding to each of the minor categories within each major functional category.
 45
 
Tables

Table 2.1 List of dsDNA viruses sampled along with the statistics on structural assignments.

No. Common name Family Total FSFs Unique

proteins assigned FSFs

1 Human adenovirus Adenoviridae 36 15 11

2 Heliothis virescens ascovirus Ascoviridae 180 39 24

3e

3 Spodoptera frugiperda Ascoviridae 123 36 22

ascovirus 1a

4 Trichoplusia ni ascovirus 2c Ascoviridae 164 49 29

5 African swine fever virus Asfarviridae 160 55 37

6 Pseudoalteromonas phage Corticoviridae 22 1 1

PM2

7 Ambystoma tigrinum virus Iridoviridae 95 33 23

8 Singapore grouper iridovirus Iridoviridae 162 48 31

9 Infectious spleen and kidney Iridoviridae 125 40 26

necrosis virus

10 Lymphocystis disease virus - Iridoviridae 239 47 30

isolate China

11 Aedes taeniorhynchus Iridoviridae 126 54 35

iridescent virus

12 Frog virus 3 Iridoviridae 99 30 22

13 Invertebrate iridescent virus 6 Iridoviridae 468 77 48

14 Lymphocystis disease virus 1 Iridoviridae 110 36 26

15 Soft-shelled turtle iridovirus Iridoviridae 105 31 22

16 Acanthamoeba polyphaga Mimiviridae 911 530 163

mimivirus
 46
 
Table 2.1 (contd.)

No. Common name Family Total FSFs Unique

proteins assigned FSFs

17 Marseillevirus Mimiviridae 856 264 61

18 Acanthocystis turfacea Phycodnaviridae 860 158 78

Chlorella virus 1

19 Emiliania huxleyi virus 86 Phycodnaviridae 472 105 58

20 Feldmannia species virus Phycodnaviridae 150 55 33

21 Paramecium bursaria Phycodnaviridae 689 155 72

Chlorella virus 1

22 Paramecium bursaria Phycodnaviridae 814 184 76

chlorella virus AR158

23 Paramecium bursaria Phycodnaviridae 849 147 71

chlorella virus FR483

24 Paramecium bursaria Phycodnaviridae 886 209 74

chlorella virus NY2A

25 Ectocarpus siliculosus virus 1 Phycodnaviridae 240 97 42

26 Ostreococcus tauri virus 1 Phycodnaviridae 230 104 56

27 Ostreococcus virus OsV5 Phycodnaviridae 264 100 58

28 Bovine papular stomatitis Poxviridae 131 42 24

virus

29 Amsacta moorei Poxviridae 294 66 37

entomopoxvirus 'L'

30 Camelpox virus Poxviridae 211 93 45

31 Canarypox virus Poxviridae 328 157 50

32 Cowpox virus Poxviridae 233 113 47

33 Crocodilepox virus Poxviridae 173 47 26

 47
 
Table 2.1 (contd.)

No. Common name Family Total FSFs Unique

proteins assigned FSFs

34 Deerpox virus W-1170-84 Poxviridae 170 73 39

35 Ectromelia virus Poxviridae 173 91 45

36 Fowlpox virus Poxviridae 261 118 44

37 Goatpox virus Pellor Poxviridae 150 66 36

38 Lumpy skin disease virus NI- Poxviridae 156 68 36

2490

39 Melanoplus sanguinipes Poxviridae 267 83 37

entomopoxvirus

40 Molluscum contagiosum virus Poxviridae 163 39 27

subtype 1

41 Monkeypox virus Zaire-96-I- Poxviridae 191 93 46

16

42 Myxoma virus Poxviridae 170 78 40

43 Orf virus Poxviridae 130 38 23

44 Pseudocowpox virus Poxviridae 134 40 22

45 Rabbit fibroma virus Poxviridae 165 79 40

46 Sheeppox virus 17077-99 Poxviridae 148 66 36

47 Swinepox virus Poxviridae 150 64 39

48 Tanapox virus Poxviridae 156 60 36

49 Taterapox virus Poxviridae 225 98 45

50 Vaccinica virus Poxviridae 223 98 46

51 Variola virus Poxviridae 197 80 44

52 Yaba monkey tumor virus Poxviridae 140 52 33

53 Yaba-like disease virus Poxviridae 152 60 36

 48
 
Table 2.1 (contd.)

No. Common name Family Total FSFs Unique

proteins assigned FSFs

54 Sulfolobus turreted icosahedral Rudiviridae 36 5 5

55 Bacillus phage Bam35c Tectiviridae 32 6 6

56 Enterobacteria phage PRD1 Tectiviridae 31 7 6

 49
 
Table 2.2 Statistical test for the enrichment of FSFs in viral taxonomic groups using hypergeometric distribution.

Significant comparisons are highlighted in bold font.

Comparison k n M N P-value Comment

ABEV_ABE 229 304 557 1739 9.80E-67 Overrepresented

BEV_BE 33 304 291 1739 0.0012 Underrepresented

BV_B 1 304 163 1739 1.80E-13 Underrepresented

BEV_BE 33 304 291 1739 0.0012 Underrepresented

ABV_AB 2 304 36 1739 0.035 Underrepresented

AV_A 1 304 22 1739 0.079 Underrepresented

AEV_AE 7 304 31 1739 0.29 Overrepresented

EV_E 25 304 335 1739 5.50E-09 Underrepresented

k, number of FSFs in viral taxonomic group; n, total number of FSFs in viral supergroup; M, number of FSFs

cellular taxonomic group; N, total number of FSFs studied (49, 66).

 50
 

CHAPTER 3: FUNCTIONAL MAKE UP OF PROTEOMES IS

CONSERVED1

Introduction

The functional repertoire of a cell is largely embodied in its proteome, the collection of

proteins encoded in the genome of an organism. The molecular functions of proteins are the

direct consequence of their structure and structure can be inferred from sequence using hidden

Markov models of structural recognition (85, 93, 104). Here we analyze the functional

annotation of protein domain structures in almost a thousand sequenced genomes, exploring the

functional and structural diversity of proteomes. We find there is a remarkable conservation in

the distribution of domains with respect to molecular functions they perform in the three

superkingdoms of life. In general, most of the protein repertoire is spent in functions related to

metabolic processes but there are significant differences in the usage of domains for regulatory

and extra-cellular processes both within and between superkingdoms. Our results support the

hypotheses that the proteomes of superkingdom Eukarya evolved via genome expansion

mechanisms that were directed towards innovating new domain architectures for regulatory and

extra/intracellular processes, functions needed for example to maintain the integrity of

multicellular structure or to interact with environmental biotic and abiotic factors (e.g., cell

signaling and adhesion, immune responses, and toxin production). Proteomes of microbial

superkingdoms, Archaea and Bacteria retained fewer numbers of domains and maintained simple

and smaller protein repertoires. Viruses appear to play an important role in the evolution of

                                                                                                               
1
This chapter has been published as manuscript in open access journal Genes (58). The rights to reprint were
retained by the authors and transferred to IEEE for parts of chapter that were communicated as poster paper in a
conference organized by IEEE. The copyright owner, IEEE has given permission to reuse text in thesis/dissertation.
© 2011 IEEE. Reprinted, with permission, from Nasir, A.; Naeem, A.; Khan, M.J; Lopez-Nicora, H.D.; Caetano-
Anolles, G, The functional make up of proteomes is remarkably conserved, Nov. 2011. (105).
 51
 
superkingdoms. We finally identify few genomic outliers that deviate significantly from the

conserved functional design. These include Nanoarchaeum equitans, proteobacterial symbionts

of insects with extremely reduced genomes, Tenericutes and Guillardia theta. These organisms

spend most of their domains on information functions, including translation and transcription,

rather than on metabolism and harbor a domain repertoire characteristic of parasitic organisms.

In contrast, the functional repertoire of the proteomes of the Planctomycetes-Verrucomicrobia-

Chlamydiae (PVC) superphyla was no different than the rest of bacteria, failing to support claims

of them representing a separate superkingdom. In turn, Protista and Bacteria shared similar

functional distribution patterns suggesting an ancestral evolutionary link between these groups.

Our results yield a global picture of the functional organization of proteomes. Results

suggest that the functional structure of proteomes is remarkably conserved across all organisms,

ranging from small bacteria to complex eukaryotes. There is also evidence for the existence of

genomic outliers that deviate from global trends. Here we explore what makes these proteomes

distinct.
 52
 
Materials and Methods

Data retrieval

We used the protein architecture assignments for a total of 965 organisms including 70

Archaea, 651 Bacteria and 244 Eukarya from the total dataset described in chapter 2. Remaining

16 organisms were excluded due to lack of information. Classification of organisms according to

their lifestyles was done manually and resulted in 592 free living (FL), 153 parasitic (P), and 158

obligate parasitic (OP) organisms.

Assigning functional categories to protein domains

The most recent domain functional annotation file for SCOP 1.73 was downloaded from

the SUPERFAMILY web server. This file describes the one-to-one mapping between FSFs and

molecular functions based on the scheme described by Vogel and Chothia (Table 3.1) (88-90).

For each genome we extracted the set of unique FSFs encoded and then mapped them to the 7

major and 50 minor functional categories (Table 3.1). We calculated both the percentage and

actual number of domains using programming implementations in Python 3.1

(http://www.python.org/download/).

Statistical analysis

The statistical significance between the numbers of functional FSFs in FL organisms of

superkingdoms was evaluated by Welch’s ANOVA in SAS (http://www.sas.com/software/sas9),

which is the appropriate test to detect differences between means for groups having unequal

variances (106). We excluded organisms with P and OP lifestyles in order to remove noise from

the data. Additionally, in order to meet asymptotic normality, we used the Log10 transformation

and rescaled the data to 0-7 using the following formula,

Nnormal = [Log10(Nxy) / Log10(Nmax)] * 7

 53
 
In this formular, Nxy is the count of a FSF in x functional category in y superkingdom;

Nmax is the largest value in the matrix; and Nnormal is the normalized and scaled score for FSF x in

y superkingdom.
 54
 
Results

General patterns in the distribution of FSF domain functions

We studied the molecular functions of 1,646 domains defined at the FSF level of

structural abstraction (SCOP 1.73) that are present in the proteomes of a total of 965 organisms

spanning the three superkingdoms. A total of 135 FSFs that could not be annotated were

excluded from the analysis. Out of the 1,646 FSFs included, approximately one-third (32.38%)

performs molecular functions related to Metabolism. Categories Other (16.58%), Intracellular

processes (ICP) (12.63%), Regulation (12.45%), and Information (12.21%) are uniformly

distributed within proteomes. In contrast, General (7.96%) and Extracellular processes (ECP)

(5.77%) are significantly underrepresented compared to the rest (Figure 3.1:A). The total number

of FSFs in each category exhibits the following decreasing trend: Metabolism > Other > ICP >

Regulation > Information > General > ECP. These patterns of FSF number and relative

proteome content are for the most part maintained when studying the functional annotation of

FSFs belonging to each superkingdom (Figure 3.1:B). However, the number of FSFs in each

superkingdom varies considerably and increases in the order Archaea, Bacteria and Eukarya, as

we have shown in earlier studies (48).

The significantly higher number of FSFs devoted to Metabolism is an anticipated result

given the central importance of metabolic networks. However, the much larger number of FSFs

corresponding to functional category Other is quite unexpected. The 273 FSFs belonging to this

category include 200 and 73 FSFs in sub-categories unknown functions and viral proteins,

respectively (Table 3.1). Viruses are generally defined as gene poor biological entities. However,

the number of domains belonging to viral proteins that are present in cellular organisms makes a

noteworthy contribution to the total pool of FSFs (4.43%). Thus, viruses have a much more rich
 55
 
and diverse repertoire of domain structures than previously thought and their association with

cellular life has contributed considerable structural diversity to the proteomic make up.

The numbers of FSFs belonging to categories Regulation, Information, and ICP are

uniformly distributed in proteomes. However, the ECP category is the least represented, perhaps

because this category is the last to appear in evolution (48, 57). Extracellular processes are more

important to multicellular organisms (mainly eukaryotes) than to unicellular organisms.

Multicellular organisms need efficient communication, such as signaling and cell adhesion. They

also trigger immune responses and produce toxins when defending from parasites and pathogens.

These ECP processes, which are depicted in the minor categories of cell adhesion, immune

response, blood clotting and toxins/defense, are needed when interacting with environmental

biotic and abiotic factors and for maintaining the integrity of multicellular structure. These

categories are also present in the microbial superkingdoms but their functional role may be

different than in Eukarya.

We note that current genomic research is highly shifted towards the sequencing of

microbial genomes, especially those that hold parasitic lifestyles and are of bacterial origin. In

fact, 67% of proteomes in our dataset belong to Bacteria. This bias can affect conclusions drawn

from global trends such as those in Figure 3.1:A, including the under-representation of ECP

FSFs, because of their decreased representation in microbial proteomes.

Distribution of FSF domain functions in the three superkingdoms of life

In order to explore whether the overall distribution of general functional categories

differs in organisms belonging to the three superkingdoms, we analyzed proteomes at the species

level and calculated both the percentage and actual number of FSFs corresponding to different

functional repertoires (Figure 3.2).

 56
 
FSF domains follow the following decreasing trend in both the percentage and actual

counts of FSFs, and do so consistently for the three superkingdoms: Metabolism > Information >

ICP > Regulation > Other > General > ECP. Note that trend lines across proteomes seldom

overlap and cross (Figure 3.2). It is noteworthy however that this trend differs from the

decreasing total numbers of FSFs we described above (Figure 3.1). Thus, no correlation should

be expected between the numbers of FSFs for individual proteomes and the total set for each

category. This suggests that variation in functional assignments across proteomes of

superkingdoms may not necessarily match overall functional patterns.

Proteomes in microbial superkingdoms Archaea and Bacteria exhibit remarkably similar

functional distributions of FSFs (Figure 3.2:A). The only exception appears to be the slight

overrepresentation of Regulation FSFs (green trend lines) and underrepresentation of ICP (black

trend lines) in Archaea compared to Bacteria (especially Proteobacteria). These distributions are

clearly distinct from those in Eukarya. Proteomic representations of FSFs corresponding to

Metabolism and Information are decreased while those of all other five functional categories are

significantly and consistently increased (Figure 3.2A). There is also more variation evident in

Eukarya; large groups of proteomes exhibit different patterns of functional use (clearly evident in

Information; red trend lines in Figure 3.2:A).

On the whole, the relative functional make up of the proteomes of individual

superkingdoms appears highly conserved (Figure 3.2:A). There is however considerable

variation in the metabolic functional repertoire of organisms, especially in Bacteria, where

Metabolism ranges 30-50% of proteomic content (100-350 FSFs). This variation is not present in

other functional repertoires.

 57
 
Consequently, tendencies of reduction in the metabolic repertoire are generally offset by

small increases in the representation of the other six repertoires, with the notable exception of

Information. In this particular case, when Metabolism goes down Information goes up. For

example, bacterial proteomes with metabolic FSF repertoires of <45% offset their decrease by a

corresponding increase in Information FSFs (generally from ~20% to ~35%, Figure 3.2:A). In all

superkingdoms, we identify groups of proteomes or few outliers that deviate from the global

trends (vertical dotted lines in Figure 3.2:A). As we will discuss below this is generally a

consequence of reductive evolution imposed by the lifestyle of organisms (discussed in detail

below). Outliers are particularly evident in Bacteria and harbor sharp increases in Information

repertoires, not always associated with corresponding decreases in Metabolism. In Archaea,

decreases of Metabolism are generally offset by increases of the Regulation category, with an

exception in Nanoarchaeum equitans (see below). In Eukarya, decreases in Metabolism go in

hand with decreases in Information, and are correspondingly offset mostly by increases in

Regulation and ECP. Apparently, the advantages of regulatory control (e.g., signal transduction

and transcriptional and posttranscriptional regulation) and multicellularity counteract the

interplay of Metabolism and Information in eukaryotes.

When we look at the actual number of FSFs within each functional repertoire (Figure

3.2:B), we observe a clear trend in domain use that matches the total trend for superkingdoms

described above (Figure 3.1). In most cases, the functional repertoires of Archaea are smaller

than those of Bacteria, and bacterial repertoires are generally smaller than those of Eukarya

(Figure 3.2:B). This holds true for all functional categories. However, the numbers of metabolic

FSFs vary 1.5-4 fold in proteomes of superkingdoms, the change being maximal in Bacteria.

While both proteomes in Eukarya and Bacteria show similar ranges of metabolic FSFs, the
 58
 
repertoire of Archaea is more constrained. Furthermore, FSFs belonging to categories Other and

ECP are significantly higher in Eukarya than in the microbial superkingdoms. These remarkable

observations suggest high conservation in the make up of proteomes of superkingdoms and at the

same time considerable levels of flexibility in the metabolic make-up of organisms. Results also

support the evolution of the protein complements of Archaea and Bacteria via reductive

evolutionary processes and Eukarya by genome expansion mechanisms (48, 83). Reductive

tendencies in microbial superkingdoms do not show bias in favor of any functional category.

Furthermore, enrichment of eukaryal proteomes with viral proteins supports theories that state

that viruses have played an important role in the evolution of Eukarya (107).

Distribution of FSF domain functions in individual phyla/kingdoms

We also explain the functional distribution of FSFs at the phyla/kingdom level for each

superkingdom (Figure 3.2). Plots describing the percentages (Figure 3.2:A) and actual number of

FSFs in proteomes (Figure 3.2:B) highlight the existence of ‘outliers’ (vertical dotted lines in

Figure 3.2:A) that deviate from the global functional trends that are typical of each

superkingdom.

In Archaea, the functional repertoires of the proteomes of Euryarachaeota,

Crenarchaeota, Korarcheota and Thaumarchaeota were remarkably conserved and consistent

with each other. Only N. equitans could be considered an outlier (insets of Figure 3.2). Its

proteome deviates from the global archaeal signature by reducing its proteomic make up (it has

only 200 distinct FSFs) and by exchanging Information for metabolic FSFs. N. equitans is an

obligate intracellular parasite (108) that is part of a new phylum of Archaea, the Nanoarchaeota

(109). N. equitans has many atypical features, including the almost complete absence of operons

and presence of split genes (110), tRNA genes that code for only half of the tRNA molecule
 59
 
(111), and the complete absence of the nucleic acid processing enzyme RNAse P (112). Some of

these features were used to propose that N. equitans is a living fossil (113), represents the root of

superkingdom Archaea and the ToL (114), and is part of a very ancient and yet to be described

superkingdom. Phylogenomic analyses of domain structures in proteomes suggest Archaea is the

most ancient superkingdom (48, 49) and has placed N. equitans at the base of the ToL together

with other archaeal species. Its ancestral nature is therefore in line with the evolutionary and

functional uniqueness of N. equitans and the very distinct functional repertoire we here report.

In Bacteria, the functional repertoires of bacterial phyla were also remarkably conserved.

Only Information and Metabolism showed significantly distinct patterns and considerable

variation in the use of FSFs. Again, decreases in representation of metabolic FSFs were generally

offset by increases in informational FSFs (Figure 3.2:A). Notable outliers include the Tenericutes

and the Spirochetes. As groups, they have the highest relative usage of Information FSFs, which

are clearly offset by a decrease in metabolic FSFs. The Tenericutes is a phylum of Bacteria that

includes class Mollicutes. Members of the Mollicutes are typical obligate parasites of animals

and plants (some of medical significance such as Mycoplasma) that lack cell walls and have

gliding motility. These organisms are characterized by small genome sizes (115) considered to

have evolved via reductive evolutionary processes (116). Because of its unique properties and

history, mycoplasmas have been used recently to produce a completely synthetic genome (117).

There were also clear outliers in the Proteobacteria. These included Candidatus Blochmannia

floridanus (symbiont of ants), Baumannia cicadellinicola (symbiont of sharpshooter insect),

Candidatus Riesia pediculicola, Candidatus Carsonella ruddii (symbiont of sap-feeding insects)

and Candidatus Hodgkinia cicadicola (symbiont of cicadas). These bacteria are generally

endosymbionts of insects (e.g., ants, sharpshooters, psyllids, cicadas) that have undergone
 60
 
irreversible specialization to an intracellular lifestyle. Candidatus Carsonella ruddii has the

smallest genome of any bacteria (118). There were also bacterial proteome groups that were

expected to be outliers but were no different than the rest. Bacteria belonging to the superphylum

Planctomycetes-Verrucomicrobia-Chlamydiae (PVC) are different from other bacterial phyla

because they have a ‘eukaryotic touch’ (119). Indeed, PVC bacteria display genetic and cellular

features that are characteristics of Eukarya and Archaea, including the presence of Histone H1,

condensed DNA surrounded by membrane, alpha-helical repeat domains and beta-propeller folds

that make up eukaryotic-like membrane coats, reproduction by budding, ether lipids and lack of

cell walls (120-122). Due to the unique nature of the PVC superphylum, it was proposed that

these organisms be identified as a separate superkingdom that contributed to the evolution of

Eukarya and Archaea (122). However, ToLs generated from domain structures in hundreds of

proteomes did not dissect the PVC superphylum into a separate group (48, 49). Functional

distributions of FSFs now show PVC proteomes appear no different from the rest of bacteria

(Figure 3.2). These results do not support PVC-inspired theories that explain the diversification

of the three cellular superkingdoms of life.

In contrast to the functional repertoires of bacterial and archaeal phyla, proteomes

belonging to individual kingdoms in Eukarya had functional signatures that were highly

conserved (Figure 3.2:A). However, these signatures differed between groups. Plants and Fungi

had functional representations that were very similar and showed little diversity. In contrast,

Metazoa functional distributions increased the representation of ECP and Regulation FSFs in

exchange of FSFs in Metabolism and Information. Protista had patterns that resemble those of

Plants and Fungi but had widely varying metabolic repertoires, very much like Bacteria. This

possible link between basal eukaryotes and bacteria revealed by our comparative analysis is
 61
 
consistent with the existence of an ancestor of Bacteria and Eukarya and the early rise of

Archaea (48). Only few outliers belonging to kingdoms Fungi (Encephalitozoon cuniculi and

Encephalitozoon intestinalis) and Protista (Guillardia theta) were identified. E. cuniculi and E.

intestinalis are amitochondriate microsporidian parasites with highly reduced genomes (123).

Similarly, Guillardia theta is a nucleomorph that has a highly compact and reduced genome with

loss of nearly all the metabolic genes (124).

When we look at the actual number of FSFs in proteomes of phyla and kingdoms (Figure

3.2:B) we observe that while the overall patterns match those of FSF representation (Figure

3.2:A), FSF number revealed considerable variation in the metabolic repertoire of Protista and

Bacteria. FSFs in these groups typically ranged 130-340, with PVC and Spirochetes exhibiting

the smallest range (130-300 FSFs). In contrast, metabolic repertoires of Archaea and the other

eukaryotic kingdoms typically ranged 200-260 FSFs and 270-350 FSFs, respectively. This

observation is significant. It provides comparative information to support a unique evolutionary

link of phyla within superkingdoms Eukarya and Bacteria. Plots of FSF number also clarified

functional patterns in outliers, revealing they did not have more numbers of FSFs in Information

but rather have reduced metabolic repertoires. This shows parasitic outliers get rid of metabolic

domains and become more and more dependent on host cells.

Effect of organism lifestyle

The analysis thus far revealed the existence of a small group of outliers within each

superkingdom. Manual inspection of lifestyles of these organisms showed that all of these

organisms are united by a parasitic or symbiotic lifestyle. For example, N. equitans is the

smallest archaeal genome ever sequenced and represents a new phylum, the NanoArchaeaota

(109). This organism interacts with Ignicoccus hospitalis, establishing the only known
 62
 
parasite/symbiont relationship of Archaea, and harbors a highly reduced genome (110).

Parasitic/symbiotic relationships with various plants and animals can be found in Tenericutes and

in the endosymbionts of insects that belong to Proteobacteria. Similarly, the Encephalitozoon

species are eukaryotic parasites that lack mitochondria and have highly reduced genomes (123).

E. cunniculi has even a chromosomal dispersion of its ribosomal genes, very much like N.

equitans, and the rRNA of the large ribosomal subunit reduced to its universal core (125). It has

also reduced intergenic spacers and shorter proteins. Similarly, Guillardia theta is a nucleomorph

that has a highly compact and reduced genome with loss of nearly all the metabolic genes (124).

Thus, all outliers exhibit extreme or unique cases of genome reduction.

In order to explore if organisms that engage in parasitic or symbiotic interactions have

general tendencies that resemble those of the outliers, we classified organisms into three different

lifestyles: free living (FL) (592 proteomes), facultative parasitic (P) (153 proteomes), and

obligate parasitic (OP) (158 proteomes). Functional distributions for the 7 major functional

categories for these proteomic sets explained the role of parasitic life on proteomic constitution

(Figure 3.3). Plots of percentages (Figure 3.3:A) and actual number of FSFs in proteomes

(Figure 3.3:B) showed FSF distribution in FL organisms were remarkably homogenous and that

the vast majority of variability within superkingdoms was ascribed to the P and OP lifestyles.

This variability was for the most part explained by a sharp decline in the number of metabolic

FSFs that are assigned to the Metabolism general category (Figure 3.3:B). Plots also support the

hypothesis that parasitic organisms have gone the route of massive genome reduction in a

tendency to lose all of their metabolic genes. This tendency makes them more and more

dependent on host cells for metabolic functions and survival (126, 127).
 63
 
Since maximum variability lies within the proteome repertoires of P and OP organisms

(Figure 3.3) and parasitism/symbiosis in these organisms is the result of secondary adaptations,

the analysis of proteomic diversity in FL organisms allows to test if differences in the functional

repertoires of superkingdoms are indeed statistically significant. The number of domains

corresponding to each major functional category in the proteomes of FL organisms increases in

the order Archaea, Bacteria and Eukarya. When compared to the total proteomic set (Figure 3.2),

Metabolism remains the predominant functional category and a large number of domains in all

the proteomes perform metabolic functions. Again, the proteomes of Eukarya have the richest

FSF repertoires, and those of Archaea the most simple. Analysis of variance showed that the

number of FSFs for each functional repertoire was consistently different between superkingdoms

(P<0.0001; Table 3.2). This supports the conclusions drawn from earlier analyses that the

microbial superkingdoms followed a genome reduction path while Eukarya expanded their

genomic repertoires (48, 83).

Analysis of minor functional categories

The seven major categories of molecular functions map to 50 minor categories (Table

3.1). We explored the distribution of FSFs corresponding to each minor category in

superkingdoms (Figure 3.4). Only category ‘not annotated’ (NONA) was excluded from analysis.

In terms of percentage (Figure 3.4:A), the overall functional signature is split into two

components: prokaryotic and eukaryotic. Prokaryotes spend most of their domain repertoire on

Metabolism and Information whereas Eukarya stand out in ECP (particularly cell adhesion,

immune response), Regulation (DNA binding, signal transduction), and all the minor functional

categories corresponding to ICP and General.

 64
 
In terms of domain counts (Figure 3.4:B), proteomes of Eukarya have the richest

functional repertoires with a significantly large number of FSFs devoted for each minor

functional category. Bacteria and Archaea work with small number of domains. However, the

number of FSFs in Bacteria is significantly higher compared to Archaea (supporting results of

Figures 3.1 and 3.2). These results are consistent with the evolutionary trends in proteomes

described previously (48, 83). Our results support the complex nature of the LUCA (49) and are

consistent with the evolution of microbial superkingdoms via reductive evolutionary processes

and the evolution of eukaryal proteomes by genome expansion (48, 83). It appears that Archaea

went on the route of genome reduction very early in evolution and was followed by Bacteria and

finally Eukarya (48). Late in evolution, the eukaryal superkingdom increased the representation

of FSFs and developed a rich proteome (48). This can explain the relatively huge and diverse

nature of eukaryal proteomes compared to prokaryotic proteomes. Finally, there appears to be no

significant difference in the distributions of FSFs corresponding to Metabolism and Information

between Bacteria and Eukarya except for minor category ‘Translation’ (green trend lines in

Figure 3.4) that is significantly higher in Eukarya compared to Bacteria. This shows that Bacteria

exhibit incredible metabolic and informational diversity despite their reduced genomic

complements. We conclude that the genome expansion in Eukarya occurred primarily for

functions related to ECP, ICP, Regulation and General.

 65
 
Discussion

Reliability of functional annotations and conclusions of this study

Our analysis depends upon the accuracy of assigning structures to protein sequences and

the SCOP protein classification and SUPERFAMILY functional annotation schemes. Databases

such as SCOP and SUPERFAMILY are continuously updated with more and more genomes and

new assignments. We therefore ask the reader to focus on the general trends in the data such as

the exact percentage or numbers of FSFs in each functional repertoire. Trends related to the

number of domains in Archaea relative to Bacteria and Eukarya and the reduction of metabolic

repertoires in parasitic organisms should be considered robust since these have been reliably

observed in previous studies with more limited datasets (48, 83). Biases in sampling of

proteomes in the three superkingdoms is not expected to over or underestimate the remarkably

conserved nature of the functional make up. We show that the conservation of molecular

functions in proteomes is only broken in genomic outliers that are united by parasitic lifestyles.

Thus equal sampling will not significantly alter the global trends described for individual

superkingdoms. In light of our results, organism lifestyle is the only factor affecting the

conserved nature of proteomes. Finally, we propose that lower or higher than expected numbers

of FSFs in any category (subcategory) can be explained either by possible limitations of the

scheme used to annotate molecular functions of FSFs or the simple nature of the functional

repertoire. For example, the number of FSFs in minor category structural proteins (major

category General) is only 7 (Table 3.3) despite the importance of structural proteins in cellular

organization. Table 3.3 lists the description of these FSFs and shows that indeed these FSF

domains play important structural roles. Their limited number indicates that the structural and

functional organization is quite limited and only very few folds are utilized for important
 66
 
structural roles. Another possibility is the ‘hidden’ overlap between FSFs and molecular

functions due to the one-to-one mapping limitations of the SUPERFAMILY functional

annotation scheme. Most of the large FSFs include many FFs and participate in multiple

pathways; for few FSFs a complete functional profile may not be intuitively obvious. This may

be one of the shortcomings of using this functional annotation scheme but dissection of such

detailed functions and pathways is a difficult task and is not described in this study. In summary,

we do not believe that the classification or annotation schemes, despite their limitations, would

undergo serious revisions or weaken our findings.

 67
 
Conclusions

Our analysis revealed a remarkable conservation in the functional distribution of protein

domains in superkingdoms for proteomes for which we have structural assignments. The average

distribution of FSFs in phyla, kingdoms, and superkingdoms reveals that the biggest proportion

of each proteome is devoted in all cases to functions related to metabolism (Figure 3.5).

Phylogenomic analysis has shown that metabolism appeared earlier than other functional group

and their structures were the first to spread in life (57). This would explain the relative large

representation of metabolism in the functional toolkit of cells. Usage of domains related to ECP

and Regulation is significantly higher in Metazoa compared to the rest. This showcases the

importance of regulation signal transduction mechanisms for eukaryotic organisms (128, 129).

Our results support the view that prokaryotes evolved via reductive evolutionary processes

whereas genome expansion was the route taken by eukaryotic organisms. Genome expansion in

Eukarya seems to be directed towards innovation of FSF architectures, especially those linked to

Regulation, ECP and General. Finally, viral structures make up a substantial proportion of

cellular proteomes and appear to have played an important role in the evolution of cellular life.

Organisms with parasitic lifestyles have simple and reduced proteomes and rely on host cells for

metabolic functions. Tenericutes are unique in this regard. They spend most of their proteomic

resources in functions linked to information (e.g., translation, replication). Remarkably, we find

that the conservation of molecular functions in proteomes is only broken in ‘outliers’ with

parasitic lifestyles that do not obey the global trends. We conclude that organism lifestyle is a

crucial factor in shaping the nature of proteomes.

 68
 
Figures

Figure 3.1 Number of FSF domains annotated for each major functional category defined in SCOP 1.73 (A) and in

the three superkingdoms (B). The functional distributions show that coarse-grained functions are conserved across

cellular proteomes and metabolism is the most dominant functional category. Numbers in parentheses indicate the

total number of FSFs annotated in each dataset. The number of FSFs increases in the order Archaea, Bacteria and

Eukarya.
 69
 
Figure 3.2 The functional distribution of FSFs in individual proteomes of the three superkingdoms. Both the

percentage (A) and actual FSF numbers (B) indicate conservation of functional distributions in proteomes and the

existence of considerable functional flexibility between superkingdoms. Dotted vertical lines indicate genomic

outliers. Insets highlight the interplay between Metabolism (yellow trend lines) and Information (red trend lines) in

N. equitans.
 70
 
Figure 3.3 The functional distribution of FSFs with respect to organism lifestyle. Both the percentage (A) and actual

FSF numbers (B) indicate that obligate parasitic (OP) and facultative parasitic (P) organisms exhibit considerable

variability in their metabolic repertoires (yellow trend lines) that is offset by corresponding increases in the

Information FSFs (red trend lines).

 71
 
Figure 3.4 The percentage (A) and number (B) of FSFs in minor functional categories across superkingdoms.

Archaea (A) and Bacteria (B) spend most of their proteomes in functions related to Metabolism and Information

whereas Eukarya (E) stands out in the minor categories of Regulation, General, ICP and ECP. In turn, the number

of FSFs increases in the order Archaea, Bacteria and Eukarya. Eukaryal proteomes have the richest functional

repertoires for Regulation, Other, General, ICP and ECP. The number of metabolic FSFs in Bacteria appears not to

be significantly different from Eukarya but still greater than Archaea. Translation (green trend lines) is the only

minor category in Information significantly enriched in eukaryotes compared to prokaryotes.

 72
 
Figure 3.4 (contd.)
 73
 
Figure 3.5 Average distribution of FSFs in phyla, kingdom, and superkingdoms suggest conservation of functional

design in proteomes. Numbers in parentheses indicate total number of proteomes analyzed for each phyla/kingdom.
 74
 
Tables

Table 3.1 Mapping between the major and minor functional categories for 1,781 protein domains defined in SCOP

1.73 and the number of FSFs corresponding to each minor category. m/tr, metabolism and transport.

Major category Minor categories No. of FSF domains

Metabolism (533 FSFs) Energy 54

Photosynthesis 20

E- transfer 31

Amino acids m/tr 20

Nitrogen m/tr 1

Nucleotide m/tr 30

Carbohydrate m/tr 30

Polysaccharide m/tr 21

Storage 0

Coenzyme m/tr 50

Lipid m/tr 17

Cell envelope m/tr 8

Secondary metabolism 11

Redox 55

Transferases 29

Other enzymes 156

General (131 FSFs) Small molecule binding 27

Ion binding 13

Lipid/membrane binding 4

Ligand binding 3

General 28

Protein interaction 49

Structural protein 7
 75
 
Table 3.1 (contd.)

Major category Minor categories No. of FSF domains

Information (201 FSFs) Chromatin structure 7

Translation 92

Transcription 24

DNA replication/repair 68

RNA processing 10

Nuclear structure 0

Other (273 FSFs) Unknown function 200

Viral proteins 73

Extracellular processes (95 FSFs) Cell adhesion 31

Immune response 19

Blood clotting 5

Toxins/defense 40

Intracellular processes (208 FSFs) Cell cycle, Apoptosis 20

Phospholipid m/tr 6

Cell motility 20

Trafficking/secretion 0

Protein modification 35

Proteases 52

Ion m/tr 21

Transport 54

Regulation (205 FSFs) RNA binding, m/tr 19

DNA-binding 66

Kinases/phosphatases 15

Signal transduction 53

Other regulatory function 34

Receptor activity 18
 76
 
Table 3.2 The comparison of functional categories across superkingdoms using Welch’s ANOVA.

Major category F-ratio DF P-value*

Metabolism 350.21 2 <0.0001

Information 582.28 2 <0.0001

ICP 1271.32 2 <0.0001

Regulation 966.75 2 <0.0001

Other 520.97 2 <0.0001

General 1043.76 2 <0.0001

ECP 263.44 2 <0.0001

*All the P-values are significant at 0.05.

 77
 
Table 3.3 Names and description of FSF domains corresponding to minor category structural proteins in the major

category General.

No. SCOP Id FSF Id Description

1 103589 g.71.1 Mini-collagen I, C-terminal domain

2 49695 b.11.1 Gamma-crystallin-like

3 51269 b.85.1 Anti-freeze protein (AFP) III-like domain

4 56558 d.182.1 Baseplate structural protein gp11

5 58002 h.1.6 Chicken cartilage matrix protein

6 58006 h.1.7 Assembly domain of catrillage oligomeric matrix protein

7 75404 d.213.1 Vesiculovirus (VSV) matrix proteins

 78
 

CHAPTER 4: STRUCTURE-FUNCTION PARTNERSHIP

EMPOWERS PHYLOGENETIC ANALYSIS AND DESCRIBES

THE PROTEOMIC MAKE UP

In our analyses, we used the structural information of hundreds of proteomes to

reconstruct the phylogenomic history of protein domains and organisms. We included viruses

with medium-to-very-large proteomes into our analyses and compared them to the cellular

organisms. To our knowledge, this is the first exercise that makes extensive use of molecular

data to study the evolution of viruses on a scale that is comparable to the cellular organisms.

Additionally, we assigned molecular functions to protein domains (grouped into FSFs) and

studied the proteomic make up. While structural information proved useful for the reconstruction

of phylogenies, functional assignments described the nature of hundreds of proteomes. This

composite exercise of linking structure and function proved highly useful and yielded significant

insights into the evolution of organisms, highlighting the conserved nature of proteomes. The

main findings of this study are numbered below:

1. The virosphere harbors a significant number of protein domains including many that are

absent in the cellular world.

2. Viruses evolved via massive reductive evolutionary processes that explain their highly

reduced genomes.

3. Viruses are more ancient than the cellular superkingdoms and predate the LUCA.

4. Viruses mediate gene transfer between cellular organisms and enhance planetary

biodiversity.
 79
 
5. The parasitic lifestyle of viruses is a late adaptation and results from massive gene loss

suffered during the early phases of the evolutionary timeline.

6. Functional studies suggest that organisms are united by a conserved functional

distribution of molecular functions.

7. Parasitic organisms deviate from the conserved trends and harbor reduced genomes with

loss of nearly all the metabolic genes.

8. Prokaryotes (Archaea & Bacteria) evolve by genome reduction (like viruses) whereas

Eukarya evolve by genome expansion.

9. Functional distributions in the genomes of Bacteria and Eukarya suggest an ancestral link

between the two superkingdoms and points to an early divergence of Archaea.

10. Organism lifestyle affects the proteomic make up.

Rationale for using molecular structure to infer phylogenies

We used molecular structure to generate trees that are rooted, i.e., they are statements that

impose an evolutionary arrow, from origins of change in their root branches to very recent

changes in its leaves. These trees describe the evolution of protein domain structures (i.e., ToDs)

and proteomes (i.e., ToPs). They are not phenetic statements (48, 49, 74, 81, 82, 87). While they

are built from multistate or quantitative valued characters, speciation in trees fulfills a molecular

clock that is compatible with paleobiology and the geological record (84). ToPs produce ToLs

that are congruent with traditional classification (49).

Rationale for using one-to-one functional mapping

We assigned a single molecular function to each FSF. Because larges FSFs include many

FFs and participate in multiple pathways, this assignment does not present a complete profile of

molecular functions (49, 58, 105). However, dissection of such detailed annotations is a difficult
 80
 
task that is not considered in our study. The focus on one-to-one mapping between structures and

functions enables us to compare the distributions of functions across hundreds of proteomes and

when linked with evolutionary information embedded in the rooted trees reveals ancestral and

derived functions.

Our analyses depend upon the schemes used for the assignment of structures and functions to

proteomes. These results should be considered robust unless the aforementioned schemes

undergo unlikely and drastic revisions.

 81
 

REFERENCES
1. Wessner DR (2010) Discovery of the giant mimivirus. Nature Education 3: 61.

2. Gibbs AJ, Calisher CH & Garcı́a-Arenal F (1995) Molecular basis of virus evolution (Cambridge university press,

London).

3. Lin CL, Chung CS, Heine HG & Chang W (2000) Vaccinia virus envelope H3L protein binds to cell surface

heparan sulfate and is important for intracellular mature virion morphogenesis and virus infection in vitro and in

vivo. J Virol 74: 3353-3365.

4. Pagaling E, et al (2007) Sequence analysis of an archaeal virus isolated from a hypersaline lake in inner

mongolia, china. BMC Genomics 8: 410.

5. La Scola B, et al (2008) The virophage as a unique parasite of the giant mimivirus. Nature 455: 100-104.

6. Mc Grath S, Fitzgerald GF & van Sinderen D (2007) Bacteriophages in dairy products: Pros and cons. Biotechnol

J 2: 450-455.

7. Prangishvili D, Forterre P & Garrett RA (2006) Viruses of the archaea: A unifying view. Nat Rev Microbiol 4:

837-848.

8. Pearson H (2008) 'Virophage' suggests viruses are alive. Nature 454: 677.

9. Claverie JM & Abergel C (2009) Mimivirus and its virophage. Annu Rev Genet 43: 49-66.

10. Abroi A & Gough J (2011) Are viruses a source of new protein folds for organisms? - virosphere structure space

and evolution. Bioessays 33: 626-635.

11. Anderson NG, Cline GB, Harris WW & Green JG (1967) in Transmission of viruses by the water route. ed Berg

G (Interscience, New York), pp 75-88.

12. Breitbart M & Rohwer F (2005) Here a virus, there a virus, everywhere the same virus?. Trends Microbiol 13:

278-284.

13. Bergh O, Borsheim KY, Bratbak G & Heldal M (1989) High abundance of viruses found in aquatic

environments. Nature 340: 467-468.

14. Rohwer F & Thurber RV (2009) Viruses manipulate the marine environment. Nature 459: 207-212.

15. Suttle CA (2007) Marine viruses--major players in the global ecosystem. Nat Rev Microbiol 5: 801-812.

16. Raoult D, et al (2004) The 1.2-megabase genome sequence of mimivirus. Science 306: 1344-1350.
 82
 
17. Arslan D, Legendre M, Seltzer V, Abergel C & Claverie JM (2011) Distant mimivirus relative with a larger

genome highlights the fundamental features of megaviridae. Proc Natl Acad Sci U S A 108: 17486-17491.

18. Koonin EV, Senkevich TG & Dolja VV (2006) The ancient virus world and evolution of cells. Biol Direct 1: 29.

19. Raoult D & Forterre P (2008) Redefining viruses: Lessons from mimivirus. Nat Rev Microbiol 6: 315-319.

20. Koonin EV & Yutin N (2010) Origin and evolution of eukaryotic large nucleo-cytoplasmic DNA viruses.

Intervirology 53: 284-292.

21. Prangishvili D, Stedman K & Zillig W (2001) Viruses of the extremely thermophilic archaeon sulfolobus.

Trends Microbiol 9: 39-43.

22. Koonin EV & Dolja VV (2006) Evolution of complexity in the viral world: The dawn of a new vision. Virus Res

117: 1-4.

23. Forterre P (2006) The origin of viruses and their possible roles in major evolutionary transitions. Virus Res 117:

5-16.

24. Bandea CI (1983) A new theory on the origin and the nature of viruses. J Theor Biol 105: 591-602.

25. Hendrix RW, Lawrence JG, Hatfull GF & Casjens S (2000) The origins and ongoing evolution of viruses.

Trends Microbiol 8: 504-508.

26. Moreira D (2000) Multiple independent horizontal transfers of informational genes from bacteria to plasmids

and phages: Implications for the origin of bacterial replication machinery. Mol Microbiol 35: 1-5.

27. Moreira D & Brochier-Armanet C (2008) Giant viruses, giant chimeras: The multiple evolutionary histories of

mimivirus genes. BMC Evol Biol 8: 12.

28. Koonin EV, Senkevich TG & Dolja VV (2009) Compelling reasons why viruses are relevant for the origin of

cells. Nat Rev Microbiol 7: 615; author reply 615.

29. Moreira D & Lopez-Garcia P (2009) Ten reasons to exclude viruses from the tree of life. Nat Rev Microbiol 7:

306-311.

30. Weiss RA (2006) The discovery of endogenous retroviruses. Retrovirology 3: 67.

31. Griffiths DJ (2001) Endogenous retroviruses in the human genome sequence. Genome Biol 2: REVIEWS1017.

32. Stanley WM (1935) Isolation of a crystalline protein possessing the properties of tobacco-mosaic virus. Science

81: 644-645.
 83
 
33. Iyer LM, Aravind L & Koonin EV (2001) Common origin of four diverse families of large eukaryotic DNA

viruses. J Virol 75: 11720-11734.

34. Iyer LM, Balaji S, Koonin EV & Aravind L (2006) Evolutionary genomics of nucleo-cytoplasmic large DNA

viruses. Virus Res 117: 156-184.

35. Van Etten JL (2003) Unusual life style of giant chlorella viruses. Annu Rev Genet 37: 153-195.

36. La Scola B, et al (2003) A giant virus in amoebae. Science 299: 2033.

37. La Scola B, Marrie TJ, Auffray JP & Raoult D (2005) Mimivirus in pneumonia patients. Emerg Infect Dis 11:

449-452.

38. Koonin EV (2005) Virology: Gulliver among the lilliputians. Curr Biol 15: R167-9.

39. Zauberman N, et al (2008) Distinct DNA exit and packaging portals in the virus acanthamoeba polyphaga

mimivirus. PLoS Biol 6: e114.

40. Suhre K, Audic S & Claverie JM (2005) Mimivirus gene promoters exhibit an unprecedented conservation

among all eukaryotes. Proc Natl Acad Sci U S A 102: 14689-14693.

41. Ghedin E & Claverie JM (2005) Mimivirus relatives in the sargasso sea. Virol J 2: 62.

42. Claverie JM & Ogata H (2009) Ten good reasons not to exclude giruses from the evolutionary picture. Nat Rev

Microbiol 7: 615; author reply 615.

43. Ludmir EB & Enquist LW (2009) Viral genomes are part of the phylogenetic tree of life. Nat Rev Microbiol 7:

615; author reply 615.

44. Woese CR & Fox GE (1977) Phylogenetic structure of the prokaryotic domain: The primary kingdoms. Proc

Natl Acad Sci U S A 74: 5088-5090.

45. Woese C (1998) The universal ancestor. Proc Natl Acad Sci U S A 95: 6854-6859.

46. Delsuc F, Brinkmann H & Philippe H (2005) Phylogenomics and the reconstruction of the tree of life. Nat Rev

Genet 6: 361-375.

47. Zuckerkandl E & Pauling L (1965) Molecules as documents of evolutionary history. J Theor Biol 8: 357-366.

48. Wang M, Yafremava LS, Caetano-Anolles D, Mittenthal JE & Caetano-Anolles G (2007) Reductive evolution of

architectural repertoires in proteomes and the birth of the tripartite world. Genome Res 17: 1572-1585.

49. Kim KM & Caetano-Anolles G (2011) The proteomic complexity and rise of the primordial ancestor of

diversified life. BMC Evol Biol 11: 140.

 84
 
50. Moreira D & Lopez-Garcia P (2005) Comment on "the 1.2-megabase genome sequence of mimivirus". Science

308: 1114; author reply 1114.

51. Mount DW (2004) Bioinformatics: Sequence and Genome Analysis, (Cold Spring Harbor Laboratory Press, Cold

Spring Harbor (NY)).

52. Rokas A, Williams BL, King N & Carroll SB (2003) Genome-scale approaches to resolving incongruence in

molecular phylogenies. Nature 425: 798-804.

53. Tekaia F, Lazcano A & Dujon B (1999) The genomic tree as revealed from whole proteome comparisons.

Genome Res 9: 550-557.

54. Clarke GD, Beiko RG, Ragan MA & Charlebois RL (2002) Inferring genome trees by using a filter to eliminate

phylogenetically discordant sequences and a distance matrix based on mean normalized BLASTP scores. J Bacteriol

184: 2072-2080.

55. Gerstein M & Hegyi H (1998) Comparing genomes in terms of protein structure: Surveys of a finite parts list.

FEMS Microbiol Rev 22: 277-304.

56. Chothia C, Gough J, Vogel C & Teichmann SA (2003) Evolution of the protein repertoire. Science 300: 1701-

1703.

57. Caetano-Anolles G, Wang M, Caetano-Anolles D & Mittenthal JE (2009) The origin, evolution and structure of

the protein world. Biochem J 417: 621-637.

58. Nasir A, Naeem A, Khan MJ, Lopez-Nicora HD & Caetano-Anolles G (2011) Functional annotation of protein

domains reveals remarkable conservation in the functional make up of proteomes across superkingdoms. Genes 2:

869-911.

59. Murzin AG, Brenner SE, Hubbard T & Chothia C (1995) SCOP: A structural classification of proteins database

for the investigation of sequences and structures. J Mol Biol 247: 536-540.

60. Riley M & Labedan B (1997) Protein evolution viewed through escherichia coli protein sequences: Introducing

the notion of a structural segment of homology, the module. J Mol Biol 268: 857-868.

61. Caetano-Anolles D, Kim KM, Mittenthal JE & Caetano-Anolles G (2011) Proteome evolution and the metabolic

origins of translation and cellular life. J Mol Evol 72: 14-33.

62. Ponting CP & Russell RR (2002) The natural history of protein domains. Annu Rev Biophys Biomol Struct 31:

45-71.
 85
 
63. Andreeva A, et al (2008) Data growth and its impact on the SCOP database: New developments. Nucleic Acids

Res 36: D419-25.

64. Gough J (2005) Convergent evolution of domain architectures (is rare). Bioinformatics 21: 1464-1471.

65. Choi IG & Kim SH (2007) Global extent of horizontal gene transfer. Proc Natl Acad Sci U S A 104: 4489-4494.

66. Forslund K, Henricson A, Hollich V & Sonnhammer EL (2008) Domain tree-based analysis of protein

architecture evolution. Mol Biol Evol 25: 254-264.

67. Yang S & Bourne PE (2009) The evolutionary history of protein domains viewed by species phylogeny. PLoS

One 4: e8378.

68. Morrison DA (2009) Why would phylogeneticists ignore computerized sequence alignment?. Syst Biol 58: 150-

158.

69. Anisimova M, Cannarozzi GM & Liberles DA (2010) Finding the balance between the mathematical and

biological optima in multiple sequence alignment. Trends Evol Biol 2: e7.

70. Maddison WP (1993) Missing data versus missing characters in phylogenetic analysis. Syst Biol 42: 576-581.

71. De Laet J (2005) in Parsimony, phylogeny and genomics. ed Albert VA (Oxford University Press, Oxford), pp

81-116.

72. Sober E & Steel M (2002) Testing the hypothesis of common ancestry. J Theor Biol 218: 395-408.

73. Caetano-Anolles G, et al (2009) The origin and evolution of modern metabolism. Int J Biochem Cell Biol 41:

285-297.

74. Wang M & Caetano-Anolles G (2009) The evolutionary mechanics of domain organization in proteomes and the

rise of modularity in the protein world. Structure 17: 66-78.

75. Sun FJ & Caetano-Anolles G (2008) The origin and evolution of tRNA inferred from phylogenetic analysis of

structure. J Mol Evol 66: 21-35.

76. Kim KM, Sung S, Caetano-Anolles G, Han JY & Kim H (2008) An approach of orthology detection from

homologous sequences under minimum evolution. Nucleic Acids Res 36: e110.

77. Theobald DL (2010) A format test of the theory of universal common ancestry. Nature 465: 219-222.

78. Zwickl DJ & Hillis DM (2002) Increased taxon sampling greatly reduces phylogenetic error. Syst Biol 51: 588-

598.

79. Kluge AG & Farris JS (1969) Quantitative phyletics and the evolution of anurans. Syst Zool 18: 1-32.
 86
 
80. Huelsenbeck JP & Nielsen R (1999) Effect of nonindependent substitution on phylogenetic accuracy. Syst Biol

48: 317-328.

81. Caetano-Anolles G & Caetano-Anolles D (2003) An evolutionarily structured universe of protein architecture.

Genome Res 13: 1563-1571.

82. Wang M & Caetano-Anolles G (2006) Global phylogeny determined by the combination of protein domains in

proteomes. Mol Biol Evol 23: 2444-2454.

83. Wang M, Kurland CG & Caetano-Anolles G (2011) Reductive evolution of proteomes and protein structures.

Proc Natl Acad Sci U S A 108: 11954-11958.

84. Wang M, et al (2011) A universal molecular clock of protein folds and its power in tracing the early history of

aerobic metabolism and planet oxygenation. Mol Biol Evol 28: 567-582.

85. Gough J & Chothia C (2002) SUPERFAMILY: HMMs representing all proteins of known structure. SCOP

sequence searches, alignments and genome assignments. Nucleic Acids Res 30: 268-272.

86. Karplus K (2009) SAM-T08, HMM-based protein structure prediction. Nucleic Acids Res 37: W492-497.

87. Caetano-Anolles D, Kim KM, Mittenthal JE & Caetano-Anolles G (2011) Proteome evolution and the metabolic

origins of translation and cellular life. J Mol Evol 72: 14-33.

88. Vogel C, Berzuini C, Bashton M, Gough J & Teichmann SA (2004) Supra-domains: Evolutionary units larger

than single protein domains. J Mol Biol 336: 809-823.

89. Vogel C, Teichmann SA & Pereira-Leal J (2005) The relationship between domain duplication and

recombination. J Mol Biol 346: 355-365.

90. Vogel C & Chothia C (2006) Protein family expansions and biological complexity. PLoS Comput Biol 2: e48.

91. Tatusov RL, et al (2003) The COG database: An updated version includes eukaryotes. BMC Bioinformatics 4:

41.

92. Claverie JM, et al (2006) Mimivirus and the emerging concept of "giant" virus. Virus Res 117: 133-144.

93. Gough J, Karplus K, Hughey R & Chothia C (2001) Assignment of homology to genome sequences using a

library of hidden markov models that represent all proteins of known structure. J Mol Biol 313: 903-919.

94. Swofford DL (2002) Phylogenomic Analysis Using Parsimony and Other Programs (PAUP*) Ver 4.0b10.

(Sinauer, Sunderland, MA).

95. Farris JS (1989) The retention index and homoplasy excess. Syst Zool 18: 406-407.
 87
 
96. Eriani G, Delarue M, Poch O, Gangloff J & Moras D (1990) Partition of tRNA synthetases into two classes

based on mutually exclusive sets of sequence motifs. Nature 347: 203-206.

97. Ibba M, Curnow AW & Soll D (1997) Aminoacyl-tRNA synthesis: Divergent routes to a common goal. Trends

Biochem Sci 22: 39-42.

98. O'Donoghue P & Luthey-Schulten Z (2003) On the evolution of structure in aminoacyl-tRNA synthetases.

Microbiol Mol Biol Rev 67: 550-573.

99. Boeke JD & Stoye JP (1997) in Retroviruses, eds Coffin JM, Hughes SH & Varmus HE (Cold Spring Harbor

Laboratory Press, Cold Spring Harbor (NY)), pp 343-435.

100. Liu H, et al (2010) Widespread horizontal gene transfer from double-stranded RNA viruses to eukaryotic

nuclear genomes. J Virol 84: 11876-11887.

101. Edwards RA & Rohwer F (2005) Viral metagenomics. Nat Rev Microbiol 3: 504-510.

102. Ludmir EB & Enquist LW (2009) Viral genomes are part of the phylogenetic tree of life. Nat Rev Microbiol 7:

615; author reply 615.

103. Belshaw R, et al (2004) Long-term reinfection of the human genome by endogenous retroviruses. Proc Natl

Acad Sci U S A 101: 4894-4899.

104. Wilson D, et al (2009) SUPERFAMILY--sophisticated comparative genomics, data mining, visualization and

phylogeny. Nucleic Acids Res 37: D380-6.

105. Nasir A, Naeem A, Khan MJ, Lopez-Nicora HD, Caetano-Anolles G (2011) The functional make up of

proteomes is remarkably conserved. IEEE International Conference on Bioinformatics and Biomedicine Workshops,

1023-1025, Nov. 2011 (© 2011 by IEEE).

106. Welch BL (1938) The significance of the difference between two means when the populationvariances are

unequal. Biometrika 29: 350-362.

107. Koonin EV, Wolf YI, Nagasaki K & Dolja VV (2008) The big bang of picorna-like virus evolution antedates

the radiation of eukaryotic supergroups. Nat Rev Microbiol 6: 925-939.

108. Das S, Paul S, Bag SK & Dutta C (2006) Analysis of nanoarchaeum equitans genome and proteome

composition: Indications for hyperthermophilic and parasitic adaptation. BMC Genomics 7: 186.

109. Huber H, et al (2002) A new phylum of archaea represented by a nanosized hyperthermophilic symbiont.

Nature 417: 63-67.

 88
 
110. Waters E, et al (2003) The genome of nanoarchaeum equitans: Insights into early archaeal evolution and

derived parasitism. Proc Natl Acad Sci U S A 100: 12984-12988.

111. Randau L, Munch R, Hohn MJ, Jahn D & Soll D (2005) Nanoarchaeum equitans creates functional tRNAs

from separate genes for their 5'- and 3'-halves. Nature 433: 537-541.

112. Randau L, Schroder I & Soll D (2008) Life without RNase P. Nature 453: 120-123.

113. Di Giulio M (2006) Nanoarchaeum equitans is a living fossil. J Theor Biol 242: 257-260.

114. Di Giulio M (2007) The tree of life might be rooted in the branch leading to nanoarchaeota. Gene 401: 108-

113.

115. Woese CR, Maniloff J & Zablen LB (1980) Phylogenetic analysis of the mycoplasmas. Proc Natl Acad Sci U S

A 77: 494-498.

116. Chambaud I, et al (2001) The complete genome sequence of the murine respiratory pathogen mycoplasma

pulmonis. Nucleic Acids Res 29: 2145-2153.

117. Gibson DG, Smith HO, Hutchison CA,3rd, Venter JC & Merryman C (2010) Chemical synthesis of the mouse

mitochondrial genome. Nat Methods 7: 901-903.

118. Nakabachi A, et al (2006) The 160-kilobase genome of the bacterial endosymbiont carsonella. Science 314:

267.

119. Forterre P & Gribaldo S (2010) Bacteria with a eukaryotic touch: A glimpse of ancient evolution?. Proc Natl

Acad Sci U S A 107: 12739-12740.

120. Devos DP & Reynaud EG (2010) Evolution. intermediate steps. Science 330: 1187-1188.

121. Kamneva OK, Liberles DA & Ward NL (2010) Genome-wide influence of indel substitutions on evolution of

bacteria of the PVC superphylum, revealed using a novel computational method. Genome Biol Evol 2: 870-886.

122. Santarella-Mellwig R, et al (2010) The compartmentalized bacteria of the planctomycetes-verrucomicrobia-

chlamydiae superphylum have membrane coat-like proteins. PLoS Biol 8: e1000281.

123. Katinka MD, et al (2001) Genome sequence and gene compaction of the eukaryote parasite encephalitozoon

cuniculi. Nature 414: 450-453.

124. Douglas S, et al (2001) The highly reduced genome of an enslaved algal nucleus. Nature 410: 1091-1096.
 89
 
125. Peyretaillade E, et al (1998) Microsporidian encephalitozoon cuniculi, a unicellular eukaryote with an unusual

chromosomal dispersion of ribosomal genes and a LSU rRNA reduced to the universal core. Nucleic Acids Res 26:

3513-3520.

126. Martin W & Herrmann RG (1998) Gene transfer from organelles to the nucleus: How much, what happens, and

why?. Plant Physiol 118: 9-17.

127. Keeling PJ & Slamovits CH (2005) Causes and effects of nuclear genome reduction. Curr Opin Genet Dev 15:

601-608.

128. Burglin TR (2008) Evolution of hedgehog and hedgehog-related genes, their origin from hog proteins in

ancestral eukaryotes and discovery of a novel hint motif. BMC Genomics 9: 127.

129. Ingham PW, Nakano Y & Seger C (2011) Mechanisms and functions of hedgehog signalling across the

metazoa. Nat Rev Genet 12: 393-406.

 90
 

APPENDIX A
Table A.1 Top ten technical and conceptual advantages of trees from protein domain abundance over standard

molecular phylogenetic trees.

Properties Trees of genes or Trees of proteomes Trees of domain

genomes (ToGs) (ToPs) structures (ToDs)

Used to derive a ToL Direct inference Direct inference Indirect inference

Phylogenetic model Parametric or non- Non-parametric Non-parametric

parametric models of model of proteome model of domain

sequence or gene evolution evolution

evolution (occurrence) (abundance) (abundance)

Focus System System Parts (i.e., domains)

Characters Genomic features Proteome Domain structures

(e.g., nucleotide site, (repertoire of domain (e,g, defined at FF,

gene) structures) FSF, F, etc)

Character states Occurrence of Abundance of Abundance of

genomic feature (e.g., domain structures domain structures

nucleotide in (e.g., defined at FF, (e,g, defined at FF,

nucleotide site, FSF, F, etc) FSF, F, etc)

presence of gene)

Leaves Genomic features of Proteomes of Domain structures

different organisms different organisms in all organisms

Trees are rooted No Yes Yes

without outgroups

Homology requires Yes No No

alignment

Tree imbalance Moderate Moderate (No, results High (No, results

(technical artifact) from abundance) from abundance)

 91
 
Table A.1 (contd.)

Properties Trees of genes or Trees of proteomes Trees of domain

genomes (ToGs) (ToPs) structures (ToDs)

Problem of Yes No No

inapplicables in

phylogenetic analysis

Saturation and Yes (deep signal is No (deep signal is No (deep signal is

phylogenetic signal lost) conserved) conserved)

loss

Affected by domain Yes (genes encoding No No

rearrangements multidomain proteins

are not evolutionary

units)

False assumption of Yes (e.g., nucleotide No (when used with No (the tree reveals

evolutionary age is heterogeneous) ToDs) the age of parts)

homogeneity of parts

Orthology/paralogy Yes No No

Convergence/HGT Maximum Minimum Minimum

Complicating effect Yes Yes No (taxa is finite)

of taxon sampling

False assumption of Yes (existence of Yes (domains interact No (proteomes

character structure) with each other) evolve largely

independence independently)
 92
 

APPENDIX B
Figure B.1 uToL reconstructed from the total proteome dataset. One optimal (P<0.01) most parsimonious

phylogenomic tree describing the evolution of 981 riboorganisms (652 Bacteria, 70 Archaea, and 259 Eukarya) and

56 viruses (51 NCLDV and 5 viruses from Archaea, Bacteria and Eukarya) generated using the census of abundance

of 1,739 FSFs (1,696 parsimoniously informative sites; 2,63,635 steps; CI = 0.043; RI = 0.800; g1 = -0.112).

Terminal leaves of Viruses (V), Archaea (A), Bacteria (B), and Eukarya (E) were colored red, blue, green and black

respectively.