Computational prediction of secretion systems and secretomes of Brucella: identification of novel type IV effectors and their interaction with the host

Abstract

Brucella spp. are facultative intracellular pathogens that cause brucellosis in various mammals including humans. Brucella survive inside the host cells by forming vacuoles and subverting host defence systems. This study was aimed to predict the secretion systems and the secretomes of Brucella spp. from 39 complete genome sequences available in the databases. Furthermore, an attempt was made to identify the type IV secretion effectors and their interactions with host proteins. We predicted the secretion systems of Brucella by the KEGG pathway and SecReT4. Brucella secretomes and type IV effectors (T4SEs) were predicted through genome-wide screening using JVirGel and S4TE, respectively. Protein–protein interactions of Brucella T4SEs with their hosts were analyzed by HPIDB 2.0. Genes coding for Sec and Tat pathways of secretion and type I (T1SS), type IV (T4SS) and type V (T5SS) secretion systems were identified and they are conserved in all the species of Brucella. In addition to the well-known VirB operon coding for the type IV secretion system (T4SS), we have identified the presence of additional genes showing homology with T4SS of other organisms. On the whole, 10.26 to 14.94% of total proteomes were found to be either secreted (secretome) or membrane associated (membrane proteome). Approximately, 1.7 to 3.0% of total proteomes were identified as type IV secretion effectors (T4SEs). Prediction of protein–protein interactions showed 29 and 36 host–pathogen specific interactions between Bos taurus (cattle)–B. abortus and Ovis aries (sheep)–B. melitensis, respectively. Functional characterization of the predicted T4SEs and their interactions with their respective hosts may reveal the secrets of host specificity of Brucella.

---

Introduction

Brucellosis is a zoonotic disease caused by Brucella spp.¹Brucella infects a wide range of mammals and it is transmitted to humans by the consumption of contaminated dairy products and direct contact with the infected animals. Human brucellosis is a debilitating disease with diverse pathological manifestations such as undulant fever, osteoarticular complications, endocarditis and several neurological disorders.²Brucella is a Gram-negative, facultative intracellular pathogen capable of infecting numerous cell types, including epithelial cells, placental trophoblasts, dendritic cells and macrophages.³Brucella resides and proliferates inside the host cells by forming Brucella-containing vacuole (BCV).

Secretion of proteins is a major virulence mechanism in bacterial pathogens and the secretion system plays a major role in the pathogen–host interactions. Bacteria secrete proteins through various types of secretion systems classified as type I to type VI. General secretory (Sec) pathway and twin-arginine translocation (Tat) pathway assist the translocation of proteins from the cytoplasm to the membrane bound secretion systems.⁴ Majority of the secreted proteins in bacteria are secreted via the Sec pathway. The type I secretion system (T1SS) consists of 3 proteins, a pore-forming outer membrane protein (OMP), a membrane fusion protein (MFP) and an inner membrane ATP-binding cassette (ABC) protein.⁵ The T1SS is involved in the transport of cyclic β-glucans, polysaccharides and a wide variety of small peptide signalling molecules including bacteriocins.⁶ The type II secretion system (T2SS) is involved in the transport of proteins from the periplasm to the outer membrane. T2SS depends on either the Sec or Tat pathway for the initial translocation of the proteins from the cytoplasm to the periplasm. Most of the hydrolases and toxins are secreted via T2SS.⁷

The type III secretion system (T3SS) is homologous to the bacterial flagellum, which can directly inject proteins into the host cells.⁸ Similarly, the type IV secretion system (T4SS), which plays major roles in bacterial pathogenesis, can directly inject proteins into the host cells.⁹ The T4SS is evolutionarily related to the conjugation system. The T4SS is an important virulence factor in several bacterial pathogens, such as Helicobacter pylori, Legionella pneumophila, Bartonella spp. and Brucella spp. (Zechner et al., 2012). The proteins specifically secreted through T4SS, called as type IV secreted effectors (T4SEs), exert important functions inside the infected host cells.¹⁰ The type V secretion system (T5SS) is the simplest protein secretory apparatus. Proteins secreted by T5SS are secreted via either the auto-transporter system (type Va) or the two partner secretion pathway (type Vb).¹¹ The type VI secretion system (T6SS) is recently identified that constitutes a phage-tail-spike-like structure, which can inject proteins directly into the host cells.¹²

Computational methods can provide timely information on secretion systems and secretome from genome sequences. For example, the SignalP tool predicts secretory proteins based on the signal peptides and transmembrane regions.¹³ SecretomeP predicts non-classical signal peptide-independent secretion proteins based on various other post-translational and localizational aspects.¹⁴ Based on the composition of reduced amino acid alphabet (RAAA) obtained from the protein blocks and increment of diversity, an ID_RAAA approach was developed for the prediction of antimicrobial peptides and defensins.¹⁵ Verma et al.¹⁶ developed a support vector machine (SVM) model based on the position specific scoring matrix (PSSM) profile for predicting secretory proteins in malaria parasite.¹⁶ Zuo and Li¹⁷ developed k-minimum increment of diversity approach to predict the secretory proteins in Plasmodium falciparum.¹⁷ Lin et al.¹⁸ predicted the secretome of Plasmodium falciparum using the iSMP-Grey approach, which is based on the pseudo amino acid composition (PseAAC) of proteins.¹⁸ Zhou et al.¹⁹ developed Lactobacillales cluster of ortholog groups (LaCOGs) for the prediction of secretomes in lactic acid bacteria by integrating protein sub-cellular location prediction and homology clustering methods.¹⁹ In addition to these general secretome prediction tools, specific tools have been developed for the prediction of proteins secreted through T3SS and T4SS. Arnold et al.²⁰ developed a type III secretion effector (T3SE) prediction tool based on T3S signal in the N-terminus of the signal peptide and frequencies of amino acids with certain physico-chemical properties.²⁰ Similarly, Meyer et al.²¹ developed a searching algorithm for type IV effector proteins (S4TE) to identify T4SEs based on 13 different parameters.²¹

Computational methods to categorize the secretion systems and potential secreted proteins followed by their functional characterization would be a valuable approach to understand the pathogen–host interactions. However, only limited information is available on the secretion systems and secretomes of Brucella. In this study, the secretion systems and the secretome of Brucella spp. were predicted through systematic screening of all the open reading frames of 39 complete Brucella genomes available in the databases. Essentially, the T4SS of Brucella plays major roles in intracellular survival and multiplication.^22,23 The T4SS coded by the virB operon is reported as essential for the intracellular survival of Brucella and the mutants that lack T4SS failed to establish infection in vitro.²⁴ T4SS is also reported as essential to escape from the host endocytic pathway.²³ However, only minimal information is available on Brucella T4SEs. Therefore, an attempt was made to identify Brucella T4SEs and their interactions with host proteins.

Materials and methods

Retrieval of genome sequences

Complete genome sequences of Brucella spp. were obtained from the National Center for Biotechnology Information (NCBI) database (http://ftp://ftp.ncbi.nih.gov/genomes/Bacteria/). A total of 39 complete genomes were used in this study.

Prediction of secretion systems

Components of the secretion systems and Sec and Tat pathways were predicted using the KEGG (Kyoto Encyclopedia of Genes and Genomes) database.²⁵ The T4SS was predicted using SecReT4 (Type IV Secretion system Resource), which annotates and locates the components of T4SS on the genome sequences.²⁶ The STRING database²⁷ was used to study possible interactions among proteins, including physical and/or functional interactions of secretion systems, Sec and Tat pathway components.

Prediction of secretome

JVirGel 2.0 was used for the simulation and analysis of proteomics data, which determines the theoretical isoelectric points (pI) and the calculated molecular weights (MW) of proteins and visualizes as a virtual two-dimensional (2D) protein map. Based on the PrediSi and JCaMelix tools available at JVirGel 2.0, secretomes and membrane proteomes, respectively, of Brucella were identified with default parameters.²⁸ T4SE were predicted using a default parameter of the S4TE tool. S4TE predicts and ranks T4SE candidates by using a combination of 13 sequence characteristics, including homology to known effectors, homology to eukaryotic domains, presence of sub-cellular localization signals or secretion signals, etc.²¹ We applied Proteinortho²⁹ and OrthoMCL-DB³⁰ tools for the comparative analysis and to compute orthologous proteins of secretome and T4SE. Similarly OrthoMCL³¹ was used to predict the orthologous group of host proteins.

Prediction of Brucella T4SE and host protein–protein interaction networks

We retrieved proteome of Bos taurus (cattle) (proteome id: UP000009136; 23 866 proteins) and Ovis aries (sheep) (proteome id: UP000002356; 23 112 proteins) from UniProt. Protein–protein interactions (PPI) between the host proteome and the respective host specific Brucella (B. abortus 2308 and B. melitensis NI) T4SEs were predicted using the Host–Pathogen Interaction DataBase (HPIDB)³² with pathogen percent identity and query coverage as 30% and host percentage identity and query coverage as 50%, respectively. The HPIDB predicts PPI based on the homology approach from the plentiful template eukaryotic–prokaryote inter-species PPIs available among 68 hosts and 602 pathogens. The PPI network analysis and its visualization were performed using cytoscape.³³ Cellular localization of the host proteins interacting with T4SEs of Brucella were predicted using EuLoc,³⁴ CELLO2GO³⁵ and AgBase.³⁶

Results

Availability of complete genome sequences of B. abortus, B. canis, B. ceti, B. melitensis, B. microti, B. pinnipedialis, B. ovis and B. suis provided an opportunity to identify the secretion systems and secretomes of these species. A total of 39 complete genome sequences of Brucella spp. were used in this study. The complete genome of B. abortus 2308 was used for the initial analysis and subsequently extended to other genomes.

Secretion systems of Brucella

Based on the comprehensive analyses of genome sequences using the KEGG database, the presence of three types of secretion systems (Types I, IV and V) were identified in Brucella spp. (Fig. 1). Genes coding for type II, III and VI secretion systems are missing in all the 39 genome sequences and thus, Brucella lacks these secretion systems. For the prediction of secretion systems, Escherichia coli O157:H7 EC4115 (NC_011353) and Yersinia pseudotuberculosis IP32953 (NC_006155) were used as query sequences for homology search against the Brucella spp. using BLASTP algorithm. The T1SS export the proteins from the cytoplasm to the extracellular environment in a single step without involving periplasmic intermediate. The type IV system is made up of a multiprotein complex, which secretes the proteins directly into the host cells and those proteins are referred as type IV secretion effectors (T4SEs). The T5SS uses the Sec or Tat pathway for crossing the cytoplasmic membrane into the periplasm and subsequently processed by the outer membrane T5SS complex for extracellular secretion.

Fig. 1 Predicted model of secretion systems in Brucella.

General secretory (Sec) and twin-arginine translocation pathway (Tat) genes

The majority of proteins to be exported across the cytoplasmic membrane or integration into the membrane are handled by the evolutionarily conserved Sec-system. The Sec-system mobilizes the proteins through the translocon in an unfolded manner. Brucella genomes contains genes coding for SecA, which contains ATPase that drives protein movement into and across the membrane, SecB, a bifunctional preprotein translocase subunit, SecD/F, preprotein translocase subunit, SecE, SecG, SecY, YajC, YidC, inner membrane protein translocase components, signal recognition particle (SRP) protein ffh and signal recognition particle-docking protein FtsY, altogether constitute the functional Sec machinery (Table 1). Most of the proteins secreted via the Sec pathway are originally synthesized with an amino-terminal extension of 15–30 amino acids called the signal or leader peptide. The signal peptides are cleaved by distinct enzymes called signal peptidases (SPases). SPases play an indispensable role in the transport of proteins across membranes. In Brucella, two types of signal peptidases were identified, namely, SPase I and Spase II. The Tat pathway translocates the protein in folded forms. It translocates proteins into the periplasm and allows integration of inner membrane proteins. TatA, TatB and TatC are located in the Brucella genome as an operon. Possible interactions among Sec, Tat and other secretion systems are shown in Fig. 2. Most of the translocations across the outer membrane involving the Sec and Tat pathway could be processed by T5SS and few proteins could be secreted via T4SS. Sec and Tat pathways transport proteins to the type V secretion system through yidC to nuoK (NADH dehydrogenase subunit K), which shows a stronger interaction with a score value of 0.861 (ESI,† file 1). Likewise, these pathways showed a weaker interaction (a score value of 0.577) with SecY-virB10, a component of the T4SS translocation channel in the Brucella inner membrane (Fig. 2).

Table 1 Components of the secretion systems identified in B. abortus 2308

Components	Reference	Locus Id	Function	Query	Locus id	Coverage (%)	Identify (%)	E value
Type I secretion system
TolC	E. coli O157:H7 EC4115	ECH74115_4349	Pore-forming outer membrane protein	B. abortus 2308	BAB1_0963	83	25	6 × 10^−25
HLyD		ECH74115_B0020	Membrane fusion protein		BAB2_0773	44	23	0.002
HlyB		ECH74115_B0019	Inner membrane ATP-binding cassette (ABC) protein		BAB1_0209	81	34	2 × 10^−75
Type IV secretion system
VirB1	E. coli O157:H7 EC4115	ECH74115_A0009	Punch holes in the cell wall	B. abortus 2308	BAB2_0068	73	35	1 × 10^−29
VirB2		ECH74115_A0010	Pilus subunit		BAB2_0067	32	26	8 × 10^−07
VirB3		ECH74115_A0011	Pass membrane protein		BAB2_0066	56	29	6 × 10^−06
VirB4		ECH74115_A0012	Energizes substrate transport		BAB2_0065	96	27	2 × 10^−83
VirB5		ECH74115_A0008	Pilus subunit		BAB2_0064	38	28	8 × 10^−08
VirB6		ECH74115_A0006	Translocation channel at the inner membrane		BAB2_0063	82	24	7 × 10^−10
VirB7		ECH74115_A0007	Component of a complex to span the cell envelope		BAB2_0062	31	58	3 × 10^−06
VirB8		ECH74115_A0013	Translocation channel at the inner membrane		BAB2_0061	70	33	8 × 10^−28
VirB9		ECH74115_A0014	Component of a complex to span the cell envelope		BAB2_0060	90	27	2 × 10^−27
VirB10		ECH74115_A0015	Translocation channel at the inner membrane		BAB2_0059	70	38	2 × 10^−48
VirB11		ECH74115_A0016	Energizes substrate transport		BAB2_0058	91	65	8 × 10^−70
Type Vc secretion system
YadA	Yersinia pseudotuberculosis IP32953	pYV0013		B. abortus 2308	BAB1_1854	74	31	8 × 10^−09
Sec pathway
SecD	E. coli O157:H7 EC4115	ECH74115_0488	Translocon-associated complex	B. abortus 2308	BAB1_0910	78	33	3 × 10^−64
SecF		ECH74115_0489	Translocon-associated complex		BAB1_1344	74	34	7 × 10^−57
SecE		ECH74115_5446	Translocation subunit		BAB1_1270	72	33	4 × 10^−10
SecG		ECH74115_4497	Translocation subunit		BAB1_1160	98	27	1 × 10^−14
SecY		ECH74115_4623	Translocation subunit		BAB1_1235	94	56	4 × 10^−157
YajC		ECH74115_0487	Translocon-associated complex		BAB1_0909	91	31	3 × 10^−18
YidC		ECH74115_5135	Membrane protein integration and folding		BAB2_0986	85	32	2 × 10^−84
SecA		ECH74115_0106	Translocation ATPase		BAB1_1946	99	54	0.0
FtsY		ECH74115_4784	SRP receptor		BAB1_1934	68	47	5 × 10^−95
SecB		ECH74115_4982	Holdase chaperone and targeting to SecA		BAB1_2073	85	31	2 × 10^−26
ffh		ECH74115_3849	SRP protein component		BAB1_1834	96	52	1 × 10^−152
Tat pathway
TatA	E. coli O157:H7 EC4115	ECH74115_5277	Translocase protein	B. abortus 2308	BAB1_0901	64	42	2 × 10^−10
TatB		ECH74115_5278	Translocase protein		BAB1_0902	45	29	2 × 10^−14
TatC		ECH74115_5279	Translocase protein		BAB1_0903	96	36	9 × 10^−51

---

Fig. 2 Interactions among proteins of Sec, Tat pathways, T4SS and T5SS in Brucella abortus 2308 identified using the STRING database. Stronger associations are represented by thicker lines, while weaker associations are represented by thinner lines.

Type I secretion system

T1SS secretes substrates in a single step without a periplasmic intermediate. T1SS is involved in the secretion of cytotoxins belonging to the RTX (repeats-in-toxin) protein family, proteases, cell surface layer proteins and lipase proteins. It is also involved in the export of non-proteinaceous substrates like cyclic β-glucans and polysaccharides. In Brucella, ABC transporters are probably the most common as well as the most wide-spread active transport systems. Over two-thirds of Brucella transporters are ABC-transporters. The presence of quite a lot of amino acid and peptide transporters is correlated to the intracellular replication niche of Brucella, where peptides and amino acids are abundant.

Type IV secretion system

The proteins involved in T4SS were predicted using SecReT4. The T4SS spans both membranes of the bacteria and translocates proteins across the BCV membrane into the cytoplasm of the host cell. The Brucella T4SS is coded by the virB operon comprising of virB1 to virB11, which is located on the chromosome II (Table 2). The predicted T4SS and the length of their gene coordinates, strands, size, protein ID, E-value that match with the subject protein are shown in the ESI,† file 2 (each sheet represents one Brucella genome). virB1 is a transglycosylase located in the periplasm that provides openings in the peptidoglycan layer for the efficient assembly of the virB2/virB5 pilus. Two ATPases, viz., virB4 and virB11, comprise the cytoplasmic component of the system and are closely associated with the inner membrane, which promotes the ATP-dependent transfer of substrates through the secretion channel. virB4 interacts with virB11 to control the pilus structure during the assembly of the secretion complex. Polytopic membrane protein virB6 has multiple transmembrane helices and forms an inner membrane complex with bitopic membrane proteins, virB8 and virB10. virB9 and the short lipoprotein virB7 form an outer membrane complex. The gene coding for short lipoprotein virB7 was found in all the genomes of Brucella except B. abortus A13334, B. melitensis M28, B. melitensis M5-90, B. ovis ATCC 25840, B. canis HSK A52141 and B. suis biovar 2. virB2 and virB5 are the major and minor extracellular pilus components respectively. virB3 is an inner membrane protein, which is also predicted to be involved in pilus assembly. Once assembled, the multi-protein T4SS can transport effector proteins into the host cells, which is essential for both the survival of Brucella and its replication in infected host cells.

Table 2 Predicted proteins involved in type IV secretion system in Brucella

Brucella strain	virB11	virB10	virB9	virB8	virB7	virB6	virB5	virB4	virB3	virB2	virB1	TrbL	TrbJ	TraI	LysM	FtsK	OmpA	SLT	FlgJ	SLT	SLT
+/− indicates the presence/absence of a protein.
B. abortus A13334	+	+	+	+	−	+	+	+	+	+	+	−	−	−	+	+	+	+	+	+	+
B. abortus S19	+	+	+	+	+	+	+	+	+	+	+	−	−	−	+	+	+	+	+	+	+
B. abortus bv. 1. 9-941	+	+	+	+	+	+	+	+	+	+	+	−	−	−	+	+	+	+	+	+	+
B. abortus 2308	+	+	+	+	+	+	+	+	+	+	+	−	−	−	+	+	+	+	+	−	+
B. abortus 3196	+	+	+	+	+	+	+	+	+	+	+	−	−	−	+	+	+	+	+	+	+
B. abortus 63 75	+	+	+	+	+	+	+	+	+	+	+	−	−	−	+	+	+	+	+	+	+
B. abortus BDW	+	+	+	+	+	+	+	+	+	+	+	−	−	−	+	+	+	+	+	+	+
B. abortus BER	+	+	+	+	+	+	+	+	+	+	+	−	−	−	+	+	+	+	+	+	+
B. abortus BFY	+	+	+	+	+	+	+	+	+	+	+	−	−	−	+	+	+	+	+	+	+
B. abortus NCTC 10505	+	+	+	+	+	+	+	+	+	+	+	−	−	−	+	+	+	+	+	+	+
B. abortus ZW053	+	+	+	+	+	+	+	+	+	+	+	−	−	−	+	+	+	+	+	+	+
B. canis ATCC 23365	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	−	−	+	+
B. canis HSK A52141	+	+	+	+	−	+	+	+	+	+	+	+	+	+	+	+	+	+	−	+	+
B. canis RM6/66	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	−	+	+
B. canis str. Oliveri	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	−	+	+
B. canis SVA13	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	−	+	+	−	+	+
B. ceti TE10759-12	+	+	+	+	+	+	+	+	+	+	+	−	−	−	+	+	+	+	+	+	+
B. ceti TE28753-12	+	+	+	+	+	+	+	+	+	+	+	−	−	−	+	+	+	+	−	+	+
B. melitensis ATCC 23457	+	+	+	+	+	+	+	+	+	+	+	−	−	−	+	+	+	+	−	+	−
B. melitensis M28	+	+	+	+	−	+	+	+	+	+	+	−	−	−	+	+	+	+	+	+	+
B. melitensis M5-90	+	+	+	+	−	+	+	+	+	+	+	−	−	−	+	+	+	+	+	+	+
B. melitensis NI	+	+	+	+	+	+	+	+	+	+	+	−	−	−	+	+	+	+	+	+	+
B. melitensis bv. 1 str. 16M	+	+	+	+	+	+	+	+	+	+	+	−	−	−	+	+	+	+	+	+	+
B. microti CCM 4915	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	−	+	+
B. ovis ATCC 25840	+	+	+	+	−	+	+	+	+	+	+	−	−	−	−	+	+	+	−	+	+
B. pinnipedialis B2/94	+	+	+	+	+	+	+	+	+	+	+	−	−	−	+	+	+	+	+	+	+
B. pinnipedialis 6/566	+	+	+	+	+	+	+	+	+	+	+	−	−	−	+	+	+	+	+	+	+
B. suis 1330	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	−	+	−
B. suis ATCC 23445	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	−
B. suis VBI22	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	−	+	−
B. suis 513UK	+	+	+	+	+	+	+	+	+	+	+	−	−	−	+	+	+	+	+	+	+
B. suis BSP	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	−	+	+
B. suis ZW043	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	−	+	+
B. suis ZW046	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	−	+	+
B. suis bv. 2 Bs143CITA	+	+	+	+	−	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+
B. suis bv. 2 Bs364CITA	+	+	+	+	−	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+
B. suis bv. 2 Bs396CITA	+	+	+	+	−	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+
B. suis bv. 2 PT09143	+	+	+	+	−	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+
B. suis bv. 2 PT09172	+	+	+	+	−	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+

---

In addition to the classical virB operon coding for T4SS, we also noticed 10 more genes, TrbL, TrbJ, TrbI, LysM, FtsK, OmpA, FlgJ and three copies of SLT, showing homology with T4SS of other organisms such as Nitrosomonas eutropha, Methylomonas methanica, Clostridium perfringens, Coxiella burnetii, Sinorhizobium medicae, Haemophilus ducreyi and Enterobacter cloacae shown in the ESI,† file 2. Location of the virB operon and other identified genes of T4SS in B. abortus 2308 are shown in Fig. 3.

Fig. 3 Genomic locations of the predicted T4SS in B. abortus 2308. The predicted virB operon and other horizontally transferred genes associated with T4SS are shown.

Chromosomal location of these genes suggests the likelihood of multiple horizontal transfer events during the course of evolution. The presence and absence of gene encoded by virB operon and accessory genes of T4SS in the complete genomes of Brucella are shown in Table 2. The accessory genes of T4SS and its similarity with genes of other organisms are shown in Table 3. TrbL could be Hamiltonella defensa. TraI, TrbL and TrbJ were found only in B. canis, B. microti and B. suis. Therefore, these genes might be acquired by the common ancestor of these three species. OmpA has critical roles in pathogenesis including adhesion, invasion, or intracellular survival as well as evasion of host defences or stimulators of proinflammatory cytokine production. FlgJ plays a major role in flagellar assembly, which acts as a protective antigen that could provoke both humoral and cell-mediated immune responses.³⁷ Lytic transglycosylases are an important class of bacterial enzymes that act on peptidoglycan with the same substrate specificity as lysozyme. Lytic transglycosylases may assist the assembly of type IV secretion systems via localized lysis of the peptidoglycan.³⁸

Table 3 List of horizontally transferred genes identified as possible components of T4SS in Brucella and their respective homolog

Accessory genes of T4SS in Brucella	Homolog	Similarity (%)
TrbL	TrbL of Nitrosomonas eutropha	40.74
TrbJ	TrbJ of Nitrosomonas eutropha	47.45
TraI	virB8 of Candidatus	31.21
LysM (lytic transglycosylase)	Tfc4 of Methylomonas methanica	33.3
FtsK (cell division protein)	TcpA of Clostridium perfringens	31.2
OmpA family protein	IcmN of Coxiella burnetii	36.3
SLT (transglycosylase SLT domain-containing protein)	virB1 of Vibrio sp.	39.6
FlgJ (flagellar protein)	Tfc22 of Haemophilus ducreyi	40.15
SLT (transglycosylase SLT domain-containing protein)	Orf169_F of Enterobacter cloacae	37.17
SLT (transglycosylase SLT domain-containing protein)	virB1 of Sinorhizobium medicae	37.56

---

Type V secretion system

The type V secretion (T5SS) system includes the autotransporter system (type Va or AT-1), the two partner secretion pathway (type Vb), and the recently described type Vc system (also termed AT-2). In Brucella, the presence of Vc T5SS was predicted, consists of the YadA protein, which promote their pathogenicity and virulence in host cells, through cell adhesion. T5SS needs Sec-SRP and TAT pathways to export the proteins across the cytoplasmic membrane to the periplasm and the outer membrane. T5SS transports the proteins in an unfolded manner. Trimeric autotransporters (TAAs) are important virulence factors in many Gram-negative pathogens. The prototype of all TAAs is YadA, which is known to bind to the collagen of the host extracellular matrix.³⁹

Secretome and membrane proteome of Brucella

Secreted proteins are responsible for sensing stress, substrate binding, adhesion, cell–cell communication, host–microbe interactions and adaptation to the environment and lifestyle of organisms. Membrane proteins are part of the interface between the intracellular and extracellular environment of the cell. The secretomes and membrane proteomes of Brucella strains are predicted using the JVirGel 2.0 tool and the results are shown in Table 4. The number of proteins predicted as secretomes ranges from 316 to 450, which constitutes approximately 10 to 15% of total proteins. The predicted secretomes of 39 Brucella genomes are given in the ESI,† file 3. The number of proteins identified as membrane proteome ranges from 592 to 796, which constitutes approximately 22 to 27% of the total proteome. The predicted membrane proteomes of 39 Brucella genomes are given in the ESI,† file 4. Some of the proteins were predicted in both secretome and membrane proteome.

Table 4 Genome-wide scan showing the number of proteins predicted as secretome, membrane proteome and T4SE in Brucella

Brucella strain	Total proteins	Secretome	Membrane proteome	T4SE	Accession^a
a Accession numbers for chromosome I and II.
B. abortus 2308	3034	382	677	80	NC_007618, NC_007624
B. abortus A13334	3338	402	735	80	NC_016795, NC_016777
B. abortus S19	3000	400	720	83	NC_010742, NC_010740
B. abortus bv. 1 str. 9-941	3084	387	692	75	NC_006932, NC_006933
B. abortus 3196	3068	412	744	81	CP007707, CP007708
B. abortus 63 75	3077	412	742	85	CP007663, CP007662
B. abortus BDW	3111	413	740	75	CP007680, CP007681
B. abortus BER	3060	417	739	80	CP007683, CP007682
B. abortus BFY	3063	418	743	79	CP007738, CP007737
B. abortus NCTC 10505	3080	413	744	66	CP007700, CP007701
B. abortus ZW053	3160	430	780	68	CP009098, CP009099
B. canis ATCC 23365	3251	409	751	80	NC_010103, NC_010104
B. canis HSK A52141	3280	406	752	80	NC_016778, NC_016796
B. canis RM6/66	3079	436	775	82	CP007758, CP007759
B. canis str. Oliveri	3288	406	736	56	HG803176, HG803175
B. canis SVA13	2950	426	753	77	CP007630, CP007629
B. ceti TE10759-12	2611	390	706	78	NC_022905, NC_022906
B. ceti TE28753-12	2376	316	592	65	NC_022907, NC_022908
B. melitensis ATCC 23457	3135	399	716	83	NC_012441, NC_012442
B. melitensis M28	3363	410	755	73	NC_017244, NC_017245
B. melitensis M5-90	3358	412	755	81	NC_017246, NC_017247
B. melitensis NI	3229	396	732	84	NC_017248, NC_017283
B. melitensis bv. 1 str. 16M	3198	328	706	74	NC_003317, NC_003318
B. microti CCM 4915	3282	425	773	71	NC_013119, NC_013118
B. ovis ATCC 25840	2890	357	679	80	NC_009505, NC_009504
B. pinnipedialis B2/94	3324	418	757	78	NC_015857, NC_015858
B. pinnipedialis 6/566	3095	436	764	93	CP007743, CP007742
B. suis 1330	3271	410	744	74	NC_004310, NC_004311
B. suis ATCC 23445	3241	410	748	86	NC_010169, NC_010167
B. suis VBI22	3270	411	745	75	NC_016797, NC_016775
B. suis 513UK	3052	440	766	73	CP007716, CP007717
B. suis BSP	3077	437	772	83	CP008756, CP008757
B. suis ZW043	3090	443	789	86	CP009094, CP009095
B. suis ZW046	3124	450	796	85	CP009096, CP009097
B. suis bv. 2 Bs143CITA	3363	422	774	78	CP007695, CP007696
B. suis bv. 2 Bs364CITA	3372	425	768	82	CP007697, CP007698
B. suis bv. 2 Bs396CITA	3374	424	767	79	CP007721, CP007720
B. suis bv. 2 PT09143	3370	421	773	77	CP007691, CP007692
B. suis bv. 2 PT09172	3288	419	769	76	CP007694, CP007693

---

Type IV secretion effectors (T4SEs)

Intracellular bacterial pathogens secrete specific effector proteins via T4SS to exploit the host cell machinery and hijack the immune responses. A genome-wide scan was performed for protein sequences of 39 complete Brucella genomes using S4TE and the numbers of T4SEs are shown in Table 4. The number of identified T4SEs ranges from 56 to 93, which constitutes approximately 1.7 to 3.0% of total proteomes. In B. abortus 2308, 80 T4SEs were predicted and their T4SE specific motifs/domains are given in the ESI,† file 5 and Table S1. Of these, only seven T4SEs were reported earlier and 73 T4SEs were predicted for the first time in this study.

Approximately, 46% of the T4SEs were predicted to contain coiled-coil motifs. Another 27% of the T4SEs were predicted based on the C-terminal charge/hydrophobicity. Eukaryotic-like domain, including the patatin family, acetyltransf_1 domain, and cNMP_binding domain were identified in approximately 14% of the identified T4SEs. The lists of T4SEs identified from the 39 complete genomes are given in the ESI,† file 6. Among the 80 T4SEs of B. abortus 2308, 15 of them were predicted as part of either secretome or membrane proteome. Therefore, these 15 proteins could be processed via the general Sec pathway and other 65 proteins could be processed directly by T4SS via the virB4,11 complex (Fig. 1). In B. abortus 2308, more than 40% of the predicted T4SEs have a coiled coil structure. Coiled coils are structural motifs of proteins with two α-helices coiled together and these proteins could have roles in the regulation of gene expression by stabilizing transcription factors.⁴⁰ Another 25% of T4SEs were predicted based on C-terminal features such as basicity, charge and hydropathy. Few predicted T4SEs such as TrkA, ATP-dependent helicase and putative monovalent cation/H+ antiporter subunit A contain a nuclear localization signal (NLS), which targets these effectors into the nucleus of the host cell. Therefore, these proteins could be involved in host genetic regulation during infection. A tetracycline resistance protein, a D-alanyl-D-alanine carboxypeptidase and few other hypothetical proteins were predicted to have a mitochondrial localization signal (MLS).

Among the predicted T4SEs, few of them have a characteristic eukaryotic-like domain, which could be involved in interactions with host proteins. Eukaryotic-like domains containing T4SEs include the patatin family, fic/doc family, acetyltransferase domain 1, acetyltransferase domain 10, cNMP binding domain, pyridoxal deC family, tetratrico peptide repeat (TPR) 11 repeat, TPR 16 repeat and SNARE associated family proteins.^41–43Brucella possesses a glutamate decarboxylase system that belongs to the virulence factor of Brucella.⁴⁴ The cyclic nucleotide-binding protein (cNMP) regulates the subunits of cNMP dependent kinases and regulatory subunits of some other protein complexes.⁴⁵Fic (filamentation induced by cAMP) domain containing proteins which mediate bacterial pathogenesis and unrecognized eukaryotic post-translational modifications may regulate key signalling events.⁴⁶ The fic domain is classified together with a second family of sequences called as doc (death on curing).⁴⁷ The patatin-like phospholipases contain a patatin domain with lipid acyl hydrolase activity.⁴⁸N-Acetyltransferase-GCN5, which comes under the acetyl transferase domain is a putative lysine acetylase that regulates metabolic enzyme activity.⁴⁹

Conserved and species specific T4SEs

The virB coded T4SS is conserved in all Brucella spp., but the T4SEs are dynamic. Therefore, a set of species specific T4SEs may contribute to the differences in virulence or host specificity of Brucella spp. The T4SEs like, the trigger factor, ribonuclease E and G, peptidyl-prolyl cis–trans isomerase, the Sel1 repeat-containing protein, the lipopolysaccharide biosynthesis protein, DNA topoisomerase IV subunit A, rhomboid family proteins and three other hypothetical proteins are conserved in all the Brucella species. Crp, GntR and TetR families of transcriptional regulatory proteins, involved in virulence and symbiosis, are also identified in all Brucella spp. The DedA family protein and cell division protein FtsQ are conserved in the human pathogenic Brucella spp., viz., B. abortus, B. melitensis, B. suis and B. canis. Cyclic β-1,2-glucan is required for intracellular Brucella to avoid fusion of the phagosome with lysosomes, which was found in all Brucella species, except B. pinnipedialis. Previously reported virB-coregulated effectors VceA and VceC were also identified in our study. VceA is conserved in all the Brucella species and VceC is missing in B. canis, B. ceti and B. suis. Similarly, previously reported T4SEs like BAB1_1043, BAB1_2005, and BAB1_1275 are conserved in all Brucella spp., while BAB2_0123 is missing in B. ovis. Cobalamin synthesing protein P47K was found only in B. abortus. Similarly, unique hypothetical proteins were identified in different species of Brucella, viz., BMEA_A2173 in B. melitensis, BOV_0178 in B. ovis, and BCAN_A0397 in B. canis. Therefore, the presence of these unique species specific T4SEs could be at least partly responsible for host specificity of Brucella.

We also analyzed the core- and pan-T4SEs of Brucella spp. From the 39 complete genomes analyzed, a total of 25 T4SEs were identified as the core-T4SEs (present in all the genomes of Brucella). Overall, 201 pan-T4SEs were identified from these genomes. Furthermore, we also examined the species level core T4SEs for B. abortus, B. melitensis, B. canis and B. suis. For these species, at least five complete genomes are available. The predicted core-T4SEs of each species and shared T4SEs among the four species are given in the Venn diagram (Fig. 4). In B. abortus, 46 T4SEs are conserved in all the 11 genomes analyzed. Similarly, 54 T4SEs are conserved in five B. melitensis genomes; 49 and 51 core-T4SEs are identified 5 B. canis and 12 B. suis strains, respectively. Overall, 25 T4SEs are conserved among all the 33 genomes of Brucella.

Fig. 4 Venn diagram showing core-T4SEs predicted from B. abortus, B. melitensis, B. canis and B. suis genomes. Number denotes the core-T4SEs of individual species and the subset of core-T4SEs shared among different Brucella spp.

Prediction of the host–pathogen protein–protein interaction (PPI) network

We predicted the interactions of B. abortus T4SEs with Bos taurus (its preferred host) proteome. We have identified 100 PPIs with the involvement of 24 T4SEs of B. abortus and 93 B. taurus proteins (Fig. 5). Brucella invades the host cell and forms a plasma membrane-derived BCV. Brucella takes a circuitous route through the endoplasmic reticulum to avoid lysosomal compartments. Brucella survives within BCV and eventually proliferates.⁵⁰Brucella replicative organelles are characterized by the presence of endoplasmic reticulum markers on the vacuolar membrane and in the absence of lysosomal-associated membrane protein 1 (LAMP-1), which is temporarily acquired by BCVs during maturation.²³ Furthermore, maturation of BCVs is characterized by the exclusion of LAMP-1 from the vacuolar membrane.⁵⁰

Fig. 5 Protein–protein interaction network between B. abortus T4SEs and B. taurus proteome. Blue and red circles represent T4SEs of B. abortus. Blue circles represent T4SEs found in both B. abortus and B. melitensis. Red circles represent B. abortus specific T4SEs. Green and pink circles represent host proteins. Green circles are proteins found in both B. taurus and O. aries. Pink circles represent B. taurus specific host proteins. ER – endoplasmic reticulum, LS – lysosome, ES – endosome, GC – golgi-complex, MT – mitochondria, NU – nucleus.

Prediction of cellular localization of B. taurus proteins identified in PPIs

The targeted proteins are located in cellular compartments that are relevant to the pathogenesis mechanism. The predicted localizations of B. taurus PPI interacting proteins were identified in the endosome, lysosome, endoplasmic reticulum, golgi-complex, mitochondria, nucleus, plasma membrane and cytoplasm (Fig. 6). A carbamoyl-phosphate synthetase ammonia chain (CarB) interacts with the nuclear factor of kappa B-cells (NF-κB) that involved in cell survival and proliferation, inflammatory response, and anti-apoptotic factors. STAT (Signal Transducer and Activator of Transcription) proteins are involved in the development and function of the immune system and play a role in maintaining immune tolerance, which interacts with acetyl-CoA hydrolase. The STAT protein regulates many aspects of growth, survival and differentiation in cells. A carbamoyl phosphate synthase small subunit interacts with the S100 calcium binding protein A9 (S100A9), which plays a prominent role in the regulation of inflammatory processes and immune response. The PAS domain-containing protein interacts with the endoplasmic reticulum resident protein 29 (ERP29) and the cytoplasmic polyadenylation element-binding protein 2(CPEB3). ERP29 plays an important role in the processing of secretory proteins. CPEB3 is involved in the regulation of mRNA translation and cell proliferation. The Peptidyl-tRNA hydrolase domain-containing protein interacts with the mitochondrial protein 28S ribosomal protein S29 (MRPS29) that involved in mediating interferon-gamma-induced cell death. Patatin interacts with the cell division control protein 42 (CDC42) that regulates cellular responses. Patatin also interacts with the ras homolog family Member J (RhoJ) protein involved in the regulation of cell morphology. The RNA polymerase sigma factor RpoD interacts with the conserved oligomeric golgi complex subunit 3 (COG3) and the tripartite motif containing 27. COG3 required for golgi morphology and localization, also involved in endoplasmic reticulum to golgi transport. The tripartite motif containing 27 (TRIM27) is involved in the differentiation of germ cells. Acetyl-CoA hydrolase interact with centriolin (CNTRL), which involved in cell progression and cytokinesis activity. The carbamoyl phosphate synthase small subunit carA interacting with carnitine palmitoyltransferase 1A (CPT1A) appears to increase the risk of a serious liver disorder that can develop during pregnancy. The Sel1 repeat containing protein interacts with the small ubiquitin-like modifier (SUMO) that involved in cellular processes, such as nuclear-cytosolic transport, transcriptional regulation, apoptosis. D-Alanyl-D-alanine carboxypeptidase interacts with olfactomedin-4 (OLFM4) and inhibits cell growth and induces cell differentiation and apoptosis. However, the roles of these PPIs in the pathogenesis of Brucella should be studied using suitable in vitro and in vivo experiments.

Fig. 6 Cellular localization of B. taurus host proteins interacting with B. abortus T4SEs. Proteins localized on the respective organelles are shown and all other proteins are localized in the cytoplasm.

Differentiation of host proteins and T4SE proteins of B. abortus and B. melitensis using PPIs

We also predicted the interactions of B. melitensis T4SEs with Ovis aries (its preferred host) proteome to differentiate the host specificity of B. abortus and B. melitensis. We identified 104 PPIs with the involvement of 25 B. melitensis T4SEs and 99 O. aries proteins (ESI,† file 5 and Fig. S1). With respect to the pathogen, 21 T4SEs are conserved in both B. abortus and B. melitensis, which showed potential PPI with respective host proteome. Three T4SEs, such as the cobalamin synthesis protein P474 (BAB1_0246), deoxyribonucleotide triphosphate pyrophosphatase (BAB1_0175) and PAS domain-containing proteins (BAB1_0640) are B. abortus specific. Similarly, T4SEs such as the pyrroline-5-carboxylate reductase (G4PlR1), multidrug resistance protein (G4PJZ8), protein translocase subunit SecA (G4PHM6), Ribose transport system permease protein rbsc (G4PHB5) are B. melitensis specific.

With respect to host proteins, 58 proteins showing potential PPI with respective pathogens were found in both host species (B. taurus and O. aries). Of these, proteins such as cell division cycle 42 and nuclear factor of kappa light polypeptide gene enhancer in B-cells 1 (NFkB) are involved in the MAPK signalling pathway associated with intracellular signalling. Collagen, type IV, and alpha 2 (col4a2) are involved in the adhesion pathway. Eukaryotic translation initiation factor 4B (EIF4B) is involved in the mTOR signalling pathway that helps in translation initiation factor activity. The protein involved in the Jak-STAT signalling pathway includes the signal transducer and activator of transcription 6, interleukin-4 induced (STAT6).

In addition, 29 interacting proteins were identified as specific to B. taurus showing interactions with B. abortus (ESI,† file 5 and Table S2). The B. taurus–B. abortus specific interaction proteins include the ABI family protein, member 3 (NESH) binding protein, HCLS1 associated protein X-1, RAN binding protein 2, cytoplasmic polyadenylation element binding protein 2 and endoplasmic reticulum protein 29, which are involved in regulation of the biological and cellular process. Likewise, 36 interacting proteins were specific to O. aries showing interaction with B. melitensis (ESI,† file 5 and Table S2). The Ras-like protein TC10 identified as specific to O. aries, which is involved in the intracellular signalling pathway. Similarly, amyloid beta (A4) precursor protein-binding, family A, member 1 and 3, which are involved in cell–cell signalling and communications were O. aries specific.

Discussion

We have identified the genes coding for the conserved Sec and Tat pathways of secretion in Brucella. The Sec pathway functions as the major protein transport machinery in bacteria. The Sec pathway has two functions; it transports secretory proteins through the inner membrane and inserts membrane proteins into the inner membrane. The Sec translocase of Brucella is composed of a membrane-embedded complex including SecY, SecE, SecG, SecD/F, YajC and YidC. The cytoplasmic SecA subunit hydrolyzes ATP to drive the protein through the translocation process.⁵¹ The signal sequence is recognized by the signal recognition particle, which targets the ribosome to translocase via the FtsY receptor. Additional secreted proteins are recognized through the ffh and SecB chaperone after the completion of translation, which targets the proteins to the translocase by binding to SecA.⁵¹ The Tat pathway transports folded proteins across the inner membrane. Tat substrates often form co-factor containing proteins and the insertions of these co-factors were restricted to the cytosol.⁵² Based on the prediction of interactions among the secretion systems, we have identified that both Sec and Tat pathway proteins are interacting with T4SS and T5SS proteins.

The Brucella virB T4SS is of crucial importance during macrophage infection, required for internalization by macropinocytosis and for fusion with the ER to establish the replicative niche.²³ The secreted T4SEs mediated by the virB T4SS plays major roles in Brucella pathogenesis.⁵³Brucella virB mutants showed attenuated phenotype in mouse models of infection.⁵⁴ The T4SEs required for proper trafficking within the cell, interaction with the ER and survival within a host.⁵⁵ The genus Brucella does not contain any plasmid and therefore it is probable that T4SS involved in protein secretion rather than conjugation.⁵³ The virB operon is conserved in all the sequenced Brucella species. virB operon was first identified in B. suis from the TnblaM mutant during an early scan for virulence factors.²² Later, the role of the B. abortus virB system in virulence was reported.⁵⁶ Previous studies described that virB1 and virB7 are not essential for the survival of Brucella.⁵⁷VceA and VceC were previously identified as T4SEs using TEM1 β-lactamase fusion assays⁵⁸ and these proteins are co-regulated with the virB genes. Translocation of VceC induces ER stress, which results in the induction of proinflammatory host cell responses during B. abortus infection.⁵⁹RicA is another reported T4SE identified by high-throughput yeast two hybrid screening.⁶⁰RicA (Rab2 interacting conserved protein A) bind to GDP-bound form of Rab2,⁶⁰ a host GTPase reported to be essential for the intracellular replication of Brucella.⁶¹ The other reported T4SE of Brucella is Btp1/TcpB containing a Toll/interleukin-1 receptor (TIR) domain that interferes with TLR2 and TLR4 signaling.⁶²

Proteins with eukaryotic like domains might be acquired by interdomain horizontal gene transfer from eukaryotes to bacteria.⁶³ Marchesini et al.⁶⁴ have identified six T4SEs through computational approaches. Among these proteins, four BEPs viz., BAB1_1043, BAB1_2005, BAB1_1275 and BAB2_0123 require a functional T4SS for their delivery into host macrophage-like cells upon infection.⁶⁴ These proteins were identified through bioinformatics approaches like homology, eukaryotic like domain, proteins related to virulence and coiled-coil. Myeni et al.⁵³ have reported VirB T4SS effector proteins using bioinformatics approaches like homology with other bacterial organisms, GC content, and the presence of eukaryotic like-motif. They have predicted 11 T4SEs, in which 10 of them were claimed as novel proteins. Of these, three T4SEs, viz., BspA, BspB and BspF predicted to inhibit host protein secretion and promoted pathogen intracellular growth and persistence in the liver of infected mice.⁵³

SEL1 domains belong to the tetratricopeptide repeat (TPR) protein family. TPR-containing proteins were found in eukaryotic organisms and reported to direct a variety of protein–protein interactions,⁴⁰ making them candidate bacterial effectors that potentially interacts with host proteins following delivery to the cytosol. SEL1 repeats are a subfamily of TPR proteins found largely in eukaryotic organisms and regulate the turnover of endoplasmic reticulum proteins.⁶⁵ The trigger factor plays a critical role in the acute phase of Brucella infection.⁶⁶Crp, GntR, LysM and TetR families of transcriptional regulator proteins are involved in virulence and symbiosis.⁶⁷

Some of the identified T4SEs in this study have been earlier reported in Brucella spp. Wang et al.^68,69 have developed the T4SS mutant strain of B. melitensis under acidic pH and observed changes in the expression level of Omp25 and Omp31.^68,69 Paredes-Cervantes et al.⁷⁰ have observed a T4SS dependent increase in the expression of Omp28, as well as a small decrease in the expression of Omp25 when comparing the mutant strain and the wild-type strain grown under acidic pH.⁷⁰ In the present study, we have predicted that Omp25 and Omp31 are secreted by all Brucella spp. Peptidyl-prolyl cis–trans isomerase and alkyl hydroperoxide reductase D are other proteins identified as secreted proteins, which were earlier reported by Delpino et al.⁷¹ Rossetti et al.⁷² observed that cultured B. melitensis is the most infectious during the late log phase, which is concurrent with the overexpression of T4SS related genes, including gene coding for outer membrane proteins, lipoproteins, amino acid and carbohydrate-specific transporters, nonspecific transporters and genes associated with Fe transport and metabolism.⁷²

We also predicted protein–protein interactions between T4SEs of Brucella and respective host proteome. Approximately, 100 PPIs were predicted in both pathogen–host systems. Overall, the predicted host proteins showing PPI with respective pathogens are involved in DNA binding, RNA binding, ion binding, signal transduction activity, transcriptional factor binding, etc. The host–pathogen PPIs were located in very relevant cellular compartments. These resemble the pathogen infection and invasion of host cells. The predicted localizations include the intracellular region, cytoplasm, cytoskeleton, endoplasmic reticulum, plasma membrane and mitochondria. The host interacting proteins are involved in pathways such as the MAPK signalling pathway, Toll-like receptor signalling pathway, apoptosis, Jak-STAT signalling pathway and T cell. Some PPIs were commonly found in both pathogen–host systems. In addition, a subset of T4SEs and proteins were found to be specific to the respective pathogen and host (B. abortus–B. taurus and B. melitensis–O. aries).

Conclusions

Secretion systems of Brucella spp. were predicted by in silico methods from the available complete genomes. To identify host cell interactions, we screened the genome of Brucella proteins that would make them good candidates for translocation, such as eukaryotic-like domains, coiled-coil, nuclear localization signal, mitochondrial localization signal, homology to known effectors in related species and structural features known to be involved in protein–protein interactions. We have identified 80 T4SEs in B. abortus with at least one of the above mentioned properties. Some of them were annotated as hypothetical proteins without any predicted function. Prediction of protein–protein interactions between the host proteome and Brucella T4SEs revealed few species specific T4SEs and their respective host target proteins (B. abortus–B. taurus and B. melitensis–O. aries). Further in vitro and in vivo experiments are required to confirm these predicted networks. However, this is the first study throwing some light on Brucella and host species specific interactions.

Availability of datasets

The predicted datasets of the secretion system, secreted proteins and T4SE for 39 complete genomes are available in the URL: http://www.dbtbrucellosis.in/BrucellaSecretionSystem.html.

Acknowledgements

The work was financially supported by the Department of Biotechnology, Govt. of India under DBT-Network Project on Brucellosis. The authors also acknowledge the UGC-CAS, NRCBS, DBT-IPLS, and DST-PURSE Programs of the School of Biological Sciences, Madurai Kamaraj University. The authors also acknowledge Dr McCarthy FM, Department of Basic Sciences, College of Veterinary Medicine, Mississippi State University, USA, for helping in HPIDB analysis.

References

1. J. P. Gorvel and E. Moreno, Vet. Microbiol., 2002, 90, 281–297.
2. M. J. Corbel, Emerging Infect. Dis., 1997, 3, 213–221.
3. J. P. Gorvel, Microbes Infect., 2008, 10, 1010–1013.
4. P. A. Lee, D. Tullman-Ercek and G. Georgiou, Annu. Rev. Microbiol., 2006, 60, 373–395.
5. P. Delepelaire, Biochim. Biophys. Acta, Mol. Cell Res., 2004, 1694, 149–161.
6. B. Arellano-Reynoso, N. Lapaque, S. Salcedo, G. Briones, A. E. Ciocchini, R. Ugalde, E. Moreno, I. Moriyón and J.-P. Gorvel, Nat. Immunol., 2005, 6, 618–625.
7. N. P. Cianciotto, Trends Microbiol., 2005, 13, 581–588.
8. G. R. Cornelis, Nat. Rev. Microbiol., 2006, 4, 811–825.
9. C. E. Alvarez-Martinez and P. J. Christie, Microbiol. Mol. Biol. Rev., 2009, 73, 775–808.
10. Z. Ding, K. Atmakuri and P. J. Christie, Trends Microbiol., 2003, 11, 527–535.
11. I. R. Henderson, F. Navarro-Garcia, M. Desvaux, R. C. Fernandez and D. Ala'Aldeen, Microbiol. Mol. Biol. Rev., 2004, 68, 692–744.
12. L. Bingle, C. Bailey and M. Pallen, Curr. Opin. Microbiol., 2008, 11, 3–8.
13. T. N. Petersen, S. Brunak, G. von Heijne and H. Nielsen, Nat. Methods, 2011, 8, 785–786.
14. J. D. Bendtsen, L. Kiemer, A. Fausbøll and S. Brunak, BMC Microbiol., 2005, 5, 58.
15. Y.-C. Zuo and Q.-Z. Li, Peptides, 2009, 30, 1788–1793.
16. R. Verma, A. Tiwari, S. Kaur, G. C. Varshney and G. P. Raghava, BMC Bioinf., 2008, 9, 201.
17. Y.-C. Zuo and Q.-Z. Li, Amino Acids, 2010, 38, 859–867.
18. W.-Z. Lin, J.-A. Fang, X. Xiao and K.-C. Chou, PLoS One, 2012, 7, e49040.
19. M. Zhou, D. Theunissen, M. Wels and R. J. Siezen, BMC Genomics, 2010, 11, 651.
20. M. A. Jehl, R. Arnold and T. Rattei, Nucleic Acids Res., 2011, 39, 591–595.
21. D. F. Meyer, C. Noroy, A. Moumène, S. Raffaele, E. Albina and N. Vachiéry, Nucleic Acids Res., 2013, 41, 9218–9229.
22. D. O'Cellaghan, C. Cazevieille, A. Allardet-Servent, M. L. Boschiroli, G. Bourg, V. Foulongne, P. Frutos, Y. Kulakov and M. Ramuz, Mol. Microbiol., 1999, 33, 1210–1220.
23. J. Celli, C. de Chastellier, D.-M. Franchini, J. Pizarro-Cerda, E. Moreno and J.-P. Gorvel, J. Exp. Med., 2003, 198, 545–556.
24. P. de Figueiredo, T. A. Ficht, A. Rice-Ficht, C. A. Rossetti and L. G. Adams, Am. J. Pathol., 2015, 185, 1505–1517.
25. H. Ogata, S. Goto, K. Sato, W. Fujibuchi, H. Bono and M. Kanehisa, Nucleic Acids Res., 1999, 27, 29–34.
26. D. Bi, L. Liu, C. Tai, Z. Deng, K. Rajakumar and H. Y. Ou, Nucleic Acids Res., 2013, 41, D660–D665.
27. A. Franceschini, D. Szklarczyk, S. Frankild, M. Kuhn, M. Simonovic, A. Roth, J. Lin, P. Minguez, P. Bork, C. Von Mering and L. J. Jensen, Nucleic Acids Res., 2013, 41, D808–D815.
28. K. Hiller, A. Grote, M. Maneck, R. Münch and D. Jahn, Bioinformatics, 2006, 22, 2441–2443.
29. M. Lechner, S. Findeiss, L. Steiner, M. Marz, P. F. Stadler and S. J. Prohaska, BMC Bioinf., 2011, 12, 124.
30. F. Chen, A. J. Mackey, C. J. Stoeckert and D. S. Roos, Nucleic Acids Res., 2006, 34, D363–D368.
31. L. Li, Jr., C. J. Stoeckert and D. S. Roos, Genome Res., 2003, 13, 2178–2189.
32. R. Kumar and B. Nanduri, BMC Bioinf., 2010, 11(suppl 6), S16.
33. R. Saito, M. E. Smoot, K. Ono, J. Ruscheinski, P. Wang, S. Lotia, A. R. Pico, G. D. Bader and T. Ideker, Nat. Methods, 2012, 9, 1069–1076.
34. T. H. Chang, L. C. Wu, T. Y. Lee, S. P. Chen, H. Da Huang and J. T. Horng, J. Comput.-Aided Mol. Des., 2013, 27, 91–103.
35. C.-S. Yu, C.-W. Cheng, W.-C. Su, K.-C. Chang, S.-W. Huang, J.-K. Hwang and C.-H. Lu, PLoS One, 2014, 9, e99368.
36. F. M. McCarthy, C. R. Gresham, T. J. Buza, P. Chouvarine, L. R. Pillai, R. Kumar, S. Ozkan, H. Wang, P. Manda, T. Arick, S. M. Bridges and S. C. Burgess, Nucleic Acids Res., 2011, 39, D497–D506.
37. X. Li, J. Xu, Y. Xie, Y. Qiu, S. Fu, X. Yuan, Y. Ke, S. Yu, X. Du, M. Cui, Y. Chen, T. Wang, Z. Wang, Y. Yu, K. Huang, L. Huang, G. Peng, Z. Chen and Y. Wang, Vet. Microbiol., 2012, 161, 137–144.
38. C. Höppner, A. Carle, D. Sivanesan, S. Hoeppner and C. Baron, Microbiology, 2005, 151, 3469–3482.
39. J. C. Leo, I. Grin and D. Linke, Philos. Trans. R. Soc., B, 2012, 367, 1088–1101.
40. A. Lupas, M. Van Dyke and J. Stock, Science, 1991, 252, 1162–1164.
41. D. De Biase and E. Pennacchietti, Mol. Microbiol., 2012, 86, 770–786.
42. E. R. Moore, D. J. Mead, C. A. Dooley, J. Sager and T. Hackstadt, Microbiology, 2011, 157, 830–838.
43. L. D. D'Andrea and L. Regan, Trends Biochem. Sci., 2003, 28, 655–662.
44. A. Walburger, C. Lazdunski and Y. Corda, Mol. Microbiol., 2002, 44, 695–708.
45. P. B. Daniel, W. H. Walker and J. F. Habener, Annu. Rev. Nutr., 1998, 18, 353–383.
46. C. A. Worby, S. Mattoo, R. P. Kruger, L. B. Corbeil, A. Koller, J. C. Mendez, B. Zekarias, C. Lazar and E. Jack, Mol. Cell, 2010, 34, 1–23.
47. L. N. Kinch, M. L. Yarbrough, K. Orth and N. V Grishin, PLoS One, 2009, 4, e5818.
48. S. Banerji and A. Flieger, Microbiology, 2004, 150, 522–525.
49. F. Dyda, D. C. Klein and A. B. Hickman, Annu. Rev. Biophys. Biomol. Struct., 2000, 29, 81–103.
50. G. Gao and J. Xu, Crit. Rev. Eukaryotic Gene Expression, 2013, 23, 65–76.
51. H. Mori and K. Ito, Trends Microbiol., 2001, 9, 494–500.
52. R. Kudva, K. Denks, P. Kuhn, A. Vogt, M. Müller and H. G. Koch, Res. Microbiol., 2013, 164, 505–534.
53. S. Myeni, R. Child, T. W. Ng, J. J. Kupko, T. D. Wehrly, S. F. Porcella, L. A. Knodler and J. Celli, PLoS Pathog., 2013, 9, e1003556.
54. M. L. Boschiroli, S. Ouahrani-Bettache, V. Foulongne, S. Michaux-Charachon, G. Bourg, A. Allardet-Servent, C. Cazevieille, J. P. Lavigne, J. P. Liautard, M. Ramuz and D. O'Callaghan, Vet. Microbiol., 2002, 90, 341–348.
55. A. A. Rambow-Larsen, E. M. Petersen, C. R. Gourley and G. A. Splitter, Trends Microbiol., 2009, 17, 371–377.
56. R. Sieira, D. J. Comerci, D. O. Sanchez and R. A. Ugalde, J. Bacteriol., 2000, 182, 4849–4855.
57. A. B. Den Hartigh, H. G. Rolán, M. F. De Jong and R. M. Tsolis, J. Bacteriol., 2008, 190, 4427–4436.
58. M. F. De Jong, Y. H. Sun, A. B. Den Hartigh, J. M. Van Dijl and R. M. Tsolis, Mol. Microbiol., 2008, 70, 1378–1396.
59. M. F. de Jong, T. Starr, M. G. Winter, A. B. den Hartigh, R. Child, L. A. Knodler, J. M. van Dijl, J. Celli and R. M. Tsolis, mBio, 2013, 4, e00418-12.
60. M. De Barsy, A. Jamet, D. Filopon, C. Nicolas, G. Laloux, J. F. Rual, A. Muller, J. C. Twizere, B. Nkengfac, J. Vandenhaute, D. E. Hill, S. P. Salcedo, J. P. Gorvel, J. J. Letesson and X. De Bolle, Cell. Microbiol., 2011, 13, 1044–1058.
61. E. Fugier, S. P. Salcedo, C. De Chastellier, M. Pophillat, A. Muller, V. Arce-Gorvel, P. Fourquet and J. P. Gorvel, PLoS Pathog., 2009, 5, e1000487.
62. G. K. Radhakrishnan, Q. Yu, J. S. Harms and G. A. Splitter, J. Biol. Chem., 2009, 284, 9892–9898.
63. J. C. Amor, J. Swails, X. Zhu, C. R. Roy, H. Nagai, A. Ingmundson, X. Cheng and R. A. Kahn, J. Biol. Chem., 2005, 280, 1392–1400.
64. M. I. Marchesini, C. K. Herrmann, S. P. Salcedo, J. P. Gorvel and D. J. Comerci, Cell. Microbiol., 2011, 13, 1261–1274.
65. P. R. E. Mittl and W. Schneider-Brachert, Cell. Signalling, 2007, 19, 20–31.
66. M. S. Roset, L. G. Fernández, V. G. DelVecchio and G. Briones, Infect. Immun., 2013, 81, 521–530.
67. V. Haine, A. Sinon, F. Van Steen, S. Rousseau, M. Dozot, P. Lestrate, C. Lambert, J.-J. Letesson and X. De Bolle, Infect. Immun., 2005, 73, 5578–5586.
68. Y. Wang, Z. Chen, F. Qiao, Z. Zhong, J. Xu, Z. Wang, X. Du, Q. Qu, J. Yuan, L. Jia, H. Song, Y. Sun and L. Huang, FEMS Microbiol. Lett., 2010, 303, 92–100.
69. Y. Wang, Z. Chen, F. Qiao, T. Ying, J. Yuan, Z. Zhong, L. Zhou, X. Du, Z. Wang, J. Zhao, S. Dong, L. Jia, X. Yuan, R. Yang, Y. Sun and L. Huang, PLoS One, 2009, 4, e5368.
70. V. Paredes-Cervantes, R. Flores-Mejía, M. C. Moreno-Lafont, H. Lanz-Mendoza, Á. T. Tello-López, J. Castillo-Vera, V. Pando-Robles, G. Hurtado-Sil, E. González-González, O. Rodríguez-Cortés, A. Gutiérrez-Hoya, M. T. Vega-Ramírez and R. López-Santiago, J. Proteomics, 2011, 74, 2959–2971.
71. M. V. Delpino, D. J. Comerci, M. A. Wagner, M. Eschenbrenner, C. V Mujer, R. A. Ugalde, C. A. Fossati, P. C. Baldi and V. G. DelVecchio, Arch. Microbiol., 2009, 191, 571–581.
72. C. A. Rossetti, C. L. Galindo, S. D. Lawhon, H. R. Garner and L. G. Adams, BMC Microbiol., 2009, 9, 81.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5mb00607d

---