An updated gene regulatory network reconstruction of multidrug-resistant Pseudomonas aeruginosa CCBH4851

Abstract

BACKGROUND

Healthcare-associated infections due to multidrug-resistant (MDR) bacteria such as Pseudomonas aeruginosa are significant public health issues worldwide. A system biology approach can help understand bacterial behaviour and provide novel ways to identify potential therapeutic targets and develop new drugs. Gene regulatory networks (GRN) are examples of in silico representation of interaction between regulatory genes and their targets.

OBJECTIVES

In this work, we update the MDR P. aeruginosa CCBH4851 GRN reconstruction and analyse and discuss its structural properties.

METHODS

We based this study on the gene orthology inference methodology using the reciprocal best hit method. The P. aeruginosa CCBH4851 genome and GRN, published in 2019, and the P. aeruginosa PAO1 GRN, published in 2020, were used for this update reconstruction process.

FINDINGS

Our result is a GRN with a greater number of regulatory genes, target genes, and interactions compared to the previous networks, and its structural properties are consistent with the complexity of biological networks and the biological features of P. aeruginosa.

MAIN CONCLUSIONS

Here, we present the largest and most complete version of P. aeruginosa GRN published to this date, to the best of our knowledge.

Key words:
Pseudomonas aeruginosa; gene regulatory network; multidrug resistance; system biology

Pseudomonas aeruginosa is a ubiquitous and opportunistic pathogen of which infections can affect the lower respiratory tract, skin, urinary tract, eyes, soft tissues, surgical wound, and gastrointestinal system, among others, leading to bacteraemia, endocarditis, and other complications, particularly in health care settings and in immunocompromised patients.^1,2,3 This Gram-negative bacteria is one of the most difficult to treat,⁴ due to its intrinsic resistance, acquisition of resistance through chromosomal gene mutations, and horizontally acquired resistance mechanisms to multiple drugs.³ Multidrug resistance (MDR) imposes the central difficulty in the selection of appropriate antibiotic treatment and reduces treatment options, especially in nosocomial settings such as healthcare-associated infections (HAI).^5,6

HAI is a severe public health issue related to high morbidity and mortality rates in hospitalised patients and high healthcare costs.⁷ Worldwide, P. aeruginosa is one of the most prevalent agents of HAI.⁸

In Brazil, the Brazilian Health Surveillance Agency⁹ ranked P. aeruginosa as the third most common causative agent of HAI in hospitalised patients in adult intensive care units (ICU) and the second in paediatric ICU, being nearly 40% of the reported strains resistant to carbapenems.⁹ This class of beta-lactam antibiotics has been widely administered worldwide for treating P. aeruginosa infections and other MDR Gram-negative bacterial infections.¹⁰ Indeed, a significantly higher mortality rate was observed among patients infected with MDR P. aeruginosa clones (44.6%) compared to those infected with non-MDR (24.8%).⁶

The most epidemiologically important mechanism of carbapenem resistance is the production of carbapenemases. Among MDR P. aeruginosa clinical isolates in Brazil, the most prevalent carbapenemase is the São Paulo metallo-β-lactamase (SPM-1).¹¹ This enzyme is encoded by the gene bla _SPM-1 , located on the P. aeruginosa chromosome,¹² and it confers resistance to almost all classes of beta-lactams. The first register of an MDR P. aeruginosa strain carrying the bla _SPM-1 gene found in Brazil is from 2003.¹³ Widely disseminated in distinct Brazilian geographic regions, SPM-1-producing P. aeruginosa is associated with the clone SP/ST277 and has been isolated from hospital sewage systems, rivers, and microbiota of migratory birds.^11,12 The strain P. aeruginosa CCBH4851 belongs to clone SP/ST277, and was involved in an endemic outbreak in Brazil in 2008.¹⁴ This strain is resistant to most antimicrobials of clinical importance, such as aztreonam, amikacin, gentamicin, ceftazidime, cefepime, ciprofloxacin, imipenem, meropenem, and piperacillin-tazobactam, being susceptible only to polymyxin B, and has several mechanisms of mobile genetic elements.^2,14

To better understand P. aeruginosa’s behaviour, more comprehensive knowledge of gene expression patterns predicted by analysing its gene regulatory network (GRN) is of great value. A GRN consists of a set of transcription factors (TF) that interact selectively and nonlinearly with each other and with other molecules in the cell to regulate mRNA and protein expression levels.¹⁵

Mathematical modelling and computational simulations are approaches for analysing the GRN and other complex cellular systems influenced by numerous factors. These models allow the construction of biological networks, predict its behaviour under unusual conditions, identify how a disease might develop, and intervene in such development to prohibit cells from reaching undesirable states.¹⁶ In addition, due to their lower cost and high accuracy, such approaches contribute to developing new drugs.¹⁷

The P. aeruginosa PAO1 strain had its genome sequence published in 2000, providing information regarding genome size, genetic complexity, and ecological versatility.¹⁸ It has been extensively studied since then.

Published in 2011 by Galán-Vasquez et al.,¹⁹ the first P. aeruginosa GRN was based on the PAO1 strain (PAO1-2011). Then, in 2019, Medeiros et al.² described a GRN reconstruction of CCBH4851 strain (CCBH-2019). Finally, in 2020, Galán-Vasquez et al.²⁰ published the updated GRN of P. aeruginosa with the PAO1 strain (PAO1-2020), which was much larger than the previous ones, containing new interactions. All works analysed the GRNs main structural properties and regulatory interactions.

This manuscript describes CCBH-2022, an updated GRN of the MDR P. aeruginosa based on the CCBH4851 strain, using as references both CCBH-2019 and PAO1-2020. We characterise regulators, target genes (TGs), transcription factors (TFs), auto-activation interactions, and influential genes of the network.

We analyse the main structural properties of the network, such as degree distribution, clustering coefficient, and relative abundance of network motifs. Finally, we compare the results of our analyses with those from previous GRNs.

MATERIALS AND METHODS

In this work we study the P. aeruginosa CCBH4851 strain, which is deposited at the Culture Collection of Hospital-Acquired Bacteria (CCBH) located at the Laboratório de Pesquisa em Infecção Hospitalar, Instituto Oswaldo Cruz/Fundação Oswaldo Cruz (Fiocruz) (WDCM947; 39 CGEN022/2010). The genome sequence is available in the GenBank database (Accession CP021380.2).¹⁴

CCBH-2019 and PAO1-2020 models were the bases for the reconstruction of this GRN. CCBH-2022 GRN results from the orthology analysis between the P. aeruginosa PAO1 and CCBH4851 gene sequences. CCBH-2022 model also inherits the orthologs between CCBH4851 and P. aeruginosa PA7²¹ and P. aeruginosa PA14²² strains, which were already present in CCBH-2019 GRN. The evolutionary histories of genes and species reconstruction are based critically on the accurate identification of orthologs.²³ Orthology refers to a specific relationship between homologous characters that arose by speciation at their most recent point of origin,^24,25 a common ancestor. One of the most common approaches to determining orthology in comparative genomics is the Reciprocal Best Hits (RBH), which relies on BLAST.²⁶ An RBH occurs when two genes from different genomes find themselves the best scoring match in the opposite genome.^27,28 Regulatory interactions between TFs and TGs in the PAO1 GRN were propagated to CCBH-2022 GRN if the TF and the TG formed an RBH. Medeiros et al.² designed and implemented an algorithm using the Python programming language to automate and generate a list of RBHs in a tabular format (available as Supplementary data). All the protein sequences from P. aeruginosa CCBH4851 (P1) and P. aeruginosa PAO1 (P2) were considered. BLAST+²⁹ was used to query the proteins from P1 against those from P2 (forward results) and P2 against P1 (reverse results). Each P1 query sequence was considered in turns, and its best match from P2 was identified from forwarding results (x). Likewise, each P2’s query sequence was considered from the reverse results, with its best match in P1 (x’). If x = x’, then they are RBH. Local BLASTP searches of each protein set against the other were executed, with the following cut-off parameters: identity ≥ 90%, coverage ≥ 90%, and E-value ≤ 1 e-5, showing the results in tabular format. If the search returned no hits, the gene was considered to have no ortholog within the opposite genome. Manual BLASTP was used to prevent false negatives, aligning these gene sequences with the opposite genome, considering the above parameters. If they still returned no hits but were present in either PAO1-2020 or CCBH-2019 models, the results were evaluated with a literature search to determine if they were accurate and whether they should be part of CCBH-2022.

The final GRN table is available as Supplementary data and is organised into six columns: Regulatory gene, Ortholog of the regulatory gene, Target gene, Ortholog of the target gene, Mode of regulation, and Reference. The first column lists the regulatory genes of P. aeruginosa CCBH4851, while the second column contains orthologs of regulatory genes in the reference strain (PAO1 and PA7 or PA14 from the exclusive interactions in CCBH-2019; the same applies to TG’s orthologs). The third column refers to the target gene in P. aeruginosa CCBH4851, while the fourth column lists orthologs of TGs in the reference strain. Finally, the fifth column describes the mode of regulation, and the sixth column indicates the corresponding data source.

The interactions between transcription factor proteins and the genes they regulate in an organism define a directed graph. For the computational analysis, the structure of GRN can be represented as a directed graph, formed by a set of vertices (or nodes) connected by a set of directed edges (or links). Basic network measurements are related to vertex connectivity, the occurrence of cycles, and the distances between pairs of nodes, among other possibilities.³⁰

The degree of vertices is the most elementary characterisation of node i, and k(i) is defined as its number of edges. In directed networks, there are incoming (k-in degree) edges and outgoing (k-out degree) edges.³¹ The degree distribution can follow a functional form P(k) = Ak ^-γ, called power-law distribution, where P(k) is the likelihood that a randomly chosen node from the network has k direct interactions, A is a constant that ensures that the P(k) values add up to one, and γ is the degree exponent.^{32,33,34,35,36} According to Albert,³⁷ this function indicates high diversity of node degrees, with the P(k) value decaying as a power law that is free of a characteristic scale, resulting in the absence of a typical node in the network that could be used to characterise the rest of the nodes. Most real networks with structural information available exhibit this scale-free behaviour, deviating from a Poisson distribution expected in a classical random network.^38,39

Studies have shown a scale-free structure in cellular metabolic networks,^32,40 protein interaction networks, including in cancer,^41,42 transcription regulatory networks, and GRN.^20,43,44,45 Following the literature,^{36,37,46,47,48} there are some qualitative and quantitative characteristics to ensure that a network is scale-free: the power-law distribution appears as a straight line on a log-log plot; the γ value usually is in the range 2<γ<3; and the presence of high-degree nodes, called hubs, the most highly connected nodes,⁴⁷ with most nodes clustered around them. The hubs demonstrate the absence of a uniform connectivity distribution in the network, presenting the 80-20 rule (also referred to as the Pareto principle), with small-degree nodes being the most abundant. However, the frequency of high-degree nodes decreases slowly.³⁷ Hubs are fundamental for determining therapeutic targets against an infectious agent.² Scale-free networks are heterogeneous,⁴⁹ so random node disruptions generally do not lead to a significant loss of connectivity. However, the loss of the hubs causes the breakdown of the network into isolated clusters.⁵⁰ Some studies validate these general conclusions for cellular networks.^51,52,53

In the GRN, determining the vertices with the highest k-out degrees is a method for identifying a hub,² The degree threshold is the exact number of interactions that characterise a hub, and this criterion differs from one study to another.⁵⁴ The degree threshold adopted in this work was the average number of connections of all nodes having at least two edges, resulting in a cut-off value of 16 connections.

Motifs are connectivity patterns, a small set of recurring regulation patterns from which the networks are built^55,56 that are associated with specific functions.⁵⁷ A triangle, i.e., three fully connected vertices, is the simplest type of motif.⁴⁷ These genes are a regulator, X, which regulates Y, and gene Z, which is regulated by both X and Y.⁵⁸ Triangles can be closed (three connections within the set) or open (two edges).³⁸ This 3-genes motif is the feedforward loop (FFL) and the most common in GRN, appearing in gene systems in bacteria and other organisms,^59,60 with the possibility of either activation or repression in each of the three regulatory interactions.⁶¹ The coherent type-1 FFL and the incoherent type-2 FFL occur more frequently in transcriptional networks.⁵⁸

The clustering coefficient is the probability that two genes with a common neighbour in a graph are also interconnected.¹⁹ This measure has two popular definitions: the local and global clustering coefficient. The local clustering coefficient of vertex i, C_i, is defined as C_i = 2e_i / k_i (k_i - 1), where e_i is the number of edges connecting node i with degree k_i, and k_i (k_i - 1) / 2 is the maximum number of edges in the neighbourhood of node i.³⁶ In GRNs, the local clustering coefficient C(i) is interpreted as the interaction between genes forming the regulatory groups.² The clustering coefficient of a network, C, is calculated by the average of C_i over all vertices.^19,62 Not considering the directionality of the edges, the global clustering coefficient is the ratio between the number of closed triangles and the total number of triangles (open or closed) in the network.² C(k) represents the mean clustering coefficient over the vertices with degree k.³⁶ Some biological networks tend to present high clustering coefficient values, e.g., in the protein-protein interaction network of S. cerevisiae, <C> ≈ 0.18.^30,47

The network density measure is the number of edges of the network over the maximum possible number of edges, measuring the interconnectivity between vertices, and is strongly correlated to the potential to generate gene expression heterogeneity.⁶³ The network diameter is the path length between the two most distant nodes.³⁶ The average path length is the measure that indicates the distances between pairs of vertices (the average of the shortest path length over all pairs of nodes in the network).⁴⁶

Several genes are connected in the GRN. When the nodes interact through a direct or an indirect link (intermediate connections), they are considered part of a connected component. These associations are the concept of network connectivity, and for this analysis in the present work, network interactions were considered undirected.²

Analysing the structural characteristics (connected components, hubs, and motifs) can help determine the best approach to disturb a network to promote a desired phenotype in the cell.⁶⁴

For CCBH-2022 structural analyses, the R programming language and RStudio were used.⁶⁵ Scales, dplyr, tibble, readr and igraph packages were used for data manipulation and plotting the structural analyses.^2,20,66 The igraph library was used to compute most properties described above: the in and out degrees, centrality, clustering coefficients, feed-forward loop motifs, connectivity, cycles, paths, and hierarchical levels analyses.⁶⁷

The illustrations of the GRN, the hub’s network, and the connectivity analysis were made in Cytoscape.⁶⁸ All figures are presented with higher resolution in the Supplementary data.

The codes for the structural analysis in R and for finding RBH in python, implemented by Medeiros et al.,² and the CCBH-2022 file in CSV format are available as Supplementary data in our Github repository (https://github.com/FioSysBio/CCBH2022).

RESULTS

CCBH-2022 consists of 5452 regulatory interactions among 3186 gene products, of which 218 were identified as regulatory genes and 2968 as target genes. Of the 218 regulatory proteins, 87 are TFs, 19 are sigma factors (SF), and 13 are RNAs. Of these 13 RNAs, 11 are SF as well. The tables containing their relations are presented in the Supplementary data.

Given the 6577 predicted protein-coding genes of P. aeruginosa CCBH4851, the model organism in this study, the current network represents roughly 50% of the genome, against 16.52% from CCBH-2019.

Specific regulatory genes and their interactions were kept as described in CCBH-2019, such as the ones resulting from the P. aeruginosa PA7 and P. aeruginosa PA14 orthology, and in dedicated biological databases and scientific literature, e.g., IHF (integration host factor). This bacterial DNA-bending protein, essential in gene expression regulation, is absent in the CCBH4851 genome. However, Delic-Attree et al.⁶⁹ demonstrated that P. aeruginosa contains the IHF protein composed of the products of the himA and himD genes. These genes act in combination as a TF for several TGs,² and all were listed as regulatory genes in CCBH-2019. Consequently, equivalent annotations to the previous CCBH4851 GRN were maintained.

CCBH-2022 has 5452 edges, and these interactions were classified into activation (“+”), repression (“-”), dual (“d”, when, depending on the conditions, the regulatory gene act as an activator or a repressor), and unknown (“?”), as described in biological databases and scientific literature. An illustration of CCBH-2022 is presented in Fig. 1.

Fig.1:
visualisation of CCBH-2022. Yellow circles indicate regulatory genes, light blue circles indicate target genes (TGs), black lines indicate an unknown mode of regulation, green lines indicate activation, and red lines indicate repression. Purple lines indicate a dual-mode of regulation. A: the gene regulatory networks (GRNs) large highly connected network component; B: all regulatory and TGs with no connections with A.

Regarding the structural measurements of the updated network, the summarised statistical results are presented in Table I. It contains the standard measures (the number of nodes and edges, number of autoregulatory motifs, network diameter, and average path length), the number of feed-forward motifs, and clustering coefficients. Also, Table I presents a comparison with data from PAO1-2011, CCBH-2019, PAO1-2020, and CCBH-2022.

TABLE I
Comparison of structural statistic measures between PAO1-2011, CCBH-2019, PAO1-2020, CCBH-2022

	PAO1-2011	CCBH-2019	PAO1-2020	CCBH-2022
Vertices	690	1046	3009	3186
Edges	1020	1576	5040	5452
Regulatory genes	76	138	173	218
Target genes	593	908	2709	2968
Positive regulation	779	772	3851	3829
Negative regulation	218	454	390	649
Dual regulation	11	13	10	19
Unknown regulation	12	337	789	955
Autoregulation (total)	29	72	50	91
Positive autoregulation	16	21	24	29
Negative autoregulation	13	39	15	46
Unknown autoregulation	-	12	11	17
Feed-forward loop motifs (total)a	137	208	702	968
Coherent type I feed-forward loop motifsa	82	79	226	239
Incoherent type II feed-forward loop motifsa	3	4	8	10
Density	2.12e-03	1.44e-03	6.07e-04	5.99e-04
Diameter	9	12	12	12
Average shortest path length	04.08	4.80	04.01	4.67
Global clustering coefficient	2.28e-02	3.2e-02	3.03e-03	4.42e-03
Local clustering coefficient	2.5e-01	1.92e-01	1.63e-01	1.87e-01

• a: number of feed-forward loop motifs determined using the igraph package.

Since both CCBH-2022 and PAO1-2020 contain significant updates from their previous counterparts, the comparison between CCBH-2022 and PAO1-2020 is most relevant. CCBH-2022 had a density of 5.99e-04, slightly lower than the density of PAO1-2020 (6.07e-04) but showed the same order of magnitude. The diameter was 12, the same as CCBH-2019 and PAO1-2020 and higher than PAO1-2011, which was 9. The average shortest path distance was 4.67, higher than PAO1-2020 (4.01) but slightly lower than CCBH-2019 (4.80). Similar to the previous GRN, CCBH-2022 was disconnected, showing one large connected component (3102 genes) and 20 small connected components.

The degree distributions of the four networks can be seen in Fig. 2A-D, with A and B being the incoming and C and D the outcoming degree distribution. Fig. 2B, D is on a log-log axis, and the straight line is consistent with a power-law distribution. For k-in, the estimated value for γ was 2.79, within the range 2<γ<3, consistent with a power law distribution. For PAO1-2020, the corresponding value was 2.67, 2.89 for CCBH-2019 and 2.71 for PAO1-2011.

Fig. 2:
graphical representation of structural measurements of CCBH-2022 (red) compared to the previously published networks: PAO1-2011 (purple), CCBH-2019 (orange), and PAO1-2020 (green). (A-B) incoming degree distribution of the four gene regulatory networks (GRNs); (C-D) outgoing distribution of the four GRNs. The distributions are plotted on a linear (A, C) and on a logarithmic scale (B, D); (E) local clustering coefficient distribution; (F) clustering coefficient by degree.

The distribution of local clustering coefficients can be seen in Fig. 2E. CCBH-2022 had a global clustering coefficient equal to 4.42e-03, higher than PAO1-2020 (3.03e-03). The scatter plot in Fig. 2F shows the correlation between the local clustering coefficient C(i) and the degree k(i).

The most frequent mode of regulation in CCBH-2022 is activation, occurring in 70,2% of the total interactions in the network, followed by roughly 12% of repression mode and 17.8% of dual or unknown mode. Autoregulation occurs when a gene regulates its expression, and the prevalence in CCBH-2022 is of negative autoregulatory motifs.

The most abundant motif in all four networks was the coherent type I FFL, with 239 in CCBH-2022 (PAO1-2011: 82; CCBH-2019: 79; PAO1-2020: 226). In addition, there were 10 incoherent type II FFL motifs in CCBH-2022 (PAO1-2011: 3; CCBH-2019: 4; PAO1-2020: 8).

Table II shows the 30 most influential hubs in CCBH-2022 and PAO1-2020.

An analysis was performed to determine whether the hubs are interconnected through direct interactions (Fig. 3).

TABLE II
The 30 most influential hubs of CCBH-2022 and PAO1-2020

CCBH-2020			PAO1-2020
Gene	Total number of connections (k-out)	Function	Gene	Total number of connections (k-out)
rpoD	740	Control of expression of housekeeping genes70	rpoD	749
rpoN	650	Nitrogen metabolism, adhesion, quorum sensing (QS), biofilm formation71	rpoN	658
algU	353	Positive regulation of response to oxidative stress72	algU	357
sigX	298	Positive regulation of cell growth73	sigX	319
rpoS	278	QS, Biofilm, virulence, antibiotic resistance74	fliA	281
fliA	270	Adhesion, flagellin biosynthesis75	rpoS	271
rpoH	184	Heat-shock response76	rpoH	194
gacA	121	Monolayer and biofilm formation77	gacA	128
algR	119	Cell motility, biofilm formation78	algR	122
amrZ	109	Cell motility, biofilm formation79	amrZ	115
lasR	106	QS, regulation of elastin catabolic process80	lasR	95
fleQ	92	Regulation of mucin adhesion and flagellar expression81	pvdS	91
fur	88	Control of expression of siderophores and exotoxin A82	sphR	90
pvdS	87	Iron metabolism, pyoverdine, virulence83,84	fleQ	85
sphR	74	Sphingosine catabolic process85	fur	69
mvfR	65	QS, regulation of lyase activity, control production of virulence factors86	vqsM	65
vqsM	61	QS, control production of virulence factors87	mvfR	62
anr	58	Regulation of oxidoreductase activity88	pchR	57
rhlR	56	QS, regulation of lipid biosynthetic and proteolysis89,90	anr	53
mexT	53	Antibiotic efflux pump91	mexT	51
pchR	47	Regulation of pyochelin siderophore, ferripyochelin receptor synthesis92	argR	46
argR	46	Controls arginine uptake and metabolism93	fecI	44
gbdR	44	Regulation of cellular amino acid metabolic process94	gbdR	42
pmrA	43	Antibiotic efflux pump95,96	rhlR	40
fecI	41	Regulation of iron ion transport97	phoB	40
soxR	40	Antibiotic efflux pump98	pmrA	40
phoB	40	Cell motility, regulation of cellular response to phosphate starvation99,100	soxR	39
vfr	37	QS, exotoxin A regulator, cell motility101	dnr	34
dnr	34	Regulation of nitrogen compound metabolic process102	himA	30
rsaL	34	QS, biofilm formation, regulation of virulence factors103	himD	30

Fig.3:
connectivity relationships among the 30 most influential hubs of CCBH-2022. Yellow circles indicate regulatory genes considered hubs, light blue circles indicate target genes, black lines indicate an unknown mode of regulation, green lines indicate activation, and red lines indicate repression. Purple lines indicate a dual-mode of regulation.

DISCUSSION

The reconstruction and analysis of the P. aeruginosa GRN contribute to a better understanding of its antibiotic resistance mechanisms. It also contributes to a greater knowledge of related cellular behaviours, such as adaptation and pathogenicity, mainly based on an MDR strain such as CCBH4851.

In this work, we have good coverage of roughly 50% of the genome on this updated network. The genome of reference strain PAO1 has 6.2Mbp, and PAO1-2020 has a coverage of 50% as well, with 5040 interactions and 3006 genes.²⁰ However, considering that the CCBH4851 genome has 6.8Mbp and has 5452 edges, and 3186 nodes, we can affirm that, to the best of our knowledge, this study presents the largest GRN of P. aeruginosa that has been assembled to date.

On the structural aspects, the charts in Fig. 2 and data in Table I make clear that CCBH-2022 represents a substantial improvement in terms of network completeness and complexity when compared with the previous P. aeruginosa GRNs since it includes more nodes, edges, and network motifs, and when comparing clustering coefficients (Fig. 2E-F). For the in silico approach, the network structural analysis is essential to understand the network architecture and performance.

The structural differences between CCBH-2022 and PA01-2020 results from additional information available due to the new version of PAO1 and recent experimental work on characterising the complete closed genome of P. aeruginosa CCBH4851.¹⁰⁴

The structural measures of CCBH-2022, such as node degree distribution and clustering coefficient, are consistent with a qualitative description of a scale-free network type. Indeed, the degree distribution followed the power-law distribution (Fig. 2B, D): a small number of nodes had many connections (the hubs) and many nodes had few connections.

The local clustering coefficient and node degree correlation (Fig. 2F) showed that nodes with lower degrees had greater local clustering coefficients than nodes with higher degrees. These characteristics are representative of several biological processes, e.g., RNA binding.^105,106

CCBH-2022 showed a lower density value than PAO1-2020. The density of both GRNs was low due to the dynamic and structural flexibility of the networks, a characteristic typical of natural phenomena-based networks,¹⁰⁷ and because the nodes were not all interconnected.² However, CCBH-2022 density was lower probably because it has 20 small connected components disconnected from the larger one (Fig. 1), while PAO1-2020 had 12 separated components. The variation in the number of connected components is plausible due to their size difference and the biological information about interactions available for the reconstruction.

All the previous P. aeruginosa GRNs are disconnected graphs, showing one large connected component and a separated few small connected components, and there may be several reasons for this disconnection in specific points. According to Medeiros et al.,² interactions among all genes are not expected since some genes in an organism are independent of each other, compartmentalised or global, constitutive or growth phase-dependent, and are triggered in different growth phases, thus resulting in a disconnected network, which corroborates with the observed low density. The reason can also be from loss of existing interactions or a gain of interactions still not fully described from additional strain-specific blocks of genes acquired by horizontal gene transfer.¹⁰⁸ The large number of connected components found in CCBH-2022 results from connectivity parameters and the global clustering coefficient. Both structural measures are affected by the same biological behaviours.¹⁰⁷

CCBH-2019 presented more negative regulations than PAO1-2011, a trend that continued between CCBH-2022 and PAO1-2020. Also, the most frequent regulatory activity in CCBH-2022 is activation, but ~50% of the autoregulation was negative, which may be a consequence of the increase in negative autoregulation in the overall network interactions compared to the previous ones. Negative auto-regulation in biological systems is commonly observed.¹⁰⁹ The Escherichia coli GRN exhibited the same pattern, with negative autoregulation prevailing concurrently with the positive regulation in the overall network.¹¹⁰ The continuity of biological processes is ensured by positive autoregulation.¹¹¹ For example, quorum sensing, biofilm formation, secretion of toxins, virulence, and resistance factors production, once initiated, must reach a final stage to have the expected effect.² In CCBH-2022, genes involved in these processes, such as lasR,⁸⁰ rhlR,⁹⁰pvdS,⁸³algU,⁷²dnr¹⁰² and anr,⁸⁸ have positive autoregulation (and are amongst the 30 principal hubs).

Negative cycles are also crucial for life-sustaining cyclic processes such as metabolic processes¹¹² and cellular homeostasis.¹¹³ In CCBH-2022, genes involved in arginine metabolism (iscR, desT, lexA, hutC, and mvat)¹¹⁰ showed a predominance of negative mode of autoregulation. Negative autoregulation is associated with cellular stability.¹¹⁴ It rapidly responds to variations in concentrations of proteins, toxins, and (or) metabolites to avoid undesired effects such as the energy cost of unneeded synthesis.¹¹⁵ In CCBH-2022, algZ (transcriptional activator of AlgD, involved in alginate production),¹¹⁶lexA (involved in the SOS response),¹¹⁷metR (involved in swarming motility and methionine synthesis),^118,119ptxR (affects exotoxin A production)¹²⁰ and rsaL (quorum-sensing repressor)¹²¹ presented negative autoregulatory interactions. Autoregulation is common among genes positioned upstream in GRN with crucial developmental functions.^122,123

The FFL motifs are essential for the modulation of cellular processes according to environmental conditions.¹²⁴ CCBH-2022 has 968 FFL motifs, which are patterns of structural structures, while PAO1-2020 has 702. There are 239 coherent type I FFL motifs in CCBH-2022, an abundant presence. According to Mangan and Alon⁶¹ these motifs act as sign-sensitive delay elements, i.e., a circuit that responds rapidly to step-like stimuli in one direction (ON to OFF) and as a delay to steps in the opposite direction (OFF to ON); the temporary removal of the stimulus ceases transcription, so the activation of expression requires a persistent signal to carry on. The incoherent type II FFL motif was less represented but also found in all the GRNs, with a total of 10 in CCBH-2022. Contrastingly with the coherent FFL, the type II FFL acts as a sign-sensitive accelerator, i.e., a circuit that responds rapidly to step-like stimuli in one direction but not in the other.⁶¹

One last characteristic revealed by the structural analysis was the presence of hubs. The hub’s network (Fig. 3) shows the connection among their interactions; they are all interconnected and belong to the largest connected component of the GRN (Fig. 1A). This connectivity reflects the importance of the influential genes. The hubs can be considered the basis of the GRN. They are crucial in searching for potential drug targets for developing new drugs, as in direct interaction with their specific targets or for an indirect interaction with the subsequent process regulation triggered by them. CCBH-2022 hubs are mainly associated with efflux pump mechanisms (mexT, pmrA, soxR),^91,95,98 alginate biosynthesis (algU, algR, rpoN),¹²⁵ and biofilm formation (rpoN, rpoS, gacA, amrZ).¹²⁶

Table II shows the 30 hubs of PAO1-2020. They are very similar to CCBH-2022 hubs, with some changes in the k-out connections. However, two CCBH-2022 hubs were not in the 30 most influential hubs of PAO1-2020: vfr, a global virulence factor regulator¹²⁷ that directly regulates 37 genes, and rsaL, associated with bacterial tolerance to antibiotics, including ciprofloxacin and carbenicillin¹²⁸ which directly regulates 34 genes. In PAO1-2020, vfr directly regulates only 12 target genes, while rsaL regulates 19, being one of these an exclusive PAO1-2020 interaction. There are 25 exclusive vfr interactions in CCBH-2022 and 16 exclusive rsaL interactions compared with PAO1-2020. These distinctions can be explained by the fact that the P. aeruginosa CCBH4851 strain is more virulent and multidrug-resistant and also because CCBH-2022 is 7.6% larger in the number of regulatory interactions (5452) than PAO1-2020 (5040), and 20,7% larger in the number of regulatory genes (218) than PAO1-2020 (173). The tables containing these exclusive interactions are in the Supplementary data. These facts strongly indicate that the operation of the main network hubs is not identical. The functioning of CCBH4851 is different, probably due to the greater influence of these two critical genes associated with multi-drug resistance and antibiotic tolerance mechanisms.

The Vfr gene’s role in regulating virulence factor production is related to the production of exotoxin A, a toxin that modifies specific target proteins within mammalian cells and induces necrosis in different tissues and organs in MDR P. aeruginosa infections.^129,130 The Vfr gene also regulates the las and consequently the rhl quorum-sensing system, two systems that together control the expression of several genes associated with virulence factor production,¹³¹ including alkaline protease, exotoxin A, pyocyanin, and rhamnolipid, as well as critical genes such as rpoS (the 5th most influential gene in CCBH-2022).¹³² The signal receptor (R gene) is one of the essential components of the las and rhl QS systems. It is necessary for coding the transcriptional activator protein (R protein).¹³³ The lasR and rhlR genes are among the 20 principal hubs. In interactions present only in CCBH-2022, vfr regulates genes associated with virulence and alginate production.¹³⁴

The rsaL gene has an important role in P. aeruginosa as a global regulator of quorum sensing, virulence, and biofilm formation.^103,135 Fan et al.¹²⁸ showed that the mutation of rsaL increased bacterial tolerance to ciprofloxacin and carbenicillin. In interactions present only in CCBH-2022, rsaL regulates mostly genes of the phz1 and phz2 clusters, showing the control that rsaL has on phenazine expression in P. aeruginosa, driving the production of phenazine-1-carboxylic acid (PCA) which is further converted in the virulence factor pyocyanin.^136,137 The pyocyanin production contributes to bacterial tolerance to ciprofloxacin and carbenicillin.¹²⁸

Pseudomonas aeruginosa evades antimicrobial activity during treatment and exerts antimicrobial resistance by mainly intrinsic resistance mechanisms. Examples of resistance mechanisms are multi-drug efflux pumps, biofilm synthesis, enzymatic inactivation/degradation, drug permeability restriction, production of beta-lactamases, acquired resistance by a mutation in drug targets, and acquisition of resistance genes via horizontal gene transfer.¹³⁸

There is a directed regulatory connection from alginate biosynthesis to iron metabolism and some antibiotic resistance mechanisms.¹³⁹ The algU, algR, rpoN, pvdS, and fecI genes are related to these processes^140,141 and are among the most influential hubs.

Pseudomonas aeruginosa has multiple efflux pump systems that prevent the antimicrobial agents from accumulating in adequate concentration to cause an effect in the cell, extruding the drug out.¹³⁸ Efflux pump systems are associated with resistance to beta-lactams, fluoroquinolones, tetracycline, chloramphenicol, macrolides, and aminoglycosides.¹⁴² Differential expression or mutations of efflux system genes are also contributing factors for carbapenem and aminoglycoside resistance.¹⁴³ The mexT, pmrA, soxR genes, related to multidrug antibiotic efflux pumps, are also amongst the most influential hubs.

The fleQ gene is also among the hubs and affects psl (polysaccharide synthesis locus) genes and regulates the efflux pump genes, mexA, mexE, and oprH, by brlR.^2,144 The psl cluster comprises 15 exopolysaccharide biosynthesis-related genes organised in tandem that are important for biofilm formation.¹⁴⁵

The mexT and soxR genes positively regulate an efflux pump system, and several virulence factors,^146,147 and pmrA regulate efflux pumps and the polymyxin B and colistin resistance.^95,148,149

Efflux pumps also help biofilm formation.¹⁵⁰ Biofilms are also related to protection from the host immune system and antibiotic penetration and tolerance, preventing them from entering the microbial population and inhibiting its action as a first-line defense mechanism.^123,151,152 The rpoN, rpoS, gacA, algR and amrZ hubs participate in the regulation of P. aeruginosa biofilm.

This system biology approach to characterise the MDR P. aeruginosa CCBH4851 regulatory network may lead to the development of strategies to disrupt the connectivity of these essential processes, thus, possibly decreasing the pathogenicity and suppressing the resistance of this bacterium.

In conclusion - This manuscript reports the reconstruction and structural analysis of the largest P. aeruginosa regulatory network available in the literature to date. This work can give new insights into identifying novel candidate antibiotic targets and contributes to an increase in our understanding of the behaviour of this bacterium.

This network’s dynamic model construction is one of our future studies, intending to help researchers working on experimental drug design and screening. The goal is to predict the dynamic behaviour better and improve the understanding of P. aeruginosa, allowing the simulation of normal and stress conditions to discover potential therapeutic targets and help develop new drugs against P. aeruginosa bacterial infection.

REFERENCES

• ¹
  Montero MM, López Montesinos I, Knobel H, Molas E, Sorlí L, Siverio-Parés A, et al. Risk factors for mortality among patients with Pseudomonas aeruginosa bloodstream infections: what is the influence of XDR phenotype on outcomes? J Clin Med. 2020; 9(2): 514.
• ²
  Medeiros Filho F, do Nascimento APB, dos Santos MT, Carvalho-Assef APD, da Silva FAB. Gene regulatory network inference and analysis of multidrug-resistant Pseudomonas aeruginosa. Mem Inst Oswaldo Cruz. 2019; 114: e190105.
• ³
  Horcajada JP, Montero M, Oliver A, Sorlí L, Luque S, Gómez-Zorrilla S, et al. Epidemiology and treatment of multidrug-resistant and extensively drug-resistant Pseudomonas aeruginosa infections. Clin Microbiol Rev. 2019; 32(4): e00031-19.
• ⁴
  Kadri SS, Adjemian J, Lai YL, Spaulding AB, Ricotta E, Prevots DR, et al. Difficult-to-treat resistance in gram-negative bacteremia at 173 US hospitals: retrospective cohort analysis of prevalence, predictors, and outcome of resistance to all first-line agents. Clin Infect Dis. 2018; 67(12): 1803-14.
• ⁵
  Tamma PD, Aitken SL, Bonomo RA, Mathers AJ, van Duin D, Clancy CJ. Infectious diseases society of America guidance on the treatment of extended-spectrum ß-lactamase producing Enterobacterales (ESBL-E), Carbapenem-Resistant Enterobacterales (CRE), and Pseudomonas aeruginosa with Difficult-to-Treat Resistance (DTR-P. aeruginosa). Clin Infect Dis. 2021; 72(7): e169-83.
• ⁶
  Matos ECO, Andriolo RB, Rodrigues YC, Lima PDL, Carneiro ICDRS, Lima KVB. Mortality in patients with multidrug-resistant Pseudomonas aeruginosa infections: a meta-analysis. Rev Soc Bras Med Trop. 2018; 51(4): 415-20.
• ⁷
  Haque M, Sartelli M, McKimm J, Bakar MA. Health care-associated infections - an overview. Infect Drug Resist. 2018; 11: 2321-33.
• ⁸
  Litwin A, Rojek S, Gozdzik W, Duszynska W. Pseudomonas aeruginosa device associated - healthcare associated infections and its multidrug resistance at intensive care unit of University Hospital: polish, 8.5-year, prospective, single-centre study. BMC Infect Dis. 2021; 21(1): 180.
• ⁹
  ANVISA - Agência Nacional de Vigilância Sanitária. Boletim segurança do paciente e qualidade em serviços de saúde nº23. Avaliação dos indicadores nacionais das infecções relacionadas à assistência à saúde (IRAS) e resistência microbiana do ano de 2020 [Internet]. Brasília, DF: Anvisa; 2020 [cited 2022 April 24]. Available from: https://app.powerbi.com/view?r=eyJrIjoiZGI3NzEwMWYtMDI1Yy00ZDE1LWI0YzItY2NiNDdmODZjZDgzIiwidCI6ImI2N2FmMjNmLWMzZjMtNGQzNS04MGM3LWI3MDg1ZjVlZGQ4MSJ9&pageName=ReportSectionac5c0437dbe709793b4b
  » https://app.powerbi.com/view?r=eyJrIjoiZGI3NzEwMWYtMDI1Yy00ZDE1LWI0YzItY2NiNDdmODZjZDgzIiwidCI6ImI2N2FmMjNmLWMzZjMtNGQzNS04MGM3LWI3MDg1ZjVlZGQ4MSJ9&pageName=ReportSectionac5c0437dbe709793b4b
• ¹⁰
  Yang P, Chen Y, Jiang S, Shen P, Lu X, Xiao Y. Association between antibiotic consumption and the rate of carbapenem-resistant Gram-negative bacteria from China based on 153 tertiary hospitals data in 2014. Antimicrob Resist Infect Control. 2018; 7: 137.
• ¹¹
  Martins WMBS, Narciso AC, Cayô R, Santos SV, Fehlberg LCC, Ramos PL, et al. SPM-1-producing Pseudomonas aeruginosa ST277 clone recovered from microbiota of migratory birds. Diagn Microbiol Infect Dis. 2018; 90(3): 221-7.
• ¹²
  Nascimento AP, Ortiz MF, Martins WM, Morais GL, Fehlberg LC, Almeida LG, et al. Intraclonal genome stability of the metallo-ß-lactamase SPM-1-producing Pseudomonas aeruginosa ST277, an endemic clone disseminated in Brazilian hospitals. Front Microbiol. 2016; 7: 1946.
• ¹³
  Toleman MA, Simm AM, Murphy TA, Gales AC, Biedenbach DJ, Jones RN, et al. Molecular characterization of SPM-1, a novel metallo-beta-lactamase isolated in Latin America: report from the SENTRY antimicrobial surveillance programme. J Antimicrob Chemother. 2002; 50(5): 673-9.
• ¹⁴
  Silveira M, Albano R, Asensi M, Carvalho-Assef AP. The draft genome sequence of multidrug-resistant Pseudomonas aeruginosa strain CCBH4851, a nosocomial isolate belonging to clone SP (ST277) that is prevalent in Brazil. Mem Inst Oswaldo Cruz. 2014; 109(8): 1086-7.
• ¹⁵
  Kitano H. Computational systems biology. Nature. 2002; 420(6912): 206-10.
• ¹⁶
  Bryce D, Kim S. Planning for gene regulatory network intervention. ACM Trans Intelli Syst Technol. 2010; 1(2): 11.
• ¹⁷
  Ryall KA, Tan AC. Systems biology approaches for advancing the discovery of effective drug combinations. J Cheminform. 2015; 7: 7.
• ¹⁸
  Stover CK, Pham XQ, Erwin AL, Mizoguchi SD, Warrener P, Hickey MJ, et al. Complete genome sequence of Pseudomonas aeruginosa PAO1, an opportunistic pathogen. Nature. 2000; 406(6799): 959-64.
• ¹⁹
  Galán-Vásquez E, Luna B, Martínez-Antonio A. The regulatory network of Pseudomonas aeruginosa. Microb Inform Exp. 2011; 1(1): 3.
• ²⁰
  Galán-Vásquez E, Luna-Olivera BC, Ramírez-Ibáñez M, Martínez-Antonio A. RegulomePA: a database of transcriptional regulatory interactions in Pseudomonas aeruginosa PAO1. Database (Oxford). 2020; 2020: baaa106.
• ²¹
  Roy PH, Tetu SG, Larouche A, Elbourne L, Tremblay S, Ren Q, et al. Complete genome sequence of the multiresistant taxonomic outlier Pseudomonas aeruginosa PA7. PLoS One. 2010; 5(1): e8842.
• ²²
  Lee DG, Urbach JM, Wu G, Liberati NT, Feinbaum RL, Miyata S, et al. Genomic analysis reveals that Pseudomonas aeruginosa virulence is combinatorial. Genome Biol. 2006; 7(10): R90.
• ²³
  Tekaia F, Yeramian E. SuperPartitions: detection and classification of orthologs. Gene. 2012; 492(1): 199-211.
• ²⁴
  Fitch WM. Distinguishing homologous from analogous proteins. Syst Zool. 1970; 19(2): 99-113.
• ²⁵
  Fitch WM. Homology a personal view on some of the problems. Trends Genet. 2000; 16(5): 227-31.
• ²⁶
  Hernández-Salmerón JE, Moreno-Hagelsieb G. Progress in quickly finding orthologs as reciprocal best hits: comparing blast, last, diamond and MMseqs2. BMC Genomics. 2020; 21(1): 741.
• ²⁷
  Overbeek R, Fonstein M, D'Souza M, Pusch GD, Maltsev N. The use of gene clusters to infer functional coupling. Proc Natl Acad Sci USA. 1999; 96(6): 2896-901.
• ²⁸
  Ward N, Moreno-Hagelsieb G. Quickly finding orthologs as reciprocal best hits with BLAT, LAST, and UBLAST: how much do we miss? PLoS One. 2014; 9(7): e101850.
• ²⁹
  Camacho C, Coulouris G, Avagyan V, Ma N, Papadopoulos J, Bealer K, et al. BLAST+: architecture and applications. BMC Bioinformatics. 2009; 10: 421.
• ³⁰
  Costa LF, Rodrigues FA, Travieso G, Boas PRV. Characterization of complex networks: a survey of measurements. Adv Phys. 2007; 56(1): 167-242.
• ³¹
  Bazanella AS, Gevers M, Hendrickx JM, Parraga A. Identifiability of dynamical networks: which nodes need be measured? In 2017 IEEE 56th Annual Conference on Decision and Control (CDC). Melbourne; 2017.
• ³²
  Jeong H, Tombor B, Albert R, Oltvai ZN, Barabási AL. The large-scale organization of metabolic networks. Nature. 2000; 407(6804): 651-4.
• ³³
  Dorogovtsev SN, Mendes JF. Scaling properties of scale-free evolving networks: continuous approach. Phys Rev E Stat Nonlin Soft Matter Phys. 2001; 63(5 Pt 2): 056125.
• ³⁴
  Strogatz SH. Exploring complex networks. Nature. 2001; 410(6825): 268-76.
• ³⁵
  Girvan M, Newman ME. Community structure in social and biological networks. Proc Natl Acad Sci USA. 2002; 99(12): 7821-6.
• ³⁶
  Almaas E, Barabási AL. Power laws in biological networks. In Koonin EV, Wolf YI, Karev GP, editors. Power laws, scale-free networks and genome biology. 2006; 1-9.
• ³⁷
  Albert R. Scale-free networks in cell biology. J Cell Sci. 2005; 118(Pt 21): 4947-57.
• ³⁸
  Barabási AL, Oltvai ZN. Network biology: understanding the cell's functional organization. Nat Rev Genet. 2004; 5(2): 101-13.
• ³⁹
  Pastor-Satorras R, Vespignani A. Evolution and structure of the Internet: a statistical physics approach. Cambridge: Cambridge University Press. 2004; 57 pp.
• ⁴⁰
  Rajula HSR, Mauri M, Fanos V. Scale-free networks in metabolomics. Bioinformation. 2018; 14(3): 140-4.
• ⁴¹
  Nacher JC, Hayashida M, Akutsu T. Emergence of scale-free distribution in protein-protein interaction networks based on random selection of interacting domain pairs. Biosystems. 2009; 95(2): 155-9.
• ⁴²
  Chen AX, Zopf CJ, Mettetal J, Shyu WC, Bolen J, Chakravarty A, et al. Scale-free structure of cancer networks and their vulnerability to hub-directed combination therapy. 2020. doi: 10.1101/2020.07.01.159657.
• ⁴³
  Wolf IR, Simões RP, Valente GT. Three topological features of regulatory networks control life-essential and specialized subsystems. Sci Rep. 2021; 11(1): 24209.
• ⁴⁴
  Farkas IJ, Jeong H, Vicsek T, Barabási AL, Oltvai ZN. The topology of the transcription regulatory network in the yeast, S. cerevisiae. Phys A: Stat Mech Appl. 2003; 318(3-4): 601-12.
• ⁴⁵
  Abdelzaher AF, Al-Musawi AF, Ghosh P, Mayo ML, Perkins EJ. Transcriptional network growing models using motif-based preferential attachment. Front Bioeng Biotechnol. 2015; 3: 157.
• ⁴⁶
  Arita M. The metabolic world of Escherichia coli is not small. Proc Natl Acad Sci USA. 2004; 101(6): 1543-7.
• ⁴⁷
  Costa LF, Rodrigues FA, Cristino AS. Complex networks: the key to systems biology. Genet Mol Biol. 2008; 31(3): 591-601.
• ⁴⁸
  Barabási AL. Linked: how everything is connected to everything else and what it means for business, science, and everyday life. 2014. 105 pp.
• ⁴⁹
  Liu Z, Lai YC, Ye N. Propagation and immunization of infection on general networks with both homogeneous and heterogeneous components. Phys Rev E Stat Nonlin Soft Matter Phys. 2003; 67(3 Pt 1): 031911.
• ⁵⁰
  Albert R, Barabasi AL. Statistical mechanics of complex networks. Rev Mod Phys. 2002; 74(1): 47-97.
• ⁵¹
  Jeong H, Mason SP, Barabási AL, Oltvai ZN. Lethality and centrality in protein networks. Nature. 2001; 411: 41-2.
• ⁵²
  Said MR, Begley TJ, Oppenheim AV, Lauffenburger DA, Samson LD. Global network analysis of phenotypic effects: protein networks and toxicity modulation in Saccharomyces cerevisiae. Proc Natl Acad Sci USA. 2004; 101(52): 18006-11.
• ⁵³
  Vogelstein B, Lane D, Levine AJ. Surfing the p53 network. Nature. 2000; 408(6810): 307-10.
• ⁵⁴
  Vandereyken K, Van Leene J, De Coninck B, Cammue BPA. Hub protein controversy: taking a closer look at plant stress response hubs. Front Plant Sci. 2018; 9: 694.
• ⁵⁵
  Milo R, Shen-Orr S, Itzkovitz S, Kashtan N, Chklovskii D, Alon U. Network motifs: simple building blocks of complex networks. Science. 2002; 298(5594): 824-7.
• ⁵⁶
  Shen-Orr SS, Milo R, Mangan S, Alon U. Network motifs in the transcriptional regulation network of Escherichia coli. Nat Genet. 2002; 31(1): 64-8.
• ⁵⁷
  Wuchty S, Oltvai ZN, Barabási AL. Evolutionary conservation of motif constituents in the yeast protein interaction network. Nat Genet. 2003; 35(2): 176-9.
• ⁵⁸
  Alon U. Network motifs: theory and experimental approaches. Nat Rev Genet. 2007; 8(6): 450-61.
• ⁵⁹
  Ma HW, Kumar B, Ditges U, Gunzer F, Buer J, Zeng AP. An extended transcriptional regulatory network of Escherichia coli and analysis of its hierarchical structure and network motifs. Nucleic Acids Res. 2004; 32(22): 6643-9.
• ⁶⁰
  Boyer LA, Lee TI, Cole MF, Johnstone SE, Levine SS, Zucker JP, et al. Core transcriptional regulatory circuitry in human embryonic stem cells. Cell. 2005; 122(6): 947-56.
• ⁶¹
  Mangan S, Alon U. Structure and function of the feed-forward loop network motif. Proc Natl Acad Sci USA. 2003; 100(21): 11980-5.
• ⁶²
  Prokhorenkova LO, Samosvat E. Global clustering coefficient in scale-free networks. In Bonato A, Graham F, Pralat P, editors. Algorithms and Models for the Web Graph. WAW 2014. Lecture notes in computer science. Vol. 8882. Springer, Cham. https://doi.org/10.1007/978-3-319-13123-8_5
  » https://doi.org/10.1007/978-3-319-13123-8_5
• ⁶³
  Bouyioukos C, Kim J. Gene regulatory network properties linked to gene expression dynamics in spatially extended systems. LNAI. 2009; 5777: 321-8.
• ⁶⁴
  Burda Z, Krzywicki A, Martin OC, Zagorski M. Motifs emerge from function in model gene regulatory networks. Proc Natl Acad Sci USA. 2011; 108(42): 17263-8.
• ⁶⁵
  RStudio Team. RStudio: integrated development for R. RStudio, PBC, Boston, MA. 2020 [cited 2022 Mar 12]. Available from: http://www.rstudio.com/
  » http://www.rstudio.com/
• ⁶⁶
  Ramos TG. Reconstrução da rede metabólica da Pseudomonas aeruginosa CCBH4851. 2018 [cited 2022 Mar 12]. Available from: https://www.arca.fiocruz.br/handle/icict/29528
  » https://www.arca.fiocruz.br/handle/icict/29528
• ⁶⁷
  Csardi G, Nepusz T. The Igraph Software package for complex network research. InterJournal Complex Systems. 2006; 1695(5): 1-9.
• ⁶⁸
  Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D, et al. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res. 2003; 13(11): 2498-504.
• ⁶⁹
  Delic-Attree I, Toussaint B, Vignais PM. Cloning and sequence analyses of the genes coding for the integration host factor (IHF) and HU proteins of Pseudomonas aeruginosa. Gene. 1995; 154(1): 61-4.
• ⁷⁰
  Aramaki H, Fujita M. In vitro transcription analysis of rpoD in Pseudomonas aeruginosa PAO1. FEMS Microbiol Lett. 1999; 180(2): 311-6.
• ⁷¹
  Caiazza NC, O'Toole GA. SadB is required for the transition from reversible to irreversible attachment during biofilm formation by Pseudomonas aeruginosa PA14. J Bacteriol. 2004; 186(14): 4476-85.
• ⁷²
  Martin DW, Schurr MJ, Yu H, Deretic V. Analysis of promoters controlled by the putative sigma factor AlgU regulating conversion to mucoidy in Pseudomonas aeruginosa: relationship to sigma E and stress response. J Bacteriol. 1994; 176(21): 6688-96.
• ⁷³
  Brinkman FS, Schoofs G, Hancock RE, De Mot R. Influence of a putative ECF sigma factor on expression of the major outer membrane protein, OprF, in Pseudomonas aeruginosa and Pseudomonas fluorescens. J Bacteriol. 1999; 181(16): 4746-54.
• ⁷⁴
  Fujita M, Tanaka K, Takahashi H, Amemura A. Transcription of the principal sigma-factor genes, rpoD and rpoS, in Pseudomonas aeruginosa is controlled according to the growth phase. Mol Microbiol. 1994; 13(6): 1071-7.
• ⁷⁵
  Starnbach MN, Lory S. The fliA (rpoF) gene of Pseudomonas aeruginosa encodes an alternative sigma factor required for flagellin synthesis. Mol Microbiol. 1992; 6(4): 459-69.
• ⁷⁶
  Benvenisti L, Koby S, Rutman A, Giladi H, Yura T, Oppenheim AB. Cloning and primary sequence of the rpoH gene from Pseudomonas aeruginosa. Gene. 1995; 155(1): 73-6.
• ⁷⁷
  Parkins MD, Ceri H, Storey DG. Pseudomonas aeruginosa GacA, a factor in multihost virulence, is also essential for biofilm formation. Mol Microbiol. 2001; 40(5): 1215-26.
• ⁷⁸
  Lizewski SE, Lundberg DS, Schurr MJ. The transcriptional regulator AlgR is essential for Pseudomonas aeruginosa pathogenesis. Infect Immun. 2002; 70(11): 6083-93.
• ⁷⁹
  Waligora EA, Ramsey DM, Pryor Jr EE, Lu H, Hollis T, Sloan GP, et al. AmrZ beta-sheet residues are essential for DNA binding and transcriptional control of Pseudomonas aeruginosa virulence genes. J Bacteriol. 2010; 192(20): 5390-401.
• ⁸⁰
  Gambello MJ, Iglewski BH. Cloning and characterization of the Pseudomonas aeruginosa lasR gene, a transcriptional activator of elastase expression. J Bacteriol. 1991; 173(9): 3000-9.
• ⁸¹
  Arora SK, Ritchings BW, Almira EC, Lory S, Ramphal R. A transcriptional activator, FleQ, regulates mucin adhesion and flagellar gene expression in Pseudomonas aeruginosa in a cascade manner. J Bacteriol. 1997; 179(17): 5574-81.
• ⁸²
  Barton HA, Johnson Z, Cox CD, Vasil AI, Vasil ML. Ferric uptake regulator mutants of Pseudomonas aeruginosa with distinct alterations in the iron-dependent repression of exotoxin A and siderophores in aerobic and microaerobic environments. Mol Microbiol. 1996; 21(5): 1001-17.
• ⁸³
  Cunliffe HE, Merriman TR, Lamont IL. Cloning and characterization of pvdS, a gene required for pyoverdine synthesis in Pseudomonas aeruginosa: PvdS is probably an alternative sigma factor. J Bacteriol. 1995; 177(10): 2744-50.
• ⁸⁴
  Beare PA, For RJ, Martin LW, Lamont IL. Siderophore-mediated cell signalling in Pseudomonas aeruginosa: divergent pathways regulate virulence factor production and siderophore receptor synthesis. Mol Microbiol. 2003; 47(1): 195-207.
• ⁸⁵
  LaBauve AE, Wargo MJ. Detection of host-derived sphingosine by Pseudomonas aeruginosa is important for survival in the murine lung. PLoS Pathog. 2014; 10(1): e1003889.
• ⁸⁶
  Déziel E, Gopalan S, Tampakaki AP, Lépine F, Padfield KE, Saucier M, et al. The contribution of MvfR to Pseudomonas aeruginosa pathogenesis and quorum sensing circuitry regulation: multiple quorum sensing-regulated genes are modulated without affecting lasRI, rhlRI or the production of N-acyl-L-homoserine lactones. Mol Microbiol. 2005; 55(4): 998-1014.
• ⁸⁷
  Liang H, Deng X, Li X, Ye Y, Wu M. Molecular mechanisms of master regulator VqsM mediating quorum-sensing and antibiotic resistance in Pseudomonas aeruginosa. Nucleic Acids Res. 2014; 42(16): 10307-20.
• ⁸⁸
  Rompf A, Hungerer C, Hoffmann T, Lindenmeyer M, Römling U, Gross U, et al. Regulation of Pseudomonas aeruginosa hemF and hemN by the dual action of the redox response regulators Anr and Dnr. Mol Microbiol. 1998; 29(4): 985-97.
• ⁸⁹
  Ochsner UA, Koch AK, Fiechter A, Reiser J. Isolation and characterization of a regulatory gene affecting rhamnolipid biosurfactant synthesis in Pseudomonas aeruginosa. J Bacteriol. 1994; 176(7): 2044-54.
• ⁹⁰
  Brint JM, Ohman DE. Synthesis of multiple exoproducts in Pseudomonas aeruginosa is under the control of RhlR-RhlI, another set of regulators in strain PAO1 with homology to the autoinducer-responsive LuxR-LuxI family. J Bacteriol. 1995; 177(24): 7155-63.
• ⁹¹
  Köhler T, Epp SF, Curty LK, Pechère JC. Characterization of MexT, the regulator of the MexE-MexF-OprN multidrug efflux system of Pseudomonas aeruginosa. J Bacteriol. 1999; 181(20): 6300-5.
• ⁹²
  Heinrichs DE, Poole K. Cloning and sequence analysis of a gene (pchR) encoding an AraC family activator of pyochelin and ferripyochelin receptor synthesis in Pseudomonas aeruginosa. J Bacteriol. 1993; 175(18): 5882-9.
• ⁹³
  Lu CD, Yang Z, Li W. Transcriptome analysis of the ArgR regulon in Pseudomonas aeruginosa. J Bacteriol. 2004; 186(12): 3855-61.
• ⁹⁴
  Wargo MJ, Ho TC, Gross MJ, Whittaker LA, Hogan DA. GbdR regulates Pseudomonas aeruginosa plcH and pchP transcription in response to choline catabolites. Infect Immun. 2009; 77(3): 1103-11.
• ⁹⁵
  Piddock LJ, Johnson MM, Simjee S, Pumbwe L. Expression of efflux pump gene pmrA in fluoroquinolone-resistant and -susceptible clinical isolates of Streptococcus pneumoniae. Antimicrob Agents Chemother. 2002; 46(3): 808-12.
• ⁹⁶
  McPhee JB, Lewenza S, Hancock RE. Cationic antimicrobial peptides activate a two-component regulatory system, PmrA-PmrB, that regulates resistance to polymyxin B and cationic antimicrobial peptides in Pseudomonas aeruginosa. Mol Microbiol. 2003; 50(1): 205-17.
• ⁹⁷
  Miyazaki H, Kato H, Nakazawa T, Tsuda M. A positive regulatory gene, pvdS, for expression of pyoverdin biosynthetic genes in Pseudomonas aeruginosa PAO. Mol Gen Genet. 1995; 248(1): 17-24.
• ⁹⁸
  Palma M, Zurita J, Ferreras JA, Worgall S, Larone DH, Shi L, et al. Pseudomonas aeruginosa SoxR does not conform to the archetypal paradigm for SoxR-dependent regulation of the bacterial oxidative stress adaptive response. Infect Immun. 2005; 73(5): 2958-66.
• ⁹⁹
  Faure LM, Llamas MA, Bastiaansen KC, de Bentzmann S, Bigot S. Phosphate starvation relayed by PhoB activates the expression of the Pseudomonas aeruginosa svreI ECF factor and its target genes. Microbiology (Reading). 2013; 159(Pt 7): 1315-27.
• ¹⁰⁰
  Blus-Kadosh I, Zilka A, Yerushalmi G, Banin E. The effect of pstS and phoB on quorum sensing and swarming motility in Pseudomonas aeruginosa. PLoS One. 2013; 8(9): e74444.
• ¹⁰¹
  Beatson SA, Whitchurch CB, Sargent JL, Levesque RC, Mattick JS. Differential regulation of twitching motility and elastase production by Vfr in Pseudomonas aeruginosa. J Bacteriol. 2002; 184(13): 3605-13.
• ¹⁰²
  Trunk K, Benkert B, Quäck N, Münch R, Scheer M, Garbe J, et al. Anaerobic adaptation in Pseudomonas aeruginosa: definition of the Anr and Dnr regulons. Environ Microbiol. 2010; 12(6): 1719-33.
• ¹⁰³
  Rampioni G, Schuster M, Greenberg EP, Zennaro E, Leoni L. Contribution of the RsaL global regulator to Pseudomonas aeruginosa virulence and biofilm formation. FEMS Microbiol Lett. 2009; 301(2): 210-7.
• ¹⁰⁴
  do Nascimento APB, Medeiros Filho F, Pauer H, Antunes LCM, Sousa H, Senger H, et al. Characterization of a SPM-1 metallo-beta-lactamase-producing Pseudomonas aeruginosa by comparative genomics and phenotypic analysis. Sci Rep. 2020; 10(1): 13192.
• ¹⁰⁵
  Zhu X, Gerstein M, Snyder M. Getting connected: analysis and principles of biological networks. Genes Dev. 2007; 21(9): 1010-24.
• ¹⁰⁶
  Hao D, Li C. The dichotomy in degree correlation of biological networks. PLoS One. 2011; 6(12): e28322.
• ¹⁰⁷
  Bales ME, Johnson SB. Graph theoretic modeling of large-scale semantic networks. J Biomed Inform. 2006; 39(4): 451-64.
• ¹⁰⁸
  Mathee K, Narasimhan G, Valdes C, Qiu X, Matewish JM, Koehrsen M, et al. Dynamics of Pseudomonas aeruginosa genome evolution. Proc Natl Acad Sci USA. 2008; 105(8): 3100-5.
• ¹⁰⁹
  Madar D, Dekel E, Bren A, Alon U. Negative auto-regulation increases the input dynamic-range of the arabinose system of Escherichia coli. BMC Syst Biol. 2011; 5: 111.
• ¹¹⁰
  Martínez-Antonio A, Janga SC, Thieffry D. Functional organisation of Escherichia coli transcriptional regulatory network. J Mol Biol. 2008; 381(1): 238-47.
• ¹¹¹
  Adhya S, Garges S. Positive control. J Biol Chem. 1990; 265(19): 10797-800.
• ¹¹²
  Shaw K. Negative transcription regulation in prokaryotes. Nature Education. 2008; 1(1): 122.
• ¹¹³
  Ferrell Jr JE. Feedback loops and reciprocal regulation: recurring motifs in the systems biology of the cell cycle. Curr Opin Cell Biol. 2013; 25(6): 676-86.
• ¹¹⁴
  Becskei A, Serrano L. Engineering stability in gene networks by autoregulation. Nature. 2000; 405(6786): 590-3.
• ¹¹⁵
  Nevozhay D, Adams RM, Murphy KF, Josic K, Balázsi G. Negative autoregulation linearizes the dose-response and suppresses the heterogeneity of gene expression. Proc Natl Acad Sci USA. 2009; 106(13): 5123-8.
• ¹¹⁶
  Baynham PJ, Wozniak DJ. Identification and characterization of AlgZ, an AlgT-dependent DNA-binding protein required for Pseudomonas aeruginosa algD transcription. Mol Microbiol. 1996; 22(1): 97-108.
• ¹¹⁷
  Cirz RT, O'Neill BM, Hammond JA, Head SR, Romesberg FE. Defining the Pseudomonas aeruginosa SOS response and its role in the global response to the antibiotic ciprofloxacin. J Bacteriol. 2006; 188(20): 7101-10.
• ¹¹⁸
  Maxon ME, Redfield B, Cai XY, Shoeman R, Fujita K, Fisher W, et al. Regulation of methionine synthesis in Escherichia coli: effect of the MetR protein on the expression of the metE and metR genes. Proc Natl Acad Sci USA. 1989; 86(1): 85-9.
• ¹¹⁹
  Yeung AT, Torfs EC, Jamshidi F, Bains M, Wiegand I, Hancock RE, et al. Swarming of Pseudomonas aeruginosa is controlled by a broad spectrum of transcriptional regulators, including MetR. J Bacteriol. 2009; 191(18): 5592-602.
• ¹²⁰
  Hamood AN, Colmer JA, Ochsner UA, Vasil ML. Isolation and characterization of a Pseudomonas aeruginosa gene, ptxR, which positively regulates exotoxin A production. Mol Microbiol. 1996; 21(1): 97-110.
• ¹²¹
  Rampioni G, Schuster M, Greenberg EP, Bertani I, Grasso M, Venturi V, et al. RsaL provides quorum sensing homeostasis and functions as a global regulator of gene expression in Pseudomonas aeruginosa. Mol Microbiol. 2007; 66(6): 1557-65.
• ¹²²
  Crews ST, Pearson JC. Transcriptional autoregulation in development. Curr Biol. 2009; 19(6): R241-6.
• ¹²³
  Hoot SJ, Brown RP, Oliver BG, White TC. The UPC2 promoter in Candida albicans contains two cis-acting elements that bind directly to Upc2p, resulting in transcriptional autoregulation. Eukaryot Cell. 2010; 9(9): 1354-62.
• ¹²⁴
  Zhang Q, Bhattacharya S, Conolly RB, Clewell HJ, Kaminski NE, Andersen ME. Molecular signaling network motifs provide a mechanistic basis for cellular threshold responses. Environ Health Perspect. 2014; 122(12): 1261-70.
• ¹²⁵
  Zielinski NA, Maharaj R, Roychoudhury S, Danganan CE, Hendrickson W, Chakrabarty AM. Alginate synthesis in Pseudomonas aeruginosa: environmental regulation of the algC promoter. J Bacteriol. 1992; 174(23): 7680-8.
• ¹²⁶
  Anupama R, Mukherjee A, Babu S. Gene-centric metegenome analysis reveals diversity of Pseudomonas aeruginosa biofilm gene orthologs in fresh water ecosystem. Genomics. 2018; 110(2): 89-97.
• ¹²⁷
  Blier AS, Veron W, Bazire A, Gerault E, Taupin L, Vieillard J, et al. C-type natriuretic peptide modulates quorum sensing molecule and toxin production in Pseudomonas aeruginosa. Microbiology (Reading). 2011; 157(Pt 7): 1929-44.
• ¹²⁸
  Fan Z, Xu C, Pan X, Dong Y, Ren H, Jin Y, et al. Mechanisms of RsaL mediated tolerance to ciprofloxacin and carbenicillin in Pseudomonas aeruginosa. Curr Genet. 2019; 65(1): 213-22.
• ¹²⁹
  Runyen-Janecky LJ, Albus AM, Iglewski BH, West SEH. The transcriptional activator Vfr binds to two apparently different binding sites in the promoters of Pseudomonas aeruginosa virulence genes. In 96th General Meeting of the American Society for Microbiology. New Orleans, LA. 1996.
• ¹³⁰
  Coppola PE, Gaibani P, Sartor C, Ambretti S, Lewis RE, Sassi C, et al. Ceftolozane-tazobactam treatment of hypervirulent multidrug resistant Pseudomonas aeruginosa infections in neutropenic patients. Microorganisms. 2020; 8(12): 2055.
• ¹³¹
  Suh SJ, Runyen-Janecky LJ, Maleniak TC, Hager P, MacGregor CH, Zielinski-Mozny NA, et al. Effect of vfr mutation on global gene expression and catabolite repression control of Pseudomonas aeruginosa. Microbiology (Reading). 2002; 148(Pt 5): 1561-9.
• ¹³²
  Glessner A, Smith RS, Iglewski BH, Robinson JB. Roles of Pseudomonas aeruginosa las and rhl quorum-sensing systems in control of twitching motility. J Bacteriol. 1999; 181(5): 1623-9.
• ¹³³
  Inat G, Siriken B, Baskan C, Erol I, Yildirim T, Çiftci A. Quorum sensing systems and related virulence factors in Pseudomonas aeruginosa isolated from chicken meat and ground beef. Sci Rep. 2021; 11(1): 15639.
• ¹³⁴
  Kanack KJ, Runyen-Janecky LJ, Ferrell EP, Suh SJ, West SEH. Characterization of DNA-binding specificity and analysis of binding sites of the Pseudomonas aeruginosa global regulator, Vfr, a homologue of the Escherichia coli cAMP receptor protein. Microbiology. 2006; 152(Pt 12): 3485-96.
• ¹³⁵
  Rampioni G, Schuster M, Greenberg EP, Bertani I, Grasso M, Venturi V, et al. RsaL provides quorum sensing homeostasis and functions as a global regulator of gene expression in Pseudomonas aeruginosa. Mol Microbiol. 2007; 66(6): 1557-65.
• ¹³⁶
  Higgins S, Heeb S, Rampioni G, Fletcher MP, Williams P, Cámara M. Differential Regulation of the phenazine biosynthetic operons by quorum sensing in Pseudomonas aeruginosa PAO1-N. Front Cell Infect Microbiol. 2018; 8: 252.
• ¹³⁷
  Sun S, Chen B, Jin ZJ, Zhou L, Fang YL, Thawai C, et al. Characterization of the multiple molecular mechanisms underlying RsaL control of phenazine-1-carboxylic acid biosynthesis in the rhizosphere bacterium Pseudomonas aeruginosa PA1201. Mol Microbiol. 2017; 104(6): 931-47.
• ¹³⁸
  Hraiech S, Brégeon F, Rolain JM. Bacteriophage-based therapy in cystic fibrosis-associated Pseudomonas aeruginosa infections: rationale and current status. Drug Des Devel Ther. 2015; 9: 3653-63.
• ¹³⁹
  Pang Z, Raudonis R, Glick BR, Lin TJ, Cheng Z. Antibiotic resistance in Pseudomonas aeruginosa: mechanisms and alternative therapeutic strategies. Biotechnol Adv. 2019; 37(1): 177-192.
• ¹⁴⁰
  Hay ID, Wang Y, Moradali MF, Rehman ZU, Rehm BH. Genetics and regulation of bacterial alginate production. Environ Microbiol. 2014; 16(10): 2997-3011.
• ¹⁴¹
  Leoni L, Orsi N, de Lorenzo V, Visca P. Functional analysis of PvdS, an iron starvation sigma factor of Pseudomonas aeruginosa. J Bacteriol. 2000; 182(6): 1481-91.
• ¹⁴²
  Tenover FC. Mechanisms of antimicrobial resistance in bacteria. Am J Med. 2006; 119(6): 3-10.
• ¹⁴³
  Jo JT, Brinkman FS, Hancock RE. Aminoglycoside efflux in Pseudomonas aeruginosa: involvement of novel outer membrane proteins. Antimicrob Agents Chemother. 2003; 47(3): 1101-11.
• ¹⁴⁴
  Baraquet C, Harwood CS. FleQ DNA binding consensus sequence revealed by studies of FleQ-dependent regulation of biofilm gene expression in Pseudomonas aeruginosa. J Bacteriol. 2015; 198(1): 178-86.
• ¹⁴⁵
  Zegans ME, Wozniak D, Griffin E, Toutain-Kidd CM, Hammond JH, Garfoot A, et al. Pseudomonas aeruginosa exopolysaccharide Psl promotes resistance to the biofilm inhibitor polysorbate 80. Antimicrob Agents Chemother. 2012; 56(8): 4112-22.
• ¹⁴⁶
  Maseda H, Saito K, Nakajima A, Nakae T. Variation of the mexT gene, a regulator of the MexEF-oprN efflux pump expression in wild-type strains of Pseudomonas aeruginosa. FEMS Microbiol Lett. 2000; 192(1): 107-12.
• ¹⁴⁷
  Sakhtah H, Koyama L, Zhang Y, Morales DK, Fields BL, Price-Whelan A, et al. The Pseudomonas aeruginosa efflux pump MexGHI-OpmD transports a natural phenazine that controls gene expression and biofilm development. Proc Natl Acad Sci USA. 2016; 113(25): E3538-47.
• ¹⁴⁸
  Webber MA, Piddock LJ. The importance of efflux pumps in bacterial antibiotic resistance. J Antimicrob Chemother. 2003; 51(1): 9-11.
• ¹⁴⁹
  Lee JY, Ko KS. Mutations and expression of PmrAB and PhoPQ related with colistin resistance in Pseudomonas aeruginosa clinical isolates. Diagn Microbiol Infect Dis. 2014; 78(3): 271-6.
• ¹⁵⁰
  Alav I, Sutton JM, Rahman KM. Role of bacterial efflux pumps in biofilm formation. J Antimicrob Chemother. 2018; 73(8): 2003-20.
• ¹⁵¹
  Leid JG. Bacterial biofilms resist key host defenses. Microbe. 2009; 4(2): 66-70.
• ¹⁵²
  Zhang Q, Bhattacharya S, Conolly RB, Clewell HJ, Kaminski NE, Andersen ME. Molecular signaling network motifs provide a mechanistic basis for cellular threshold responses. Environ Health Perspect. 2014; 122(12): 1261-70.

• Financial support: INOVA-FIOCRUZ, FAPERJ, CAPES.

Publication Dates

• Publication in this collection
  14 Oct 2022
• Date of issue
  2022

History

• Received
  12 May 2022
• Accepted
  09 Sept 2022