Comparative genomic analysis of Brucella abortus vaccine strain 104M reveals a set of candidate genes associated with its virulence attenuation

Abstract

The Brucella abortus strain 104M, a spontaneously attenuated strain, has been used as a vaccine strain in humans against brucellosis for 6 decades in China. Despite many studies, the molecular mechanisms that cause the attenuation are still unclear. Here, we determined the whole-genome sequence of 104M and conducted a comprehensive comparative analysis against the whole genome sequences of the virulent strain, A13334, and other reference strains. This analysis revealed a highly similar genome structure between 104M and A13334. The further comparative genomic analysis between 104M and A13334 revealed a set of genes missing in 104M. Some of these genes were identified to be directly or indirectly associated with virulence. Similarly, a set of mutations in the virulence-related genes was also identified, which may be related to virulence alteration. This study provides a set of candidate genes associated with virulence attenuation in B.abortus vaccine strain 104M.

Introduction

Brucellae are Gram-negative, facultative intracellular bacteria that can cause brucellosis in many animals and humans.¹ Brucellosis is among the most common zoonotic infectious disease epidemics worldwide, leading to great economic and public health problems,² particularly in developing countries. Human brucellosis is transmitted through direct animal contact and the consumption of contaminated food products of animal origin. Over the past 100 years, human brucellosis has been controlled by vaccination and culling animals,^2-4 but more than 500,000 new cases are reported annually worldwide.⁵ China is a major stock-raising country. Over the last decade, human brucellosis has increased quickly, and human cases have been reported in all provinces.⁶ This disease causes chronic infections with common animal outcomes of abortion and sterility. In humans, it can cause a severe acute fever and a febrile illness if untreated, producing focal lesions in joints, the genitourinary tract, and other organs.⁷

The most rational approach for preventing and controlling brucellosis is vaccination. Some live, attenuated vaccine strains are useful for the control and elimination of animal infections to decrease the rate of human brucellosis, such as Brucella abortus S19^8,9 and RB51,^10,11 which are used in cattle, and Brucella melitensis Rev1¹² and M5,¹³ which are used in sheep and goats.

In addition to these vaccine strains, Brucella abortus (B. abortus) 104M ¹⁴ is a vaccine strain that is used in cattle and is widely applied for brucellosis prevention in China. This strain was first isolated from a sick cow's placenta in a former Soviet republic. It exhibits typical properties of biotype I, low and stable virulence on experimental animals, and a strong immunogenicity.^15-18 In certain conditions, the strain 104M was used as a live attenuated vaccine for humans. After 1965, the Chinese FDA approved the live, attenuated strain of 104M as a vaccine for use in human. Though these vaccine strains play an irreplaceable role in the control and prevention of brucellosis, the underlying molecular and physiological mechanisms causing attenuated virulence are unclear. Recently, several comparative genomic analyses have been conducted to detect the factors affecting virulence in Brucella vaccine strains, such as S19¹⁹ and M5,¹³ but none were related to 104M.

We sequenced the Brucella abortus vaccine strain 104M genome with the goal of elucidating the genetic mechanisms underlying virulence attenuation. As a result, a set of genes, which were associated with virulence or virulence attenuation, was identified through the comparison of this newly sequenced genome with its virulent counterparts.

Results and Discussion

General genome features

The complete genome of the vaccine strain, 104M, was sequenced as described in the Material and Methods, and 2 complete chromosome sequences were obtained: one is 2122848 bp long, and the other is 1162581 bp long. The average GC content of the 2 chromosomes is 57.2%. This newly sequenced genome was annotated using the RAST server,²⁰ and 3192 protein-encoding genes, 65 RNA genes, including 10 rRNAs and 55 tRNAs, were identified. The detailed genome properties and 3 other Brucella abortus genomes are listed in Table 1. The genome of strain 104M showed remarkable similarity in various properties to those of its relatives, B. abortus A13334, S19, 9–941 and 2308.

Table 1.

Genome properties of the newly sequenced B. abortus vaccine strain 104M genome in comparison with the known genome sequences of 3 strains

	B. abortus 104M		B. abortus A13334		B. abortus S19		B. abortus 9–941		B. abortus 2308
Feature	Chr1	Chr2	Chr1	Chr2	Chr1	Chr2	Chr1	Chr2	Chr1	Chr2
Size (Mbp)	2.12	1.16	2.12	1.16	2.12	1.16	2.12	1.16	2.12	1.16
ORFs	2098	1094	2185	1153	2005	1057	2030	1055	2000	1182
tRNA	40	15	40	14	41	14	41	14	41	14
rRNA	7	3	6	3	6	3	6	3	6	3
GC(%)	57.2	57.3	57.2	57.3	57.2	57.3	57.2	57.3	57.2	57.3
Average length	859.5	917.8	830.9	883.5	890.4	942.6	845.7	900	852	904.2

A multiple genomes alignment was constructed among 104M and the other sequenced Brucella genomes to detect this newly sequenced strain's phylogenetic relationships. A maximum likelihood phylogenetic analysis of the 18 strains using SNP data was conducted, as shown in Figure 1. All Brucella strains were divided into 5 clades, as previously reported:²¹ B. melitensis-B. abortus clade; B. ovis clade; B. pinnipedialis clade; B. microti clade; and B. suis-B. canis clade. Each internal node received 100% bootstrap support. In the tree, strain 104M was located in the B. abortus sub-clade and diverged earlier than other B. abortus strains. Despite an evolutionary branching order, strain 104M appeared to be very close to other B. abortus strains for the short internal branch length.

Figure 1.

Phylogenetic tree of 18 Brucella strains. The maximum likelihood tree was constructed using SNP data. Almost all nodes received 100% bootstrap support and were labeled on each node.

A genome-wide comparison was conducted among 104M and its relatives, B. abortus A13334, 9–941, S19 and 2308, to detect the whole-genome similarities. The result is showed in Figure 2: 104M exhibits a perfect genome-wide colinearity with other B.abortus strains. The only difference is the location of blue block in each genome, which could be due to the different starting loci of genome sequencing. Thus, further comparisons in gene content between 104M and other strains were needed. The results, including the number of strain-specific genes, were shown in Figure 3. It is clear to see that the smallest differences lie in the comparison of 104M and A13334.

Figure 2.

Mauve alignment of both chromosomes from the 5 complete Brucella abortus genomes. Each genome is laid out horizontally and homologous segments are shown as colored blocks: red and olivine blocks constitute the sequence of chromosome I, and blue and green blocks constitute the sequence of chromosome II.

Figure 3.

The Venn diagrams showing the distribution of strain-specific genes between the genomes of 104M and other B.abortus strains.

The virulent strain A13334 is derived from the foetal gastric fluid of an infected dairy cow in Gyeonggi, Yangpyeong, Republic of Korea,²² which is near China. Considering the high similarity in genome and close distance in geography, we conducted the comparative genomic analysis between 104M and A13334 to detect the virulence or virulence-attenuated associated genes.

Identification of virulence-associated gene missing between 104M and A13334

Although the genome alignment results showed high similarity between the genomes of 104M and A13334, there are genetic differences leading to the attenuated virulence in 104M. The main target of this work was to identify all of the genetic differences between 104M and its virulent relative A13334, which would provide a complete basis for the virulence attenuation. Pairwise and reciprocal comparison was conducted between the 2 proteomes to identify genes that are 100% identical, non-identical with any differences and unique in either of the strains. A total of 99 and 52 genes, as showed in Figure 3, were identified to be unique in A13334 and 104M, respectively. The remaining genes in either of the genomes had their corresponding homologs in the other genome.

Gene loss associated with virulence

The unique genes in either of the genomes might imply the loss of the same genes in the other genome, suggesting a set of potential targets leading to the 2 strain's virulence differences. In addition to deletion and horizontal gene transfer (HGT), the other possible reasons for the existence of the unique genes could be de novo gene formation or frameshift mutation. However, the unique genes in either genome did not show a continuous distribution along the genome. Considering the close evolutionary relationship between the 2 strains, they may not be the result of repeated gene gain, gene loss or de novo gene formation events. Additionally, the gene sequences were used to perform a genes-to-genome search, and nearly all of the unique genes were identical to those in the corresponding genome. Thus, the unique genes might be the results of frameshift mutations.

However, 62 strain-specific genes in A13334 and 45 strain-specific genes in 104M were annotated as hypothetical. For the remaining genes with definite functional assignments, literature searches were conducted to determine whether the genes were related to virulence.^23,24

For the A13334-specific genes, the detailed functional annotation for each gene was listed in Table 2. Among the 37 A13334-specific genes with definite function assignment, more than half encode various enzymes that are involved in the fundamental life cycle, such as ATP-dependent ligase, diguanylate cyclase and fumarase. Among them, several genes might be associated with virulence.

Table 2.

A list of genes missing in the attenuated strain 104M compared with the virulent strain, A13334

Gene	Function	KO	Pathway
BAA13334_I00134	ABC transporter	K06147	ATP-binding cassette
BAA13334_II01500*	ABC transporter	K02032	Peptides/nickel transport system
BAA13334_II00879	ATP-dependent DNA ligase
BAA13334_I02291	Diguanylate cyclase
BAA13334_I01435	Fe-S protein
BAA13334_II00407*	Flagellar motor switch protein G	K02410	Bacterial chemotaxis
	Flagellar assembly
BAA13334_I01019	Fumarase
BAA13334_I01734*	Gluconolactonase
BAA13334_I01361*	Glucose-methanol-choline oxidoreductase	K00108	Glycine, serine and threonine metabolism
BAA13334_II00868*	Glyceraldehyde-3-phosphate dehydrogenase/erythrose-4-phosphate dehydrogenase
BAA13334_II01395	Glycosyl hydrolase family protein
BAA13334_I03093	Integrase, catalytic domain-containing protein
BAA13334_II01601	Leucyl-tRNA synthetase
BAA13334_I03416*	LuxR family 2 component transcriptional regulator
BAA13334_I03052	Major facilitator superfamily transporter	K03449	MFS transporter
BAA13334_I02644	Membrane protein related to metalloendopeptidase
BAA13334_I00309	Membrane-associated Zn-dependent proteases 1
BAA13334_I01622	Mg2+ and Co2+ transporter
BAA13334_I01425*	Myosin-14
BAA13334_I01865	N-acetylglucosamine-6-phosphate deacetylase
BAA13334_I01533*	N-acetyltransferase GCN5
BAA13334_II01619*	Nitrite transporter	K02575	Nitrogen metabolism
BAA13334_I01778*	NTP pyrophosphohydrolase
BAA13334_II00969*	Outer membrane receptor protein
BAA13334_I00149	Phosphatidylglycerophosphate synthase
BAA13334_I00176*	Phosphoadenosine phosphosulfate reductase	K00390	Sulfur metabolism
BAA13334_I02247	Prophage antirepressor
BAA13334_I00833*	Radical SAM domain-containing protein	K10026	Folate biosynthesis
BAA13334_II00036	Retron-type reverse transcriptase
BAA13334_II01216*	Ribonuclease R	K12573	RNA degradation
BAA13334_I02254*	Site-specific recombinase, phage integrase family protein
BAA13334_I01159*	Sulfate permease
BAA13334_I03390*	Superoxide dismutase
BAA13334_II01445	Tellurite resistance protein-related permease
BAA13334_I02277	UDP-glucose 4-epimerase
BAA13334_I01801*	Urease accessory protein UreF	K03188	urease accessory protein
BAA13334_II00218*	WD40 repeat protein
BAA13334_I00025	Hypothetical protein
BAA13334_I00054	Hypothetical protein
BAA13334_I00130	Hypothetical protein
BAA13334_I00328	Hypothetical protein
BAA13334_I00441	Hypothetical protein
BAA13334_I00661	Hypothetical protein
BAA13334_I00777	Hypothetical protein
BAA13334_I00781	Hypothetical protein
BAA13334_I00883*	Hypothetical protein
BAA13334_I00994*	Hypothetical protein
BAA13334_I01009	Hypothetical protein
BAA13334_I01155	Hypothetical protein
BAA13334_I01185	Hypothetical protein
BAA13334_I01224	Hypothetical protein
BAA13334_I01234	Hypothetical protein
BAA13334_I01281	Hypothetical protein
BAA13334_I01332	Hypothetical protein
BAA13334_I01544	Hypothetical protein
BAA13334_I01558	Hypothetical protein
BAA13334_I01585	Hypothetical protein
BAA13334_I01834	Hypothetical protein
BAA13334_I01840	Hypothetical protein
BAA13334_I01849	Hypothetical protein
BAA13334_I01922	Hypothetical protein
BAA13334_I01965	Hypothetical protein
BAA13334_I01967*	Hypothetical protein
BAA13334_I01998	Hypothetical protein
BAA13334_I02213*	Hypothetical protein
BAA13334_I02339	Hypothetical protein
BAA13334_I02438	Hypothetical protein
BAA13334_I02447	Hypothetical protein
BAA13334_I02529	Hypothetical protein
BAA13334_I02539	Hypothetical protein
BAA13334_I02563	Hypothetical protein
BAA13334_I02641	Hypothetical protein
BAA13334_I02697	Hypothetical protein
BAA13334_I02716	Hypothetical protein
BAA13334_I02805	Hypothetical protein
BAA13334_I02838	Hypothetical protein
BAA13334_I03094	Hypothetical protein
BAA13334_I03113	Hypothetical protein
BAA13334_I03253	Hypothetical protein
BAA13334_I03297	Hypothetical protein
BAA13334_I03335	Hypothetical protein
BAA13334_II00091	Hypothetical protein
BAA13334_II00282	Hypothetical protein
BAA13334_II00299	Hypothetical protein
BAA13334_II00408	Hypothetical protein
BAA13334_II00469	Hypothetical protein
BAA13334_II00654	Hypothetical protein
BAA13334_II00682	Hypothetical protein
BAA13334_II00816	Hypothetical protein
BAA13334_II01072	Hypothetical protein
BAA13334_II01240	Hypothetical protein
BAA13334_II01292	Hypothetical protein
BAA13334_II01303	Hypothetical protein
BAA13334_II01316	Hypothetical protein
BAA13334_II01365	Hypothetical protein
BAA13334_II01482	Hypothetical protein
BAA13334_II01580	Hypothetical protein
BAA13334_II01804	Hypothetical protein
BAA13334_II01806	Hypothetical protein

^*The genes that are found to be homologous in the vaccine strain S19.

Gene BAA13334_I03390 encodes a superoxide dismutase, which can directly detoxify the reactive oxygen intermediates (ROIs) in bacteria. The ROIs, including superoxide, hydrogen peroxide and hydroxyl radicals, damage macromolecular structure and macrophages and utilize this process to limit the intracellular replication of the Brucellae.^25,26 Thus, the absence of this gene in 104M could have an effect on the intracellular replication, which might indirectly decrease the strain's virulence.

Gene BAA13334_I01801 (ureF) encodes a urease accessory protein. In bacteria, ureases are multi-subunit metalloenzymes that hydrolyse urea to form carbonic acid and 2 molcules of ammonia, the latter of which could cause the pH to increase.²⁷ Recent studies have found that urease protects B. abortus during their passage through the stomach when acquired by the oral route in BALB/c mice, which is the major route of infection in human brucellosis.²⁸ Urease accessory protein is essential for incorporation of nickel into the active center of urease.²⁹ At least 3 urease accessory proteins, including UreF, are required for the activation of urease.^30-32 Thus, the lack of ureF can lead to the defective in urease activity, which might contribute to the attenuated virulence in 104M.

Gene BAA13334_I03416 encodes a 2-component transcription regulator, belonging to the LuxR family. The two component regulatory system in Brucella is predicted to be essential for sensing the phagosomal environment and changing from an extracellular to intracellular life style.³³ Recent experiments showed that the bvrS/bvrR mutants are attenuated in mice and show reduced invasiveness in cells.^33,34 Though this gene is not affiliated with the system, it is related with the system. The absence of this gene in 104M might have an effect on 2-component regulatory system functions.

In addition to the enzyme-coding genes, there are some genes encoding membrane proteins or transporters. For example, 2 genes (BAA13334_I00134 and BAA13334_II01500) encode the ABC transporter, which can play important roles in the Brucella strain virulence and influence the Brucella trafficking to compartments associated with the endoplasmic reticulum.³⁵ There are also other genes encoding the Mg²⁺ and Co²⁺ transporter, nitrite transporter and sulfate permease, all of which are involved in iron transport. These genes might not be directly related to the Brucella virulence, but all are related to the fundamental metabolism of the cell life cycle, contributing much to the survival in niches. These genes might play a role in the virulence attenuation of 104M.

There are 3 genes directly associated with the classical bacterial virulence: the gene BAA13334_I03052 encodes a MFS transporter, which functions as part of the bacterial efflux system and has an apparent effect on the virulence and resistance to antimicrobial agents. The gene BAA13334_II01445 encodes a Tellurite resistance protein-related permease, and the last gene BAA13334_II00407 encodes the flagellar motor switch protein G, which plays a central role in switching.

Thus, some of the A13334-specific genes are directly or indirectly associated with virulence. The absence of these genes might be one of the reasons for the virulence attenuation in 104M.

For the 104M-specific genes, only 7 genes were annotated, which encode an acid-shock protein, an AzlC family protein, a diguanylate cyclase, a non-classified regulator and 3 mobile element proteins. None of the functions were associated with bacterial virulence.

As a result, the absence of A13334-specific genes in 104M provides a candidate set of genes that resulted in the attenuated virulence in 104M.

Different virulence attenuation mechanisms between the vaccine strains 104M and S19

S19 was the other B. abortus vaccine strain. To determine whether the genes identified above are also the potential reason for virulence attenuation in S19, we conducted a search of the 99 genes against the genome of S19. As a result, 23 genes were found to be homologous in S19, which were marked with asterisk in Table 2. These genes contained most of the virulence-associated genes described above, suggesting that S19 possessed the candidate genes that might result in the attenuated virulence in 104M. It also suggested that there existed different sets of genes leading to the attenuation of 104M and S19.

To confirm it, a pairwise and reciprocal comparison between A13334 and S19 was conducted. A total of 428 and 95 genes were identified to be unique in A13334 and S19, respectively. The number of strain-specific genes was greater in the comparison between A13334 and S19 than in the comparison between A13334 and 104M. Moreover, the genes specific in A13334 and missing in S19, listed in Table S1, included a set of genes associated with virulence, such as multidrug resistance transporters, flgI and flhA, most of which were present in the genome of 104M. These differences further indicated that the 2 vaccine strains had different attenuation mechanisms. Thus, this paper would focus on the differences between 104M and A13334 to detect the potential genes associated with virulence attenuation in 104M.

Identification of Gene change associated with virulence between 104M and A13334

The comprehensive comparison also revealed 465 genes that are non-identical between the 2 genomes. These differences are mainly due to site mutations, premature stops and indels. Similarly, a literature search was also conducted to associate these genes with the published virulence-related genes,^23,24 and the identified virulence-associated genes with the corresponding function information were listed in Table 3.

Table 3.

A list of virulence-associated genes with mutations in 104M when compared with A13334

Gene	Mutation	Function category
wbkA	S228A, K349N	LPS
per	Q245K
virB10	A100V	Secretion system
BAA13334_I02537	Insertion: 460KKKAAP466	transporter system
BAA13334_I00045	P619L	Flagellar
cgs	V2412A	Intracellular survival
BAA13334_I1773	P76T	Amino acid metabolism
BAA13334_I02590	C-terminus mutation
eyrB	R194S	Sugar metabolism
galcD	Gene fusion
rpsA	S552R	DNA/RNA metabolism
aidB	A470V
rpoA	A219T
purF	S425G
cobB	K100Q	Vitamins/Cofactors
dxpS	S571G
BAA13334_ I02969	N117T	Stress
BAA13334_ I00578	V203G
dsbA	Q70K	Oxidoreduction
BAA13334_ I00370	N19T	Unknown
BAA13334_ II01604	R181Q

LPS

Lipopolysaccharide is vital to the structural and functional integrity of the Gram-negative bacteria outer membrane. The LPS of Brucella constitutes a key virulence mechanism for intracellular survival and replication. In Brucella, the genes in the wbk region of chromosome I are involved in the synthesis of the O-PS and its translocation to the periplasm.³⁶ In the genomes of A13334 and M104, the core genes of this region (wzm, wzt and wbkB) are missing. In this O-antigen gene cluster, gmd, per and wbkC use GDP-mannose as substrate to synthesize the GDP-Nf-Per, and here, wbkC acts as an Nf-Per-transferase.³⁷ However, compared with A13334, several mutations were present in the core genes, wbkA [T1012G(S338A), A1047C(K349N)] and per [C733A(Q245K)]. These data could suggest that the mutation of these 3 sites might contribute to the reduced virulence of M104.

Secretion system VirB or transport system

The T4SS in Brucella is typified by the virB operon encoding 12 proteins. Brucella uses T4SS for the translocation of virulence factors into mammalian cells. In the comparison, we identified that Both M104 and A13334 lost the OMP gene virB7. In addition, there is a mutation occurred in the inner membrane protein virB10, C299T(A100V).

In addition to the mutations in T4SS, there were 2 gene mutants in the transporter system: BAA13334_I02537 encodes the trigger factor, and the homologous protein in M104 has a fragment insertion 460KKKAAP466.

One non-synonymous mutation was also identified in the flagellar gene BAA13334_I00045[C1856T(P619L)]. Due to the stop codon mutation, the orthologous genes of BAA13334_II00088, BAA13334_II00089 and BAA13334_II00090 were merged to form a new gene in 104M. It has also been shown that Brucella mutants in fliF, flhA(BAA13334_II00089), flgI, flgE, fliC and motB are attenuated in BALB/mice.³⁸ These data may suggest that virulence attenuation may be related to the mutations in secretion and flagellar assembly systems.

Intracellular survival

Osmoregulated periplasmic glucans (OPGs) can be divided into 4 families (I-IV) on the basis of backbone organization and are constituents of the envelopes of gram-negative bacteria. Mutants deficient in OPG synthesis show altered chemotaxis and motility and reduced outer membrane stability and synthesis of exopolisaccharides.³⁹ Moreover, the mutants are unable to establish successful pathogenic or symbiotic association with their animal or plant hosts.⁴⁰ Brucella CβGs encoded by the cgs gene are considered family II OPGs and are neither O-substituted nor osmoregulated.⁴¹ In the comparison, a mutation was identified between the gene cgs of A13334 and M104, BAA13334_I00290[T7235C(V2412A)], which suggests that the mutation in cgs is among the reasons leading to the attenuated virulence.

Amino acid metabolism

Some studies^23,42,43 have reported that Brucella encounters an environment poor in nutrients before reaching the replicative compartment, and failure to resist this sudden lack of nutrients would result in attenuation. In the amino acid metabolism pathway, there were 2 genes identified with mutations. One is the gene BAA13334_I1773, which encodes an aminotransferase with a C226A(P76T) mutation. The other is the gene BAA13334_I02590, which encodes an extracellular solute-binding protein. Its homolog in 104M has a different 17 amino acid region in C-terminus, which might be caused by a site mutation and premature stop.

Sugar metabolism

Eryrthritol metabolism by Brucella has been identified as a trait associated with the capability of the pathogen to cause livestock abortions. The preferential growth of Brucella in the foetal tissues of cattle, sheep, goats and pigs was shown to be due to the high erythritol concentration. Although Brucella infects and causes brucellosis in other organisms such as humans, rats, rabbits and guinea pigs, overwhelming infection of the placental and foetal tissues is not observed, which is also associated with low concentrations of erythritol. In the M104 genome, the eryB gene has a mutation (R194S).

In addition to the erythritol metabolism, the galcD, which encodes D-galactarate dehydratase, merged with its upstream gene, which encodes a galactarate dehydratase, to form a new gene in 104M at the same location due to the stop codon mutation.

DNA/RNA metabolism

Purine/Pyrimidine synthesis, repair and regulatory genes are essential for Brucella intracellular replication. Mutants affecting RNA helicase or DNA gyrase activities suggest that Brucella might also use other categories as well to regulate virulence genes.⁴⁴ In this gene cluster, 4 genes were identified with non-synonymous mutations: gene rpsA, encoding the ribosomal protein S1, changed with A1564C(S522R); aidB, functioning as protection against to alkylation damage to DNA, changed with C1409T(A470V); the third gene rpoA, encoding the RNA polymerase α subunit, changed with G655A(A219T), and the last gene purF, encoding the purines synthesis, changes with S425G.

Vitamins/Cofactors

There were also mutations occurring in 2 genes involving vitamins or cofactor metabolism. cobB is a cobalamin-synthesis gene with the mutation K100Q; and dxps, taking part in thiamine synthesis, has a mutation of S571G.

Stress

Genes encoding stress proteins have been an obvious choice for directed mutagenesis for virulence studies and vaccine generation. Within the M104 genome, genes encoding protease (BAA13334_ I02969) and chaperone (BAA13334_ I00578) showed one non-synonymous mutation, namely, A350C(N117T) and T608G(V203G), respectively.

Oxidoreduction

DsbA has been shown to be involved in toxin production, adhesion, motility, extracellular enzyme production, and the type III secretion system.^45,46 An A208T(Q70K) mutation was detected in the dsbA gene (BAA13334_ I02476) of M104.

Unknown function

Two ORFs encoding hypothetical proteins were slightly different between the attenuated strain M104 and the virulent strain A13334. One is an orphan gene (BAA13334_ I00370) with the mutation of A56C(N19T), and the other is the unknown function gene (BAA13334_ II01604) with a mutation of G389A(R181Q).

Conclusion

The B. abortus vaccine strain 104M was developed in China, and this vaccine was proven to be stable in antigenic structure, virulence, and immunogenicity. Since 1965, it has been widely used to vaccinate cattle and human against brucellosis in China.

In this study, we determined the whole-genome sequence of this vaccine strain 104M and found high genome similarities between 104M and A13334. A comprehensive comparative analysis against the whole-genome sequence of the virulent strain A13334 reveals that the genome sequences are highly linearly conserved with more than 97% identity. Meanwhile, a set of genome differences, including the difference of gene content and mutation of orthologous genes, was found between both strains, many of which were in genes associated with virulence and cell survival.

The unique genes identified for one genome mean the loss of the same genes in the other genome. Thus, the lack of A13334-specific genes in 104M provides a direct basis for the virulence attenuation. Especially in the 99 genes, some are directly or indirectly related to virulence and survival of the Brucella strain. In contrast, the 104M-specific genes have no clear association with virulence, but can act as a set of potential genes for identification of vaccine strain 104M. In addition tothe strain-specific genes, many mutations in the genes shared by both genomes were identified. Similarly, some of them are virulence genes or virulence-associated genes.

Collectively, this study presents a set of candidate genes that might be associated with virulence and will helpful for the elucidation of mechanisms of Brucella virulence attenuation.

Methods and Materials

Isolation extraction and identification

B. abortus 104M was obtained from the LanZhou Institute of Biological Products in China. The total genomic DNA was extracted and purified by the modification of a previously described method.²¹ An aliquot of the DNA was subjected for analysis using the BioAnalyzer (Agilent Technologies) and was confirmed for no degradation.

Genome sequencing, assembly and annotation

Paired-end sequencing was conducted through the high throughput Illumina sequencing technology. We first constructed 2 DNA libraries with 500 and 2000 bp fragments. Then, pair-end 90-bp reads were collected. A total of 170 Mb of clean reads were obtained, resulting in more than 47X coverage of this strain's whole genome. These quality filtered reads were then assembled into scaffolds using the SOAP de-novo software and a total of 14 scaffolds were formed.

The 14 scaffolds were aligned to the whole genome sequences of B. abortus A13334 to identify the putative gaps not sequenced in the whole genome of 104M. Primers were then designed and the obtained draft genome of 104M was used as a template in PCR to amplify the segments across the gaps. The purified PCR amplicons were used as templates in sequencing. The newly generated sequences, together with the scaffolds, were used to determine the complete genome sequence.

The final complete genome sequences of B. abortus vaccine strain 104M is available in GenBank with the accession number CP009625-CP009626, which correspond to the sequences of 2 chromosomes. The following genome sequence annotation was conducted using the RAST server, an automated genome annotation pipeline.

Comparative genomic analysis

Whole genome sequences and the corresponding gene sets of B. abortus A13334 and other B. abortus strains were downloaded from NCBI database. The multiple genome alignment was based on the software of Mauve to detect the SNPs and the genome structure similarity. Pair-wise and reciprocal comparisons were performed to identify potential genes that differ between the attenuated strain, M104 and virulent strain, A13334. The detailed variations in virulence-associated genes, including SNPs, deletions, insertions, and premature stops were further identified by Muscle alignment and then calculated using Python scripts.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Authors' Contributions

JY, SL and BW formulated this study. DY performed the research. YH, XZ and JX performed the sequencing of the whole genome of Brucella abortus vaccine strain 104M. DY, LL and JY analyzed the data. DY and JY wrote the paper. All authors read and approved the final manuscript.

Funding

This work was supported by National Science and Technology Major Project of the Ministry of Science and Technology of China (2012ZX10003008002 to S. Li), National Science and Technology Major Properties for Major New Drugs Innovation and Development (20132×09304101).

Supplemental Material

Supplemental data for this article can be accessed on the publisher's website.