draft - university of toronto t-space · 12 characteristic simplification of metabolic pathways...

Draft

Ureaplasma diversum protein interaction networks: evidence of horizontal gene transfer and evolution of

reduced genomes among the Mollicutes

Journal: Canadian Journal of Microbiology

Manuscript ID cjm-2018-0688.R3

Manuscript Type: Article

Date Submitted by the Author: 09-Apr-2019

Complete List of Authors: Silva, Joana; Universidade Federal da Paraiba, Biologia MolecularMarques, Lucas; Universidade Federal da BahiaTimenetsky, Jorge; Instituto de Ciências Biomédicas, Universidade de São Paulo, Microbiologiade Farias, Savio; Universidade Federal da Paraiba, Molecular Biology

Keyword: protein interaction networks, horizontal gene transfer, Ureaplasma, Mollicutes

Is the invited manuscript for consideration in a Special

Issue? :Not applicable (regular submission)

https://mc06.manuscriptcentral.com/cjm-pubs

Canadian Journal of Microbiology

Draft

1

1 Ureaplasma diversum protein interaction networks: evidence of horizontal gene

2 transfer and evolution of reduced genomes among Mollicutes

3

4 Joana Kästle Silva1, Lucas Miranda Marques2,3, Jorge Timenetsky3, Sávio Torres de

5 Farias1

6

7 1 Department of Molecular Biology, Federal University of Paraíba, João Pessoa, Brazil

8 2 Multidisciplinary Institute of Health, Universidade Federal da Bahia, Vitória da

9 Conquista, Brazil

10 3 Department of Microbiology, Institute of Biomedical Science, University of São

11 Paulo, São Paulo, Brazil

12

13 Correspondence and reprints: Savio Torres de Farias – [email protected]

14

15 Authors correspondence

16

17 Joana Kästle – [email protected]

18 Lucas Miranda Marques – [email protected]

19 Jorge Timenetsky – [email protected]

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

Page 1 of 39



mailto:[email protected]



Draft

2

1

2 Abstract 3 Ureaplasma diversum is a member of the Mollicutes class and causes urogenital tract infection in

4 cattle and small ruminants. Studies indicate that the process of horizontal gene transfer, the

5 exchange of genetic material among different species, has a crucial role in mollicute evolution,

6 affecting the group’s characteristic genomic reduction process and simplification of metabolic

7 pathways. Using bioinformatics tools and the String database of known and predicted protein

8 interactions, we constructed the protein-protein interaction network of U. diversum and compared

9 it with the networks of other members of the Mollicutes class. We also investigated horizontal

10 gene transfer events in sub-networks of interest involved in purine and pyrimidine metabolism

11 and urease function, chosen due to their intrinsic importance for host colonization and virulence.

12 We identified horizontal gene transfer events among Mollicutes and from Ureaplasma to S.

13 aureus and Corynebacterium, bacterial groups that colonize the urogenital niche. The overall

14 tendency of genome reduction and simplification in the Mollicutes echoes in their protein

15 interaction networks, which tend to be more generalistic and less “selective.” Our data suggest

16 that the process was permitted (or enabled) by an increase in host dependence and the available

17 gene repertoire in the urogenital tract shared via horizontal gene transfer.

18

19 Keywords: protein interaction network - horizontal gene transfer – Ureaplasma - Mollicutes.

20

21

22

Page 2 of 39



Draft

3

1 1. Introduction

2

3 Mollicutes are characterized by reduced genomes and as a result, by simplified

4 metabolic pathways (Dordet-Frisoni et al. 2014). Lacking major genes required for

5 survival and reproduction, these cell wall-deficient bacteria rely upon the host for

6 metabolite demand. Studies with 16S rRNA (Woese et al. 1985) determined that the

7 Mollicutes class evolved from a gram-positive bacterium with a low G+C content through

8 a genomic reduction process. The genes "lost" during evolutionary history of the

9 Mollicutes are responsible for amino acid, cofactors and fatty acid biosynthesis (Razin

10 1985; Maniloff 1996) and can be completely absent in some species of this taxa.

11 The host-dependent life-style of mycoplasmas and ureaplasmas reflects the

12 characteristic simplification of metabolic pathways that accompanies the group’s

13 evolutionary history (Glass et al. 2017). A continuous process of genome reduction,

14 which left the Mollicutes with a natural minimal set of genes, causes such simplification,

15 showcased in incomplete or even absent metabolic pathways. The adaptation to a highly

16 specific and stable niche of multicellular hosts permits the intake of precursors,

17 metabolites and even finished products through the cellular membrane, making it possible

18 for ureaplasmas and mycoplasmas to conserve energy and gene repertoire in biosynthetic

19 pathways. The metabolic processes and gene expression patterns in mycoplasmas are

20 curiously similar to those proposed by studies that aim at constructing artificial, minimal

21 cells: transcription, protein translation and cellular replication through DNA and RNA

22 precursor intake takes place, but little else (Coyle et al. 2016).

23 An important incomplete pathway in Ureaplasma and Mycoplasma is the de novo

24 synthesis of purines and pyrimidines (Bizarro and Schuck 2007). This is embedded in one

25 interesting phenomenon observed in these bacterial groups: function of metabolic

26 pathways is observed even without the corresponding enzymes and proteins being

27 encoded by the genome (Marques et al. 2016). In addition to this, experimental assays

28 have shown that Mollicutes are capable of presenting enzyme function without the

29 specific enzyme being present in their genomic annotation (Arraes et al. 2007). This has

30 raised questions about alternative pathway use and metabolite intake to supply the

31 nutritional and metabolic demand of these bacteria. How ureaplasma species use

32 alternative pathways in their metabolisms and supply the demand for proteins and

33 enzymes for which corresponding genes are not present in their genomes remains an open

34 question.

Page 3 of 39



Draft

4

1 As it is the case of many mycoplasmas, Ureaplasma diversum is a substantial

2 pathogen of the reproductive tract in cattle, especially affecting young females (Cardoso

3 et al. 2000). Among the symptoms caused by U. diversum infection are granular

4 vulvovaginitis, salpingitis, endometritis, abortion and neonatal death, in addition to

5 infertility and seminal vesiculitis in males (Doig et al. 1979; Sanderson et al. 2000). These

6 characteristics make the infection caused by U. diversum and other mycoplasmosis an

7 important economic factor due to its impact on livestock development and production. U.

8 diversum prevalence in cattle was reported in South America (Cardoso et al. 2000;

9 Oliveira Filho et al. 2005), Australia (Argue et al. 2013), United States (Rae et al. 1993;

10 Sanderson et al. 2000) and Kenya (Mulira et al. 1989), reflecting damage caused by this

11 pathogen on an international scale. One of Ureaplasma’s distinctive virulence factors is

12 urease, an enzyme involved in urea hydrolysis that supplies the energetic demand (Ligon

13 and Kenny 1991). Urea hydrolysis generates ammonia, which in turn is responsible for

14 tissue damage caused in the host due to pH changes (Kokkayil and Dhawan 2015). The

15 urease cluster in U. diversum and human-colonizing ureaplasmas share the same

16 organization (Neyrolles et al. 1996) and this similarity is conserved in codifying

17 sequence-level coding regions (Marques et al. 2016).

18 To have a better understanding of protein function it is important to consider that

19 proteins rarely execute their biological role alone, being, instead, connected in functional

20 networks. These networks are constructed from different metabolic pathways and

21 perform their actions collectively. It has been reported that the functions fulfilled by a

22 protein are less due to their amino acid sequence and more to the variety of interactions a

23 specific protein carries with other proteins used by the organism, suggesting that complex

24 cellular processes are better understood via large-scale protein-protein interaction

25 networks (PPI) (Raman 2010). By studying these protein interaction networks, it is

26 possible to identify key-proteins that regulate diverse cellular mechanisms and metabolic

27 pathways. PPI networks have been used for determining unknown protein function (Deng

28 et al. 2003) and enabling a broad vision of cellular processes when analyzed on the

29 genomic scale. Studying PPI networks more closely, it is possible to determine the

30 proteins occupying central positions of metabolic pathways and how their presence in the

31 genomic repertoire controls cell function (Wuchty 2014). It is also enables the

32 determination of functional modules in metabolic pathways (Erten et al. 2009),

33 pinpointing essential proteins for cell survival and their integration with others, which

Page 4 of 39



Draft

5

1 can then be used to indicate metabolic pathway robustness and even antibiotic treatment

2 targets.

3 In microbiology, PPI networks were used in studies regarding Bacillus

4 licheniformis (Han et al. 2016), Escherichia coli and Saccharomyces cerevisae (Wuchty

5 and Uetz 2014) and the human gastric pathogen Helicobacter pylori (Rain et al. 2001) as

6 target species, but to our knowledge no PPI network study was performed with

7 ureaplasmas until this moment. With access to the genomic annotation of Ureaplasma

8 diversum (Marques et al. 2016), this became possible. The comparison of PPI networks

9 of different species on a large scale can be used to observe evolutionary processes (Erten

10 et al. 2009), (Consortium 2011). The evolution of the Mollicute minimal genome has been

11 targeted in many studies (Dordet-Frisoni et al. 2014), (Juhas et al. 2011), (Delaye and

12 Moya 2010), but none of them considered the characteristic gene loss of this group based

13 on the complexity of their PPI networks. We aimed to test if the protein-protein

14 interaction networks reflect the genome reduction process in Ureaplasma and

15 Mycoplasma, forcing them to compensate for the gene loss during their evolutionary

16 history.

17

18 2. Materials and Methods

19

20 2.1 Construction of PPI networks in Ureaplasma diversum

21 Protein sequences obtained from the genomic annotation of U. diversum (Marques

22 et al. 2016) were submitted to String (Szklarczyk et al. 2014) to construct initial PPI

23 networks. The networks were clustered according to cellular mechanisms and metabolic

24 pathways of their respective proteins. Still using String, we observed the general

25 information about the nature of the predicted interactions generated by the PPI database,

26 assembled in categories ranging from curated experimental databases to predictions based

27 on gene neighborhood, co-occurrence and homology. The networks were visualized in

28 Cytoscape (Shannon et al. 2003) for analysis of individual characteristics of each network

29 and their correlated parameters. The NetworkAnalyzer plugin (Assenov et al. 2007) was

30 used for analyzing network topology parameters. The ClusterOne plugin was used for

31 identification and visualization of interaction clusters inside the global network and to

32 perform sub-network arrangements.

33 To ensure high confidence in the predicted interactions, a 0.4 confidence interaction

34 score cut-off was used while assembling the predicted networks in the String database.

Page 5 of 39



Draft

6

1 Predicted interactions with a confidence score lower than medium were excluded from

2 further analysis. For network robustness assessment the NetworkRandomizer plugin

3 (Tosadori et al. 2016) was used on Cytoscape to generate randomized networks whose

4 centrality parameters were compared to those of the predicted U. diversum network for

5 network validation.

6

7 2.2 Construction of PPI networks of species used in network complexity comparison

8 Ureaplasma urealyticum (txid565575), Ureaplasma parvum (txid273119),

9 Mycoplasma genitalium (txid243273), Mycoplasma pneumoniae (txid722438),

10 Mycoplasma hominis (txid347256), Haloplasma contractile (txid1033810),

11 Mycobacterium tuberculosis (txid419947), Corynebacterium urealyticum (txid504474),

12 Corynebacterium glucurolyticum (txid548477) and Staphylococcus aureus (txid93062)

13 were used for network comparison. These species were selected due to their phylogenetic

14 relationship both in and outside of the Mollicutes. The proteomes of the species used for

15 comparison were taken from the curated UniProt (The UniProt Consortium, 2017)

16 database. The proteins were also visualized oinn the Kyoto Encyclopedia of Genes and

17 Genomes (Kanehisa and Goto, 2000) platform and assigned to their corresponding

18 metabolic pathways. The proteomes were submitted to String for genomic PPI network

19 prediction and observation of general characteristics. The networks were then submitted

20 to Cytoscape for topological and conformational analysis. The ClusterOne plugin was

21 used for identifying interaction clusters inside the global network and for sub-network

22 arrangement.

23

24 2.3 Comparison of PPI networks of U. diversum and reference species

25 The PPI networks of the reference species U. urealyticum, U. parvum, M.

26 genitalium, M. pneumoniae, M. hominis, H. contractile, M. tuberculosis, C. urealyticum

27 and S. aureus were loaded on Cytoscape where the topological characteristics and

28 conformation of the PPI networks were assessed and the comparative profile of the sub-

29 networks between these species and U. diversum was established.

30

31 2.4 Construction of Mollicute phylogeny according to sub-networks complexity and

32 interactions

33 The proteins composing PPI sub-networks of U. diversum and reference species of

34 interest were organized on a multi-FASTA file and submitted to Mafft (Katoh and

Page 6 of 39



Draft

7

1 Standley 2013) for global alignment and submitted to ProtEST (Cuff et al. 2000) for the

2 generation of the best-fitted evolutionary models. A maximum likelihood analysis was

3 performed using PhyML for tree inference on the T-REX webserver (Culman et al. 2009).

4 Still using T-REX, we assessed horizontal gene transfer rates among the studied species

5 regarding the sub-networks of metabolic and pathogenic interest, the sub-networks of

6 interest for U. diversum regard purine and pyrimidine metabolism and urease function.

7 The phylogenetic trees obtained by these two distinct methods were then compared to a

8 16S rRNA based tree for validation. The 16S rRNA sequences were obtained from the

9 genomic annotation for U. diversum and from the NCBI database of nucleotide sequences

10 (Coordinators 2016) for the remaining species. The best suited evolutionary model was

11 calculated on Mega (Tamura et al. 2007) and the parameters obtained from this analysis

12 were inserted on T-REX (Culman et al. 2009) for the 16S rRNA tree construction and

13 visualization.

14

15

16

17 3. Results

18

19 3.1 Characteristics and general parameters of U. diversum PPI network

20 The PPI network of U. diversum was constructed using a medium confidence score

21 cut-off of 0.4. From the 10.404 predicted interactions of the global network 20.71%

22 (2.155 interactions) obtained a very high confidence score (>0.9), 35.86% (3.732

23 interactions) had a high confidence score (0.6-0.899) and 43% (4.517 interactions) had a

24 medium confidence score (0.4-0.599).

25 The 301 nodes of the PPI network of U. diversum were categorized and colored

26 according to protein function: proteins involved in gene expression are shown in blue;

27 proteins involved inter-cellular processes, protein transport, protein fate and cellular

28 envelope construction and maintenance shown in purple; proteins with regulatory

29 function, proteins involved in the metabolism of cofactors and carriers, fatty acid and

30 energy metabolism related functions are depicted in orange; the proteins and enzymes

31 shown in yellow are related to purine and pyrimidine metabolism and the proteins

32 involved with the intermediary metabolism and central metabolism of nitrogen (including

33 the urease cluster) are illustrated in pink (Figure 1).

Page 7 of 39



Draft

8

1 Connectivity distribution, also known as node degree (k), refers to the number of

2 nodes to which each node in a determined network is connected. It can also be seen as a

3 representation of the interconnections inside the network. The node degree of the PPI

4 network of U. diversum follows a power law distribution, a characteristic behavior of

5 biological networks (Figure 2A). When distribution of nodes inside a network follows a

6 power law, the majority of the nodes interact with few proteins and a low number of

7 central nodes (the so-called hub proteins) concentrate the bulk of the network’s

8 connections (Raman 2010). In a biological context, this means that few proteins in U.

9 diversum interact with many nodes (proteins), providing the foundation of cellular

10 processes, while the majority of the proteins agglomerate in smaller groups compromising

11 specific interactions. This behavior can also be observed in the remaining topological

12 parameters. Closeness centrality increases according to the number of neighbors to which

13 a node is connected (Figure 2B). Proteins with a high connection to their proteomic

14 neighborhood will automatically have a higher closeness centrality than proteins with few

15 interactions. Similarly, the average clustering coefficient also follows the number of

16 neighboring nodes connected to a given node (Figure 2C). The shortest path length refers

17 to the velocity in which the information flows inside a PPI network, representing the

18 shortest path between two nodes. In the U. diversum PPI network, the most frequent

19 shortest path length is two nodes, showing that the information flux is very quick inside

20 the PPI network (Figure 2D). In a metabolic context, this means that the PPI network of

21 U. diversum shows a tendency to a high connection among the integrating proteins,

22 suggesting a rapid message flux in the metabolism (Raman 2010).

23 Network robustness can be assessed by comparing the conformation and centrality

24 parameters of a PPI network to networks randomly generated by its data (Tosadori et al.

25 2016). This can also be used to validate biological networks as a representation of organic

26 processes. The Cytoscape NetworkRandomizer plugin generates random networks using

27 a simple shuffle algorithm in the studied network (Tosadori et al. 2016). To validate the

28 PPI network of U. diversum, 10 random networks were generated using their interactions

29 and nodes as base and comparison of the subsequent centrality measures were performed.

30 Only the clustering coefficient and network centrality scores were used due to the rest of

31 the parameters remaining the same since the random networks were constructed based on

32 the predicted network data (Supplementary information – Table 1). The clustering

33 coefficient was considerably lower in the random networks compared to the predicted

34 PPI network of U. diversum (0.613 in U. diversum and an average of 0.230 in the random

Page 8 of 39



Draft

9

1 networks). This translates into a lower tendency among the nodes of the random networks

2 to form clusters. The network centrality scores of the random networks are also smaller,

3 showing a decrease from 0.359 in U. diversum to an average of 0.074 in the random

4 networks. These scores show that the random networks are poorly connected, which

5 reinforces their classification as non-representative of a biological system and scores the

6 PPI network of U. diversum as a robust organic network.

7

8 3.2 Central Proteins of the U. diversum PPI network

9 The two most used central parameters to determine the relevance of a node to its

10 surrounding PPI network are 1) the node degree (k) which calculates the number of

11 connections a given node has to other nodes in the network, being therefore a local

12 centrality measurement, and 2) the node intersection, a parameter that represents the

13 number of shortest paths of the network crossing it (Azevedo, Moreira-Filho, 2015),

14 depicting the information flow inside the network. Organic networks such as PPI

15 networks respond poorly to the removal of nodes with the high centrality measurements,

16 often referred to as hub proteins, since much of the global network’s foundation relies on

17 them and their role in “stitching” the neighborhood proteins together. Proteins and

18 enzymes with critical function in specific metabolic pathways and cellular processes, such

19 as urease and its role in supplying the energetic demand of ureaplasmas, are also prone to

20 cause network collapse and cell death upon removal from the PPI network.

21 To identify protein hubs in the U. diversum PPI network, node degree (k) and node

22 intersection of all nodes composing the network were accessed. To pinpoint the cellular

23 processes in which they are involved we searched for experimental studies in the literature

24 and in the Kyoto Encyclopedia of Genes and Genomes (KEGG) database (Kanehisa,

25 GOTO, 2000) (Table 1).

26 3.3 Comparison of the global PPI network of U. diversum and correlated species

27 We aimed to determine whether the genomic reduction, simplification and their

28 related generalization of the metabolic pathways in Ureaplasma (and Mycoplasma) are

29 reflected in the conformation and general parameters of their PPI networks. To do so, we

30 compared the PPI networks of these Mollicutes to the networks of distantly related species

31 with a higher metabolism complexity, such as S. aureus and Corynebacterium species

32 (Table 2).

Page 9 of 39



Draft

10

1 The scores analyzed in this study show a tendency among the Mollicutes to interact

2 more in clusters than phylogenetically distant species, which is shown by their higher

3 clustering coefficient rates. The average clustering coefficient among the Mollicutes is

4 0.550, while Corynebacterium species show an average of 0.357 and the analyzed S.

5 aureus populations, averaged 0.361. This can be due to the generalization of function in

6 the remaining proteins left after the genomic reduction process in the Mollicutes, forcing

7 them to interact at higher frequency rates. The studied Mycoplasma and Ureaplasma

8 species also show a higher centrality index in their PPI networks and increased

9 connectivity scores (depicted by their higher average number of neighbor scores). The

10 genomic reduction and simplification of the metabolic networks in the Mollicutes class

11 appears to be followed by a generalization process in protein function and interaction.

12

13 3.4 Horizontal Gene Transfer in Ureaplasma

14 Our horizontal gene transfer results based on the PPI network of U. diversum shows

15 gene transfer events in the sub-networks related to urease function and purine and

16 pyrimidine metabolism (Figure 3). The bioinformatic analysis indicates two events of

17 horizontal gene transfer in the urease metabolic pathway (Figure 3A), depicting a flow of

18 the urease operon between Ureaplasma and Corynebacterium and Ureaplasma and S.

19 aureus. In the purine subnetwork, analysis showed the horizontal transfer between

20 Mycoplasma and Ureaplasma (Figure 3B), two Mollicutes genera sharing the same

21 ecological niche in their multicellular hosts. In the pyrimidine metabolism subnetworks,

22 two events of horizontal gene transfer were indicated by our analysis (Figure 3C).

23

24 3.5 Horizontal Gene Transfer and network structure

25 The urease PPI network conformation is highly similar in the analyzed species,

26 showing slight differences in the proteins that interact with the urease structural subunits

27 and accessory proteins. All the investigated species present the structural urease subunits

28 (ureA, ureB, ureC) as well as the accessory proteins (ureE, ureF, ureG) in their PPI

29 network. This can be observed in the urease intersection network (Supplementary

30 information – Figure 1). The confidence index of the interactions among the urease

31 proteins was predicted as very high (>0.9) in the studied species, depicted by their colored

32 edges. The comparison of the urease subnetwork in U. diversum, U. urealyticum and U.

33 parvum (Supplementary information - Figure 2) illustrates a high similarity among the

34 interacting partners, edges and positions inside the PPI network. The network

Page 10 of 39



Draft

11

1 conformation is also similar, reinforced by proximal interaction parameters among the

2 ureaplasma species that differ in the proteins comprising the individual networks. Figure

3 4 shows the urease subnetwork of the species targeted in Ureaplasma horizontal gene

4 transfer events: U. diversum (A), C. glucurolyticum (B), C. urealyticum (C) and S. aureus

5 (D). The interactions of both networks are similar, differentiating in the proteins that

6 interact with the urease structural and accessory proteins. The comparison of these

7 networks and their associated topological parameters (Supplementary information - Table

8 2) suggest that the urease subnetwork is conserved structurally in these species.

9 Our analysis also showed a horizontal gene transfer event during the evolution of

10 Mycoplasma and Ureaplasma in proteins involved in purine metabolism, driving us to

11 focus our studies on their related subnetworks and corresponding network parameters. In

12 figure 5, the purine subnetworks of the species with suggested gene flow by lateral gene

13 transfer are shown, U. diversum (A), M. genitalium (B), M. pneumoniae (C) and M.

14 hominis (D). One of the differences in the PPI network composition in the studied

15 Mycoplasma species is the absence of the deoD gene in the subnetwork of M. hominis

16 (the only representative of the hominis group in this study), which occurs in the

17 subnetworks of M. genitalium and M. pneumoniae. The studied Ureaplasma species (U.

18 diversum, U. urealyticum, U. parvum) present similar purine metabolism PPI networks

19 in structure and composition (Supplementary information – Figure 3). Mycoplasma and

20 Ureaplasma share some of the proteins in their purine metabolism PPI network. As we

21 did in the analysis of the urease PPI networks, we also calculated the topological

22 parameters in the constructed purine metabolism subnetworks of the species involved in

23 the horizontal gene transfer events (Supplementary information - Table 3). Again, the

24 correlated network parameters remained very similar between the analyzed species.

25 The pyrimidine subnetwork analysis suggested two horizontal gene transfer events:

26 the first from Mycoplasma to Ureaplasma and the second from Mycoplasma to

27 Corynebacterium. In Figure 6, the pyrimidine subnetworks of the species involved in

28 horizontal gene transfer events are depicted.

29 The PPI networks are similar between these groups, which is reflected in their

30 corresponding network parameters (Supplementary information - Table 4). The centrality

31 measurements of the pyrimidine metabolism PPI networks in Mycoplasma, Ureaplasma

32 and Corynebacterium share similarities regarding clustering coefficient, network

33 centrality and characteristic path length parameters, which are almost identical in the

34 analyzed species (Supplementary information - Table 4). Similar to the purine

Page 11 of 39



Draft

12

1 metabolism networks, the pyrimidine PPI networks of the studied species share protein

2 composition and topological parameters.

3 The horizontal gene transfer events in the urease, purine and pyrimidine metabolism

4 subnetworks appointed by our bioinformatic analysis is reflected in a highly similar PPI

5 network composition and organization between the studied groups. The centrality

6 parameters suggest a conservation in network topology taking place during Ureaplasma,

7 Mycoplasma, Corynebacterium and S. aureus evolution. This can be due to the similar

8 lifestyles shared by these urogenital tract colonizers.

9

10 4. Discussion

11

12 4.1 U. diversum reduced metabolism

13 Mollicutes are known for their reduced genomes and simplified metabolic

14 pathways resulting from gene loss in the course of their evolutionary history. Gene loss

15 is often regarded as a deleterious or detrimental process to the organism undergoing it or

16 a neutral factor resulting simply from genetic drift, but this is not always the case. In fact,

17 the loss of specific gene groups can prove to be a major evolutionary mechanism, driving

18 species to colonize new environments, saving energy by ‘throwing out’ dispensable genes

19 and adjusting the metabolism to accommodate niche adaptation. Spontaneous loss of

20 genes was observed to increase bacterial fitness in different growth conditions in previous

21 studies (Koskiniemi et al. 2012). Thus, gene loss might play a major role in the transition

22 to new lifestyles such as host colonization and pathogenicity development in a similar

23 way to the acquisition of specific genes through processes such as, for instance, horizontal

24 gene transfer events (Albalat and Cañestro 2016). Both gene loss and acquisition act as

25 valuable sources of genetic diversity, which in turn can contribute to adaptive phenotypic

26 variety.

27 Regarding the gene losses of the mycoplasma group, some species lost almost all

28 genes involved in specific biosynthetic pathways. This distinctive characteristic also

29 applies to M. pneumoniae and M. genitalium. Both lost all genes of the amino acid

30 production pathway, relying upon the host or environment for meeting their amino acid

31 needs (Razin 2006). The group also lost the majority of genes compromised with fatty

32 acid synthesis and cofactor production (Weisburg et al. 1989). Comparatively, U.

Page 12 of 39



Draft

13

1 diversum also has an incomplete fatty acid metabolism and the majority of U. diversum

2 annotated proteins are involved in gene expression and regulation.

3 The genomic reduction in U. diversum can be easily visualized in the number of

4 nodes observed in the global PPI network carrying all of its known proteins. The process

5 of genomic reduction in U. diversum was accompanied by a rearrangement in its protein

6 interaction networks. Our data suggests an increase in connectivity between the proteins

7 left behind after the progressive gene loss process during the evolutionary history of this

8 bacterium. The indicated connectivity increase can be related to the maintenance of

9 essential metabolic pathways where remaining proteins had to partially or completely

10 assume the function executed by lost proteins during the genomic reduction process. Our

11 data indicates a mechanism of adaptation to genomic reduction by rearrangement of

12 protein-protein interaction networks.

13

14 4.2 Central Proteins of the U. diversum PPI network: signs of horizontal gene

15 transfer importance for minimal genomes

16 It is interesting to note that all proteins with the highest connections in the global

17 network of U. diversum perform functions in DNA replication and transcription,

18 contributing to the flow of genetic information inside and outside of the bacterial cell.

19 Among these, RecA (recombinase A), dnaK and a DNA helicase (UUR10_0577 also

20 known as PcrA) are also involved in stress response pathways, activation of bacterial

21 S.O.S response, homologous recombination and horizontal gene transfer.

22 The role of RecA, a DNA-dependent ATPase and recombinase, in homologous

23 recombination has positioned it among the main factors involved in the capture of

24 exogenous genetic material, especially in the processes of conjugation and

25 transformation, known mechanisms of bacterial horizontal gene transfer (Cox, 1991). The

26 process of DNA transfer mediated by RecA requires the participation of several proteins,

27 about 40 in Bacillus subtilis (Saito; Akamatsu, 2006), which may explain the high

28 connectivity of this node in the PPI network of U. diversum. In B. subtilis and E. coli,

29 RecA interacts with exogenous single-stranded DNA (ssDNA), forming nucleofilaments

30 capable of recognizing recombination sites in the bacterial chromosome, where it

31 promotes the exchange of genetic information (Prentiss et al., 2015). This process has

32 been observed in a variety of bacterial species, including Neisseria gonorrhoeae,

33 Hemophilus influenzae, Streptococcus pneumoniae, Streptococcus mutans and

Page 13 of 39



Draft

14

1 Helicobacter pylori, suggesting that the mechanism for incorporation of exogenous DNA

2 acquired through RecA-mediated horizontal gene transfer is of great importance for

3 introducing genetic diversity in these bacteria (Michod et al., 2008). Ishag et al. (2017)

4 experimentally determined that RecA activity increases homologous recombination rates

5 in Mycoplasma hyorhinis and considering the high connectivity of RecA in the PPI

6 network of U. diversum, which may also be the case in ureaplasmas and other

7 mycoplasmas.

8 RecA is a highly conserved protein in biological systems with widespread

9 functional homologs among eubacteria, archaebacteria and eukaryotes (Brendel et al.

10 1997), illustrating the crucial role it plays in cell-life maintenance. In eubacteria, invariant

11 RecA segments compromise ATP, DNA and LexA-binding sites, where a high amino

12 acid residue conservation assures proper RecA function and regulation (Karlin and

13 Brocchieri 1996). Recent studies have expanded known RecA functions in an array of

14 cellular processes, pointing out its role in signaling pathways leading to programmed cell

15 death (PCE) in bacteria (Peeters and de Jonge 2017) and the underlying molecular

16 mechanisms involved in its contribution to genomic integrity (Bell and Kowalczykowski

17 2016).

18 PcrA is a helicase responsible for relieving tension in the DNA strands generated

19 by replication and transcription processes. Studies indicate that helicases, including PcrA,

20 regulate the rate of genetic recombination in B. subtilis (Ehrlich; Petit, 2002) promoted

21 by proteins such as RecA. A similar regulatory system may take place in U. diversum,

22 with the RecA-mediated homologous recombination activity being under the control of

23 PcrA.

24

25 4.3 Horizontal gene transfer in Ureaplasma

26 Horizontal gene transfer has been discussed as a possible explanation for the

27 distinct metabolic and phylogenetic characteristics among the Mollicutes. Jain et al.

28 (1999) reported that the abundant frequency of horizontal transfer events taking place

29 between prokaryotes occur at higher rates among operational (housekeeping) genes than

30 genes involved in genome expression. The degenerative evolution mechanism of the

31 Mollicutes does not contribute to addition of new genes (Cordova et al. 2016), suggesting

32 that horizontal gene transfer may play an important part in gene acquisition and genomic

33 diversity in this group. Xiao et al. (2011) observed horizontal gene transfer among

Page 14 of 39



Draft

15

1 different serotypes of human ureaplasmas. García-Castillo et al. (2008) showcased the

2 horizontal acquisition of biofilm formation genes by ureaplasma strains from other

3 bacterial groups and previous studies reported horizontal gene transfer occurring in

4 species occupying the same ecological niche (Sirand-Pugnet et al. 2007). We also know

5 that M. pulmonis strains share genes horizontally (Teachman et al. 2002). Horizontal gene

6 transfer also seems to be more prevalent in some Mollicutes than other bacteria taxa. As

7 an example, 18% of the M. agalactiae genome was acquired by this mechanism from the

8 Mycoides group (Dordet-Frisoni et al. 2014). Even though the exact mechanisms by

9 which gene transfer among Mollicutes occur are still unclear, this process appears to not

10 only guarantee gene acquisition and exchange among Mollicutes (Cordova et al. 2016)

11 but also upgrades the metabolic machinery of Mollicute species for an increased

12 survivability against host defense mechanisms and could also enable the colonization of

13 new hosts (Sirand-Pugnet et al. 2007). During the analysis and interpretation of horizontal

14 gene transfer events, the observation of the niches occupied by the species taking part in

15 it is an important tool revealing the ecological contextualization of the obtained data.

16 Species that occupy the same niche, such as an abiotic environment or a multicellular

17 host, are in direct contact and therefore contribute to the local gene flow and genetic

18 repertoire.

19

20 4.4 Urease cluster horizontal gene transfer events

21 The urease cluster in the PPI network of U. diversum is among the sub-networks

22 with the highest confidence scores detected in this study. Analyses have shown that urea

23 hydrolysis is responsible for generating the majority (95%) of ATP demand in

24 Ureaplasmas (Smith et al. 1993; Neyrolles et al. 1996). No urea and nickel transporters

25 were found in U. diversum genome annotation (Marques et al. 2016) and the same occurs

26 in other Ureaplasmas (Pollack 2001). Urea could be acquired directly from the

27 environment by diffusion across the cell membrane, but it is still not clear if this process

28 is responsible for urea intake in U. diversum.

29 The frequency and impact of horizontal gene transfer are high among

30 Corynebacterium, especially regarding pathogenicity and virulence factors (Oliveira et

31 al. 2017). Urease can supply the energetic demand of an organism through urea

32 hydrolysis, but it is also an important virulence factor of pathogenic bacteria. Horizontal

33 gene transfer has already been detected in Corynebacterium species occurring among

34 classic pathogenicity islands, such as adhesins, secreted toxins and proteins involved in

Page 15 of 39



Draft

16

1 iron uptake from the host (Ruiz et al. 2011). The Corynebacterium species analyzed in

2 this study were C. glucuronolyticum and C. urealyticum. The 16S rRNA trees generated

3 as a control can be visualized in Figure 5 of the Supplementary information.

4 Corynebacterium glucuronolyticum is an opportunistic pathogen known for causing

5 urogenital infections in humans under certain clinical conditions (Gherardi et al. 2015)

6 and it is also the causative agent of urologic infections and reproductive disorders among

7 swine (Devriese et al. 2000) and ovine (Takahashi et al. 1997). Corynebacterium

8 urealyticum is a common isolated species from urological infection samples (Salem et al.

9 2015), and studies suggest that co-infection by Ureaplasma urealyticum can occur in

10 these cases (Trinchieri 2014), aggravating the infection and promoting kidney stone

11 formation.

12 S. aureus is one of the most studied human pathogens and is the causative agent of

13 many infections of clinical importance (Tong et al. 2015). S. aureus, as well as

14 Corynebacterium and Ureaplasma, colonizes the urogenital tract niche. Differing from

15 Ureaplasma and Corynebacterium infections, the urogenital tract diseases caused by S.

16 aureus are recurrent in immunosuppressed hospital patients with urinary catheters, which

17 facilitate bacterial colonization and dissemination. Classic studies show that

18 Staphylococcus urease has similar biochemical properties to those present in other

19 bacterial groups and there is topological conservation in the urease operon among urease

20 producing bacteria (Gatermann and Marre 1989).

21 S. aureus infections in cattle are of economic relevance since they affect product

22 quality (Guinane et al. 2011). The fact that specific S. aureus clones are capable of

23 infecting animal hosts such as poultry and cattle, but are not isolated from human samples,

24 suggests a necessary adaptation to new niches enabled by the acquisition of crucial genes

25 in these bacterial populations (Guinane et al. 2011). Specific genes have been discovered

26 as being transferred horizontally from human-colonizing S. aureus populations and

27 regarded as facilitators in the colonization of poultry (Lowder et al. 2009) and cattle hosts

28 (Herron-Olson et al. 2007). Guinane et al. (2011) using population genetics, comparative

29 genomics and a functional analytical approach estimated that the horizontal gene transfer

30 event (or events) enabling S. aureus populations to colonize domestic animals such as

31 poultry and small and big ruminants occurred 11 thousand years ago, when domestication

32 of these animals was in its early stages (Zeder 2008). The colonizing “jump” made by

33 specific S. aureus populations (later referred to as serovars) to ruminants and poultry

34 continued in the adaptation process enabled by allelic diversification, genetic loss and

Page 16 of 39



Draft

17

1 mobile genetic elements acquisition (Guinane et al. 2011). The latter can compromise 15

2 to 20% of the S. aureus genome (Lindsay 2014). Compared to human-colonizing S.

3 aureus, the serovars infecting poultry and ruminants present an increase and

4 diversification in virulence factors, adhesins and metabolic machinery that ensure

5 survivability in these new hosts. These data suggest that during evolution, S. aureus had

6 to adapt to new hosts, which is also true for Ureaplasma serovars that exclusively

7 colonize domestic animals.

8 In U. diversum, poorly characterized hypothetical proteins occupy the urease

9 network (UUR10_0212, UUR10_0168 and purines UUR10_0577) (Figure 4), while in

10 U. urealyticum and U. parvum the same spots are occupied (including the interactions

11 with urease accessory proteins) by DNA directed RNA polymerase (rpoB), fructose

12 biphosphate aldolase (fba), acetate kinase (ackA) (Supplementary information – Figure

13 2), suggesting that the functions of the hypothetical proteins in U. diversum may be

14 similar to those performed by these proteins. When placed in the same network, they

15 continue to interact with the remaining ureaplasmal protein repertoire, indicating that the

16 biological function configured by the interactions between these proteins is similar. C.

17 urealyticum have DNA polymerase in their network (cu1143), a gmp synthase (guaA)

18 and a carboxylate dehydrogenase (cu1505), while these locations in C. glucurolyticum

19 are occupied by hypothetical proteins and gatA, a transcriptional factor. As in U.

20 diversum, it may be possible that the hypothetical proteins in C. glucurolyticum have

21 functions similar to the characterized proteins in the C. urealyticum PPI network.

22 Mycoplasmas and ureaplasmas are known for low genomic GC content compared

23 to more distantly related bacterial groups. Relying on the host for intake of metabolites

24 and precursors, these bacteria were able to eliminate genes related to biosynthetic

25 pathways mainly through genetic drift, a process facilitated by restriction to specific hosts

26 and consequent lack of recombination between different strains (McCutcheon and Moran

27 2012). The genomic GC content of different bacterial species has been linked to lineage-

28 specific mutational patterns and to selection of diverse genome-wide properties (Sueoka

29 1962; Muto and Osawa 1987; Rocha and Feil 2010). The high AT bias found in the

30 Mollicutes also characterizes other minimal and reduced genomes. As discussed by

31 McCutcheon and Moran (2012), the AT bias in small genomes can be explained by the

32 characteristic loss of DNA repair genes in these groups, leading to an accumulation of

33 AT mutations. This is the case, for example, of cytosine deamination and guanine

34 oxidation, leading to (G or C)→(A or T) changes in the genome, especially in symbiotic

Page 17 of 39



Draft

18

1 bacteria (Lind and Andersson 2008). The accumulation of deleterious substitutions can

2 increase the rate of genomic evolution at nucleotide and amino acid levels (McCutcheon

3 and Moran 2012).

4 During evolution, changes in the so-called “universal genetic code” are more common

5 than initially assumed, leading to coding divergence among organisms (Knight et al.

6 2001). Amid these changes, the reassignment of UGA from a stop codon to a tryptophan

7 codon in bacterial lineages such as the Mollicutes is one of the most common (Maniloff

8 2002). This coding property could be a significant obstacle for horizontal gene transfer

9 events in which mollicute species act as donors, since a transferred gene could lead to a

10 premature translation stop in organisms with universal codon usage. On the other hand,

11 transfer of mollicute genes to more distantly related species with higher GC content is

12 expected to leave a low GC “fingerprint” behind. In our study, this was observed in the

13 urease cluster, determined by our analysis as being transferred from Ureaplasma species

14 to S. aureus, C. urealyticum and C. glucuronolyticum. In table 3, the average GC content

15 of the transferred genes among the bacterial species and the overall mean genomic GC

16 percentage are listed. As can be seen, while S. aureus GC content does not change

17 significantly between the urease cluster and the whole genome, this is not the case for the

18 Corynebacterium species analyzed in our study (Riegel et al. 1995). Both show a decrease

19 in GC content in the urease cluster related to the genome wide percentage, reinforcing the

20 genetic transfer taking place between these groups.

21

22 4.5 Purine and Pyrimidine metabolism horizontal gene transfer events

23 Ureaplasma, as well as the majority of mycoplasmas, is unable to realize de novo

24 biosynthesis of purines and pyrimidines. To compensate for the lack of metabolic genes,

25 Ureaplasma relies on the host to import a high number of precursors and even finished

26 metabolites. In the PPI network of U. diversum, 79 proteins are involved in transport and

27 binding processes and the purine and pyrimidine pathway is incomplete (Marques et al.

28 2016). Studies have shown that incorporation of nucleotides, nucleosides and free bases

29 can occur in many prokaryotic organisms (Pandey 1984; Fraser et al. 1995). The demand

30 of purines and pyrimidines can be supplied by intracellular nucleotides in the host cells,

31 and Razin and Avigan (1963) observed that mollicute strains use the host tissues and

32 medium as nutritional source of nucleotide precursors (Razin 1985). This is most probably

33 also the case in U. diversum.

Page 18 of 39



Draft

19

1 In species such as the mammalian intestinal tract parasite Cryptosporidium parvum,

2 alternative enzymes acquired through horizontal gene transfer events assist in purine and

3 pyrimidine synthesis using precursors imported from the host (Striepen et al. 2004). U.

4 diversum has three of these alternative enzymes that assist in nucleotide salvage: uridine

5 kinase, uracil phosphoribosyltransferase and thymidine kinase. In C. parvum thymidine

6 kinase was acquired through horizontal gene transfer from a proteobacteria (Striepen et

7 al. 2004).

8 Mollicutes rely upon intake of pre-existing nucleosides and bases from the host and

9 environment, using either nucleobases or nucleosides to synthesize required nucleoside

10 triphosphates and deoxynucleoside triphosphates (Arraes et al. 2007). Purine nucleoside

11 phosphorylase is an enzyme important to both purine and pyrimidine metabolism and has

12 a nucleoside ribosyltransferase function (Mombach et al. 2006). M. genitalium and M.

13 pneumoniae do not have nucleobase or nucleoside transporters (Paulsen et al. 2000) in

14 their genomes. Other transporters with a broad variety of substrates identified in both

15 species could be active in the transport of nucleotide precursor metabolites, such as the

16 family of ATP-binding cassette (ABC) proteins and the Major Facilitator Superfamily

17 (MSF) primary active transporter. Minion et al. (1993) tested 20 mycoplasma species for

18 external membrane-associated nuclease activity, which could be critical for nucleotide

19 precursor transport across bacterial membranes and found these reactions to take place.

20 A few years later, mnuA, a membrane nuclease gene, was first cloned and isolated in M.

21 pulmonis (Jarvill-Taylor et al. 1999). Membrane nucleases can be involved in virulence,

22 but also contribute to the degradation of host RNA and DNA, which could also be a source

23 of free bases and nucleotide precursors for Mycoplasma species (Razin et al. 1998).

24 In the pyrimidine pathway, uracil phosphoribosyltransferase is responsible for the

25 phosphorylation of uracil or cytidine (Tu and Turnbough 1997; Lundegaard and Jensen

26 1999) and is present in the genome annotation of U. diversum. Uridylate kinase catalyzes

27 the phosphorylation of UMP (Schultz et al. 1997) and, since cytidylate kinase is absent in

28 the genome annotation of U. diversum, it should be the protein responsible for UMP

29 phosphorylation (Bucurenci et al. 1996). Ribonucleotide reductase is a ubiquitous enzyme

30 responsible for deoxyribonucleotide synthesis and as expected, is also present in U.

31 diversum. Reichard (2002) showed that this single protein is capable of reducing all four

32 common ribonucleotides. Thymidylate kinase is part of the nucleoside monophosphate

33 kinase super family and is responsible for the reaction of thymidine monophosphate

34 phosphorylation, using ATP as its preferred phosphoribosyl donor. As in other organisms,

Page 19 of 39



Draft

20

1 the regulation of thymidylate kinase in U. diversum is essential due to its participation in

2 the overall control of DNA synthesis adjustments in the dTTP pool.

3 In the purine metabolic enzymes, U. diversum relies on guanylate kinase in its

4 metabolic machinery. Guanylate kinase is responsible for the reversible phosphoryl

5 transfer form ATP to GMP (Li et al. 1996). The genomic annotation of U. diversum

6 showed the absence of thymidine phosphorylase, which points to thymidine as the

7 precursor imported for dTTP production or alternatively an unknown enzyme could be

8 responsible for this function (Santos et al. 2011).

9 Horizontal gene transfer has an important role in the evolution of the Mollicutes,

10 especially when contextualized with the genomic reduction process that characterizes this

11 group. A high flow of genetic information is expected, as well as a horizontal gene

12 transfer rate not limited to more distantly related species. A continuous gene flow between

13 Mycoplasma, Ureaplasma and Corynebacterium could have been facilitated by the direct

14 interaction among these groups in the urogenital tract of their hosts, since these bacteria

15 are prevalent in urogenital tract infections (Russo et al. 1981).

16

17 5. Conclusion

18 Ureaplasma and Mycoplasma are among the microorganisms with the smallest

19 known genomes resulting from a genome reduction process that led these groups

20 throughout their evolution to present natural “almost minimal” gene repertoires. This

21 reflects in their protein-protein interaction networks. While phylogenetically distant

22 bacteria containing elaborate metabolic pathways such as S. aureus and Corynebacterium

23 urealyticum have global PPI networks with higher compartmentation compromised with

24 complete metabolic pathways, our data suggests that U. diversum, other ureaplasmas and

25 mycoplasmas go in the opposite direction. Even though they present protein clusters that

26 interact at a higher frequency, these groups have proteins that connect with each other in

27 a more simple and general way, performing the role occupied by many proteins in more

28 distantly related species. Urease is a niche gene crucial for energy production during

29 urogenital tract colonization by Ureaplasma, Corynebacterium and S. aureus, and this

30 enzyme is shared between these groups via horizontal gene transfer. The process of

31 genome reduction among the Mollicutes therefore seems to have favored more loosely

32 connected PPI networks, in which higher interaction among the remaining proteins

33 compensates for the gene loss suffered by these groups throughout their evolutionary

34 history.

Page 20 of 39



Draft

21

1

2

3

4 Conflict of interest

5 - The authors declare that there is no conflict of interest.

6

7 Acknowledgements

8

9 Joana Kästle Silva received a Masters degree fellowship from Coordenação de

10 Aperfeiçoamento de Pessoal de Nível Superior (CAPES), Brazil. Jim Hesson edited the

11 manuscript (https://www.academicenglishsolutions.com).

12

13 6. References

14

15 Albalat, R., and Cañestro, C. 2016. Evolution by gene loss. Nat. Rev. Genet. 17(7): 379.

16 Nature Publishing Group.

17 Argue, B., Chousalkar, K.K., and Chenoweth, P.J. 2013. Presence of Ureaplasma

18 diversum in the Australian cattle population. Aust. Vet. J. 91(3): 99–101. Wiley

19 Online Library.

20 Arraes, F., Carvalho, M.J.A. de, Maranhão, A.Q., Brígido, M.M., Pedrosa, F.O., and

21 Felipe, M.S.S. 2007. Differential metabolism of Mycoplasma species as revealed

22 by their genomes. Genet. Mol. Biol. 30(1): 182–189. SciELO Brasil.

23 Assenov, Y., Ramírez, F., Schelhorn, S.-E., Lengauer, T., and Albrecht, M. 2007.

24 Computing topological parameters of biological networks. Bioinformatics 24(2):

25 282–284. Oxford University Press.

26 Bell, J.C., and Kowalczykowski, S.C. 2016. RecA: regulation and mechanism of a

27 molecular search engine. Trends Biochem. Sci. 41(6): 491–507. Elsevier.

28 Bizarro, C.V., and Schuck, D.C. 2007. Purine and pyrimidine nucleotide metabolism in

29 Mollicutes. Genet. Mol. Biol. 30(1): 190–201. SciELO Brasil.

30 Brendel, V., Brocchieri, L., Sandler, S.J., Clark, A.J., and Karlin, S. 1997. Evolutionary

31 comparisons of RecA-like proteins across all major kingdoms of living organisms.

32 J. Mol. Evol. 44(5): 528–541. Springer.

33 Bucurenci, N., Sakamoto, H., Briozzo, P., Palibroda, N., Serina, L., Sarfati, R.S.,

34 Labesse, G., Briand, G., Danchin, A., and Bârzu, O. 1996. CMP kinase from

Page 21 of 39



Draft

22

1 Escherichia coli is structurally related to other nucleoside monophosphate kinases.

2 J. Biol. Chem. 271(5): 2856–2862. ASBMB.

3 Cardoso, M.V., Scarcelli, E., Grasso, L.M.P.S., Teixeira, S.R., and Genovez, M.É.

4 2000. Ureaplasma diversum and reproductive disorder in Brazilian cows and

5 heifers; first report. Anim. Reprod. Sci. 63(3): 137–143. Elsevier.

6 Consortium, A.I.M. 2011. Evidence for network evolution in an Arabidopsis

7 interactome map. Science (80-. ). 333(6042): 601–607. American Association for

8 the Advancement of Science.

9 Coordinators, N.R. 2017. Database resources of the national center for biotechnology

10 information. Nucleic Acids Res. 44(Database issue): D7. Oxford University Press.

11 Cordova, C.M.M., Hoeltgebaum, D.L., Machado, L.D.P.N., and SANTOS, L.D.O.S.

12 2016. Molecular biology of mycoplasmas: from the minimum cell concept to the

13 artificial cell. An. Acad. Bras. Cienc. 88: 599–607. SciELO Brasil.

14 Coyle, M., Hu, J., and Gartner, Z. 2016. Mysteries in a Minimal Genome. ACS

15 Publications.

16 Cuff, J.A., Birney, E., Clamp, M.E., and Barton, G.J. 2000. ProtEST: protein multiple

17 sequence alignments from expressed sequence tags. Bioinformatics 16(2): 111–

18 116. Oxford University Press.

19 Culman, S.W., Bukowski, R., Gauch, H.G., Cadillo-Quiroz, H., and Buckley, D.H.

20 2009. T-REX: software for the processing and analysis of T-RFLP data. BMC

21 Bioinformatics 10(1): 171. BioMed Central.

22 Delaye, L., and Moya, A. 2010. Evolution of reduced prokaryotic genomes and the

23 minimal cell concept: variations on a theme. Bioessays 32(4): 281–287. Wiley

24 Online Library.

25 Deng, M., Zhang, K., Mehta, S., Chen, T., and Sun, F. 2003. Prediction of protein

26 function using protein–protein interaction data. J. Comput. Biol. 10(6): 947–960.

27 Mary Ann Liebert, Inc.

28 Devriese, L.A., Riegel, P., Hommez, J., Vaneechoutte, M., De Baere, T., and

29 Haesebrouck, F. 2000. Identification of Corynebacterium glucuronolyticum strains

30 from the urogenital tract of humans and pigs. J. Clin. Microbiol. 38(12): 4657–

31 4659. Am Soc Microbiol.

32 Doig, P.A., Ruhnke, H.L., MacKay, A.L., and Palmer, N.C. 1979. Bovine granular

33 vulvitis associated with ureaplasma infection. Can. Vet. J. 20(4): 89. Canadian

34 Veterinary Medical Association.

Page 22 of 39



Draft

23

1 Dordet-Frisoni, E., Sagné, E., Baranowski, E., Breton, M., Nouvel, L.X., Blanchard, A.,

2 Marenda, M.S., Tardy, F., Sirand-Pugnet, P., and Citti, C. 2014. Chromosomal

3 transfers in mycoplasmas: when minimal genomes go mobile. MBio 5(6): e01958-


5 Erten, S., Li, X., Bebek, G., Li, J., and Koyutürk, M. 2009. Phylogenetic analysis of

6 modularity in protein interaction networks. BMC Bioinformatics 10(1): 333.

7 BioMed Central.

8 Fraser, C.M., Gocayne, J.D., White, O., Adams, M.D., Clayton, R.A., Fleischmann,

9 R.D., Bult, C.J., Kerlavage, A.R., Sutton, G., and Kelley, J.M. 1995. The minimal

10 gene complement of Mycoplasma genitalium. Science (80-. ). 270(5235): 397–404.

11 American Association for the Advancement of Science.

12 García-Castillo, M., Morosini, M.-I., Gálvez, M., Baquero, F., del Campo, R., and

13 Meseguer, M.-A. 2008. Differences in biofilm development and antibiotic

14 susceptibility among clinical Ureaplasma urealyticum and Ureaplasma parvum

15 isolates. J. Antimicrob. Chemother. 62(5): 1027–1030. Oxford University Press.

16 Gatermann, S., and Marre, R. 1989. Cloning and expression of Staphylococcus

17 saprophyticus urease gene sequences in Staphylococcus carnosus and contribution

18 of the enzyme to virulence. Infect. Immun. 57(10): 2998–3002. Am Soc Microbiol.

19 Gherardi, G., Di Bonaventura, G., Pompilio, A., and Savini, V. 2015. Corynebacterium

20 glucuronolyticum causing genitourinary tract infection: Case report and review of

21 the literature. IDCases 2(2): 56–58. Elsevier.

22 Glass, J.I., Merryman, C., Wise, K.S., Hutchison, C.A., and Smith, H.O. 2017. Minimal

23 Cells—Real and Imagined. Cold Spring Harb. Perspect. Biol. 9(12): a023861. Cold

24 Spring Harbor Lab.

25 Guinane, C.M., Penadés, J.R., and Fitzgerald, J.R. 2011. The role of horizontal gene

26 transfer in Staphylococcus aureus host adaptation. Virulence 2(3): 241–243. Taylor

27 & Francis.

28 Han, Y.-C., Song, J.-M., Wang, L., Shu, C.-C., Guo, J., and Chen, L.-L. 2016.

29 Prediction and characterization of protein-protein interaction network in Bacillus

30 licheniformis WX-02. Sci. Rep. 6: 19486. Nature Publishing Group.

31 Herron-Olson, L., Fitzgerald, J.R., Musser, J.M., and Kapur, V. 2007. Molecular

32 correlates of host specialization in Staphylococcus aureus. PLoS One 2(10): e1120.

33 Public Library of Science.

34 Jain, R., Rivera, M.C., and Lake, J.A. 1999. Horizontal gene transfer among genomes:

Page 23 of 39



Draft

24

1 the complexity hypothesis. Proc. Natl. Acad. Sci. 96(7): 3801–3806. National

2 Acad Sciences.

3 Jarvill-Taylor, K.J., VanDyk, C., and Minion, F.C. 1999. Cloning of mnuA, a

4 Membrane Nuclease Gene of Mycoplasma pulmonis, and Analysis of Its

5 Expression in Escherichia coli. J. Bacteriol. 181(6): 1853–1860. Am Soc

6 Microbiol.

7 Juhas, M., Eberl, L., and Glass, J.I. 2011. Essence of life: essential genes of minimal

8 genomes. Trends Cell Biol. 21(10): 562–568. Elsevier.

9 Kanehisa, M., and Goto, S. 2000. KEGG: kyoto encyclopedia of genes and genomes.

10 Nucleic Acids Res. 28(1): 27–30. Oxford University Press.

11 Karlin, S., and Brocchieri, L. 1996. Evolutionary conservation of RecA genes in

12 relation to protein structure and function. J. Bacteriol. 178(7): 1881–1894. Am Soc

13 Microbiol.

14 Katoh, K., and Standley, D.M. 2013. MAFFT multiple sequence alignment software

15 version 7: improvements in performance and usability. Mol. Biol. Evol. 30(4):

16 772–780. Society for Molecular Biology and Evolution.

17 Knight, R.D., Freeland, S.J., and Landweber, L.F. 2001. Rewiring the keyboard:

18 evolvability of the genetic code. Nat. Rev. Genet. 2(1): 49. Nature Publishing

19 Group.

20 Kokkayil, P., and Dhawan, B. 2015. Ureaplasma: current perspectives. Indian J. Med.

21 Microbiol. 33(2): 205–14. Medknow Publications and Media Pvt. Ltd.

22 doi:10.4103/0255-0857.154850.

23 Li, Y., Zhang, Y., and Yan, H. 1996. Kinetic and thermodynamic characterizations of

24 yeast guanylate kinase. J. Biol. Chem. 271(45): 28038–28044. ASBMB.

25 Ligon, J. V, and Kenny, G.E. 1991. Virulence of ureaplasmal urease for mice. Infect.

26 Immun. 59(3): 1170–1171. Am Soc Microbiol.

27 Lind, P.A., and Andersson, D.I. 2008. Whole-genome mutational biases in bacteria.

28 Proc. Natl. Acad. Sci. 105(46): 17878–17883. National Acad Sciences.

29 Lindsay, J.A. 2014. Staphylococcus aureus genomics and the impact of horizontal gene

30 transfer. Int. J. Med. Microbiol. 304(2): 103–109. Elsevier.

31 Lowder, B. V, Guinane, C.M., Zakour, N.L. Ben, Weinert, L.A., Conway-Morris, A.,

32 Cartwright, R.A., Simpson, A.J., Rambaut, A., Nübel, U., and Fitzgerald, J.R.

33 2009. Recent human-to-poultry host jump, adaptation, and pandemic spread of

34 Staphylococcus aureus. Proc. Natl. Acad. Sci. 106(46): 19545–19550. National

Page 24 of 39



Draft

25

1 Acad Sciences.

2 Lundegaard, C., and Jensen, K.F. 1999. Kinetic mechanism of uracil

3 phosphoribosyltransferase from Escherichia coli and catalytic importance of the

4 conserved proline in the PRPP binding site. Biochemistry 38(11): 3327–3334.

5 ACS Publications.

6 Maniloff, J. 1996. The minimal cell genome:" on being the right size". Proc. Natl. Acad.

7 Sci. 93(19): 10004–10006. National Acad Sciences.

8 Maniloff, J. 2002. Phylogeny and evolution. In Molecular biology and pathogenicity of

9 mycoplasmas. Springer. pp. 31–43.

10 Marques, L.M., Rezende, I.S., Barbosa, M.S., Guimarães, A.M.S., Martins, H.B.,

11 Campos, G.B., do Nascimento, N.C., dos Santos, A.P., Amorim, A.T., and Santos,

12 V.M. 2016. Ureaplasma diversum Genome Provides New Insights about the

13 Interaction of the Surface Molecules of This Bacterium with the Host. PLoS One

14 11(9): e0161926. Public Library of Science.

15 McCutcheon, J.P., and Moran, N.A. 2012. Extreme genome reduction in symbiotic

16 bacteria. Nat. Rev. Microbiol. 10(1): 13. Nature Publishing Group.

17 Minion, F.C., Jarvill-Taylor, K.J., Billings, D.E., and Tigges, E. 1993. Membrane-

18 associated nuclease activities in mycoplasmas. J. Bacteriol. 175(24): 7842–7847.

19 Am Soc Microbiol.

20 Mombach, J.C.M., Lemke, N., da Silva, N.M., Ferreira, R.A., Isaia, E., and Barcellos,

21 C.K. 2006. Bioinformatics analysis of mycoplasma metabolism: Important

22 enzymes, metabolic similarities, and redundancy. Comput. Biol. Med. 36(5): 542–

23 552. Elsevier.

24 Mulira, G.L., Saunders, J.R., and Kariuki, D.P. 1989. Isolation of Ureaplasma diversum

25 from bovine granular vulvitis in Kenya. Bull. Anim. Heal. Prod. Africa= Bull. la

26 sante la Prod. Anim. en Afrique.

27 Muto, A., and Osawa, S. 1987. The guanine and cytosine content of genomic DNA and

28 bacterial evolution. Proc. Natl. Acad. Sci. 84(1): 166–169. National Acad Sciences.

29 Neyrolles, O., Ferris, S., Behbahani, N., Montagnier, L., and Blanchard, A. 1996.

30 Organization of Ureaplasma urealyticum urease gene cluster and expression in a

31 suppressor strain of Escherichia coli. J. Bacteriol. 178(3): 647–655. Am Soc

32 Microbiol.

33 Oliveira, A., Oliveira, L.C., Aburjaile, F., Benevides, L., Tiwari, S., Jamal, S.B., Silva,

34 A., Figueiredo, H.C.P., Ghosh, P., and Portela, R.W. 2017. Insight of Genus

Page 25 of 39



Draft

26

1 Corynebacterium: Ascertaining the Role of Pathogenic and Non-pathogenic

2 Species. Front. Microbiol. 8: 1937. Frontiers.

3 Oliveira Filho, B.D., Porto, R.N.G., Gambarini, M.L., Kunz, T.L., Ferraz, H.T., Viu,

4 M.A.O., Lopes, D.T., and Sousa, A.P.F. 2005. Isolamento do Ureaplasma

5 diversum em muco vulvovaginal de vacas leiteiras repetidoras de estro no estado

6 de Alagoas Brasil. Arch. Vet. Sci. 10(2).

7 Pandey, N.K. 1984. Transport of nucleic acid bases and nucleosides across the bacterial

8 membrane. FEMS Microbiol. Lett. 21(1): 11–14. Elsevier.

9 Paulsen, I.T., Nguyen, L., Sliwinski, M.K., Rabus, R., and Saier, M.H. 2000. Microbial

10 genome analyses: comparative transport capabilities in eighteen prokaryotes. J.

11 Mol. Biol. 301(1): 75–100. Elsevier.

12 Peeters, S.H., and de Jonge, M.I. 2017. For the greater good: Programmed cell death in

13 bacterial communities. Microbiol. Res. Elsevier.

14 Pollack, J.D. 2001. Ureaplasma urealyticum: an opportunity for combinatorial

15 genomics. Trends Microbiol. 9(4): 169–175. Elsevier.

16 Rae, D.O., Chenoweth, P.J., and Brown, M.B. 1993. Ureaplasma infection in the

17 bovine. Arch. STD HIV Res. 7: 239. reproductive health center.

18 Rain, J.-C., Selig, L., De Reuse, H., Battaglia, V., Reverdy, C., Simon, S., Lenzen, G.,

19 Petel, F., Wojcik, J., and Schächter, V. 2001. The protein–protein interaction map

20 of Helicobacter pylori. Nature 409(6817): 211–215. Nature Publishing Group.

21 Raman, K. 2010. Construction and analysis of protein–protein interaction networks.

22 Autom. Exp. 2(1): 2. BioMed Central.

23 Razin, S. 1985. Molecular biology and genetics of mycoplasmas (Mollicutes).

24 Microbiol. Rev. 49(4): 419. American Society for Microbiology (ASM).

25 Razin, S. 2006. The genus Mycoplasma and related genera (class Mollicutes). In The

26 prokaryotes. Springer. pp. 836–904.

27 Razin, S., Yogev, D., and Naot, Y. 1998. Molecular biology and pathogenicity of

28 mycoplasmas. Microbiol. Mol. Biol. Rev. 62(4): 1094–1156. Am Soc Microbiol.

29 Reichard, P. 2002. Ribonucleotide reductases: the evolution of allosteric regulation.

30 Arch. Biochem. Biophys. 397(2): 149–155. Elsevier.

31 Riegel, P., Ruimy, R., De Briel, D., Prévost, G., Jehl, F., Bimet, F., Christen, R., and 32 Monteil, H. 1995. Corynebacterium seminale sp. nov., a new species associated 33 with genital infections in male patients. J. Clin. Microbiol. 33(9): 2244–2249. Am 34 Soc Microbiol.

35 Rocha, E.P.C., and Feil, E.J. 2010. Mutational patterns cannot explain genome

Page 26 of 39



Draft

27

1 composition: are there any neutral sites in the genomes of bacteria? PLoS Genet.

2 6(9): e1001104. Public Library of Science.

3 Ruiz, J.C., D’Afonseca, V., Silva, A., Ali, A., Pinto, A.C., Santos, A.R., Rocha,

4 A.A.M.C., Lopes, D.O., Dorella, F.A., and Pacheco, L.G.C. 2011. Evidence for

5 reductive genome evolution and lateral acquisition of virulence functions in two

6 Corynebacterium pseudotuberculosis strains. PLoS One 6(4): e18551. Public

7 Library of Science.

8 Russo, J.F., Coppola, K., and Furness, G. 1981. Mycoplasma hominis, Ureaplasma

9 urealyticum, and Corynebacterium genitalium recovered from the lower genital

10 tracts of adolescent women. Int. J. Gynecol. Obstet. 19(6): 461–466. Elsevier.

11 Salem, N., Salem, L., Saber, S., Ismail, G., and Bluth, M.H. 2015. Corynebacterium 12 urealyticum: a comprehensive review of an understated organism. Infect. Drug 13 Resist. 8: 129. Dove Press.

14 Sanderson, M.W., Chenoweth, P.J., Yeary, T., and Nietfeld, J.C. 2000. Prevalence and

15 reproductive effects of Ureaplasma diversum in beef replacement heifers and the

16 relationship to blood urea nitrogen level. Theriogenology 54(3): 401–408. Elsevier.

17 Santos, A.P., Guimaraes, A.M.S., do Nascimento, N.C., SanMiguel, P.J., Martin, S.W.,

18 and Messick, J.B. 2011. Genome of Mycoplasma haemofelis, unraveling its

19 strategies for survival and persistence. Vet. Res. 42(1): 102. BioMed Central.

20 Schultz, C.P., Ylisastigui-Pons, L., Serina, L., Sakamoto, H., Mantsch, H.H., Neuhard,

21 J., Bârzu, O., and Gilles, A.-M. 1997. Structural and Catalytic Properties of CMP

22 Kinase from Bacillus subtilis: A Comparative Analysis with the Homologous

23 Enzyme from Escherichia coli. Arch. Biochem. Biophys. 340(1): 144–153.

24 Elsevier.

25 Shannon, P., Markiel, A., Ozier, O., Baliga, N.S., Wang, J.T., Ramage, D., Amin, N.,

26 Schwikowski, B., and Ideker, T. 2003. Cytoscape: a software environment for

27 integrated models of biomolecular interaction networks. Genome Res. 13(11):

28 2498–2504. Cold Spring Harbor Lab.

29 Sirand-Pugnet, P., Lartigue, C., Marenda, M., Jacob, D., Barré, A., Barbe, V.,

30 Schenowitz, C., Mangenot, S., Couloux, A., and Segurens, B. 2007. Being

31 pathogenic, plastic, and sexual while living with a nearly minimal bacterial

32 genome. PLoS Genet. 3(5): e75. Public Library of Science.

33 Smith, D.G., Russell, W.C., Ingledew, W.J., and Thirkell, D. 1993. Hydrolysis of urea

34 by Ureaplasma urealyticum generates a transmembrane potential with resultant

Page 27 of 39



Draft

28

1 ATP synthesis. J. Bacteriol. 175(11): 3253–3258. Am Soc Microbiol.

2 Striepen, B., Pruijssers, A.J.P., Huang, J., Li, C., Gubbels, M.-J., Umejiego, N.N.,

3 Hedstrom, L., and Kissinger, J.C. 2004. Gene transfer in the evolution of parasite

4 nucleotide biosynthesis. Proc. Natl. Acad. Sci. U. S. A. 101(9): 3154–3159.

5 National Acad Sciences.

6 Sueoka, N. 1962. On the genetic basis of variation and heterogeneity of DNA base

7 composition. Proc. Natl. Acad. Sci. U. S. A. 48(4): 582. National Academy of

8 Sciences.

9 Szklarczyk, D., Franceschini, A., Wyder, S., Forslund, K., Heller, D., Huerta-Cepas, J.,

10 Simonovic, M., Roth, A., Santos, A., and Tsafou, K.P. 2014. STRING v10:

11 protein–protein interaction networks, integrated over the tree of life. Nucleic Acids

12 Res. 43(D1): D447–D452. Oxford University Press.

13 Takahashi, T., Mori, Y., Kobayashi, H., Ochi, M., Kikuchi, N., and Hiramune, T. 1997.

14 Phylogenetic positions and assignment of swine and ovine corynebacterial isolates

15 based on the 16S rDNA sequence. Microbiol. Immunol. 41(9): 649–655. Center

16 For Academic Publications Japan.

17 Tamura, K., Dudley, J., Nei, M., and Kumar, S. 2007. MEGA4: molecular evolutionary

18 genetics analysis (MEGA) software version 4.0. Mol. Biol. Evol. 24(8): 1596–

19 1599. Oxford University Press.

20 Teachman, A.M., French, C.T., Yu, H., Simmons, W.L., and Dybvig, K. 2002. Gene

21 transfer in Mycoplasma pulmonis. J. Bacteriol. 184(4): 947–951. Am Soc

22 Microbiol.

23 The UniProt Consortium, 2017. UniProt: the universal protein knowledge database.

24 Nucleic Acids Res. 46(5) 2699. Oxford University Press.

25 Tong, S.Y.C., Davis, J.S., Eichenberger, E., Holland, T.L., and Fowler, V.G. 2015.

26 Staphylococcus aureus infections: epidemiology, pathophysiology, clinical

27 manifestations, and management. Clin. Microbiol. Rev. 28(3): 603–661. Am Soc

28 Microbiol.

29 Tosadori, G., Bestvina, I., Spoto, F., Laudanna, C., and Scardoni, G. 2016. Creating,

30 generating and comparing random network models with NetworkRandomizer.

31 F1000Research 5. Faculty of 1000 Ltd.

32 Trinchieri, A. 2014. Urinary calculi and infection. Urologia 81(2): 93–98.

33 Tu, A.H., and Turnbough, C.L. 1997. Regulation of upp expression in Escherichia coli

34 by UTP-sensitive selection of transcriptional start sites coupled with UTP-

Page 28 of 39



Draft

29

1 dependent reiterative transcription. J. Bacteriol. 179(21): 6665–6673. Am Soc

2 Microbiol.

3 Weisburg, W.G., Tully, J.G., Rose, D.L., Petzel, J.P., Oyaizu, H., Yang, D., Mandelco,

4 L., Sechrest, J., Lawrence, T.G., and Van Etten, J. 1989. A phylogenetic analysis

5 of the mycoplasmas: basis for their classification. J. Bacteriol. 171(12): 6455–


7 Woese, C.R., Stackebrandt, E., and Ludwig, W. 1985. What are mycoplasmas: the

8 relationship of tempo and mode in bacterial evolution. J. Mol. Evol. 21(4): 305–

9 316. Springer.

10 Wuchty, S. 2014. Controllability in protein interaction networks. Proc. Natl. Acad. Sci.

11 111(19): 7156–7160. National Acad Sciences.

12 Wuchty, S., and Uetz, P. 2014. Protein-protein Interaction Networks of E. coli and S.

13 cerevisiae are similar. Sci. Rep. 4: 7187. Nature Publishing Group.

14 Xiao, L., Paralanov, V., Glass, J.I., Duffy, L.B., Robertson, J.A., Cassell, G.H., Chen,

15 Y., and Waites, K.B. 2011. Extensive horizontal gene transfer in ureaplasmas from

16 humans questions the utility of serotyping for diagnostic purposes. J. Clin.

17 Microbiol. 49(8): 2818–2826. Am Soc Microbiol.

18 Zeder, M.A. 2008. Domestication and early agriculture in the Mediterranean Basin:

19 Origins, diffusion, and impact. Proc. Natl. Acad. Sci. 105(33): 11597–11604.

20 National Acad Sciences.

21

22

23

24

25

26 Figure Caption

27

28 Figure 1. General view of the PPI network in U. diversum. The colored modules correspond to

29 the following metabolic pathways and cellular processes: Blue – Gene expression, Purple –

30 Cellular processes, Transport proteins and Protein fate, Cellular envelope; Orange – Regulatory

31 functions, fatty acid and energy metabolism, cofactors and carriers; Yellow – Purine and

32 Pyrimidine metabolism; Pink – Intermediary metabolism and central Nitrogen metabolism

33 (urease cluster).

Page 29 of 39



Draft

30

1 Figure 2. Node degree distribution (A) Closeness centrality (B); Average clustering coefficient 2 (C) and Shortest path length (D) distribution in the U. diversum PPI network.

3

4 Figure 3. Horizontal gene transfer events. In (A) urease function subnetwork, (B) purine 5 metabolism subnetwork and (C) pyrimidine metabolism subnetwork. The red arrow indicates the 6 direction of the horizontal gene transfer.

7 Figure 4. Urease subnetworks in (A) U. diversum, (B) C. glucurolyticum, (C) C. urealyticum 8 and (D) S. aureus

9

10 Figure 5. Purine subnetworks in (A) U. diversum, (B) M. genitalium, (C) M. pneumoniae and 11 (D) M. hominis

12

13 Figure 6. PPI networks compromised with pyrimidine metabolism in (A) U. diversum, (B) M. 14 hominis, (C) M. genitalium, (D) M. pneumoniae. and (E) C. urealyticum

15

16

17

18

19

20

21

22

23

24

25

26

27

Page 30 of 39



Draft

General view of the PPI network in U. diversum. The colored modules correspond to the following metabolic pathways and cellular processes: Blue – Gene expression, Purple – Cellular processes, Transport proteins

and Protein fate, Cellular envelope; Orange – Regulatory functions, fatty acid and energy metabolism, cofactors and carriers; Yellow – Purine and Pyrimidine metabolism; Pink – Intermediary metabolism and

central Nitrogen metabolism (urease cluster).

Page 31 of 39



Draft

Node degree distribution (A) Closeness centrality (B); Average clustering coefficient (C) and Shortest path length (D) distribution in the U. diversum PPI network.

Page 32 of 39



Draft

Horizontal gene transfer events. In (A) urease function subnetwork, (B) purine metabolism subnetwork and (C) pyrimidine metabolism subnetwork. The red arrow indicates the direction of the horizontal gene transfer

Page 33 of 39



Draft

Urease subnetworks in (A) U. diversum, (B) C. glucurolyticum, (C) C. urealyticum and (D) S. aureus

Page 34 of 39



Draft

Purine subnetworks in (A) U. diversum, (B) M. genitalium, (C) M. pneumoniae and (D) M. hominis

Page 35 of 39



Draft

PPI networks compromised with pyrimidine metabolism in (A) U. diversum, (B) M. hominis, (C) M. genitalium, (D) M. pneumoniae. and (E) C. urealyticum

Page 36 of 39



Draft

Table 1. Centrality parameters of the hub proteins in the U. diversum PPI network

Proteins Node degree (k) Node Intersection Cellular process

recA 176 0.021 Genome integrity

maintenance

dnaK 172 0.019 Transcription

topA 164 0.040 Replication/Transcription

ligA 154 0.022 Replication/DNA repair

UUR10_0577 69 0.023 Replication/DNA repair

Page 37 of 39



Draft

Table 2. Topological parameters of the PPI networks in the analyzed species.

N.of nodes

N. of edges

Clust. Coeff.

CentralityAvg. n.

of neighbors

Path length

U. diversum 301 10.404 0.613 0.359 69.130 1.970M. pneumoniae FH 629 14.418 0.460 0.253 45.844 2.406M. hominis ATCC 23114 523 12.706 0.558 0.276 48.589 2.512M. genitalium G37 461 14.197 0.533 0.307 61.592 2.179U. urealyticum ATCC 33699 646 16.964 0.576 0.259 52.520 2.616U. parvumATCC 700970 614 17.062 0.564 0.253 55.577 2.484C. urealyticum DSM 7109 2.022 30.153 0.364 0.224 29.825 3.091C. glucurolyticum ATCC 51867 2.645 53.828 0.351 0.181 40.702 2.981S. aureus MSHR1132

2.510 41.222 0.367 0.190 32.846 3.105S. aureus COL 2.615 52.435 0.377 0.185 40.103 2.995S. aureus Mu50 2.730 57.290 0.340 0.204 41.971 2.918

Page 38 of 39



Draft

Table 3. the organisms used in the HGT analysis and their respective genome CG contente.

Species GC% DNA length Mean GC content %Ureaplasma urealyticum

30,489362 5.700 25-27

Ureaplasma parvum 30,279305 5.687 25Ureaplasma

diversum 35.878301 5.22628,2

Staphylococcus aureus

35.308254 5.025 32,7

Corynebacterium urealyticum

56.420846 5.085 64,2

Corynebacterium glucuronolyticum

52.500913 5.478 59

Page 39 of 39



draft - university of toronto t-space · 12 characteristic simplification of metabolic pathways...

Documents