Adaptive Changes in Mycobacterium avium Gene Expression Profile Following Infection of Genetically Susceptible and Resistant Mice

Full Text

INTRODUCTION Infectious diseases caused by intracellular pathogenic bacteria represent a significant challenge in health care. The course of the infection depends not only on the protective mechanisms (native and acquired immune response, and mucous barriers), but also on the specific expression of bacterial genes. Altered expression as a response to the immune reaction of the host organism is critical for the survival and functioning of pathogenic bacteria. An analysis of these changes is important for understanding how infectious diseases proceed and developing effective approaches towards their treatment. Mycobacterium avium are widespread mycobacteria that become intracellular pathogens in humans in the absence of normal T-cell-mediated immunity [1, 2]. These bacteria are found in approximately 70% of incurable AIDS patients and are believed to be the main cause of death in such patients [3]. In patients with weakened immunity (older people and children), M. avium may cause chronic lung diseases [4-6]. Experiments modelling the infection in mice of the C57BL/6 (B6) and derivative strains with knockout mutations in genes essential for immunity showed that T-cell-medi-ated immune response to M. avium had both defensive, as well as pathogenetic functions. In such an infection, the balance between the immune response and pathogenic processes in lung tissue is very similar to that of tuberculosis [7-9]; therefore, we can assume that the diseases caused by these mycobacteria are similar not only in their immune system mechanisms, but also in the mechanisms employed by the pathogens to overcome this defence. It has been shown that mice of the I/St (I/StSnEgY-Cit) and B6 strains differ in their ability to resist an M. avium-induced infection [10]. Respiratory infection in B6 mice leads to a prolonged infiltration of lung tissue by macrophages and neutrophils, leading to the formation of necrotic lung granulomas and death. In contrast, in the I/St mice the infection is controllable, produces moderate infiltration of lung tissue, leading to small and medium granulomas without a necrotic centre, and the animals survive. The susceptibility of B6 mice to the M. avium-induced infection was shown to result from the presence of the nonfunctional allele of the Nrampl (natural resistance-associated macrophage protein-1) gene in their genome. The protein coded for by that gene consists of 12 transmembrane domains and is expressed at the membranes of late lysosomes and phagosomes. Nramp1 functions by removing divalent cations (Fe2+, Mn2+, etc.) from phagosomes, thus depriving the mycobacteria of important metabolites [10]. Table 1. Oligonucleotides and primers used for coincidence cloning. Name 5’-3’ structure Suppressive adapter 1A (resulted from anneal of equi-molar mixture of 1A long and 1A short ) 1A long GTAATACGACTCACTATAGGGCAGCGTGGTCGCGGCCGAGAG 1A short CTCTCGGCCG Suppressive adapter 1B 1B long GTAATACGACTCACTATAGGGCAGGGCGTGGTGCGGAGGGCGGC 1B short GCCGCCCTCC Suppressive adapter 2A 2A long GTAATACGACTCACTATAGGGCAGGCAGGCGGTGGTGGGCAGGC 2A short GCCTGCCCAC Suppressive adapter 2B 2B long GTAATACGACTCACTATAGGGCAGCGGAGGCGGTAGGAGGCGGA 2B short TCCGCCTCCT External primer T7 GTAATACGACTCACTATAGGGC Internal primers pr 1A AGCGTGGTCGCGGCCGAGAG pr 1B AGGGCGTGGTGCGGAGGGCGGC pr 2A AGGCAGGCGGTGGTGGGCAGGC pr 2B AGCGGAGGCGGTAGGAGGCGGA The B6 immune response is characterized by an increased production of IFN-y, TNF-a, and, especially, IL-12. We suppose that the differences in the immune response to M. avium infection are manifested in the differences in the pathogen expression in the lungs and lymphoid organs of mice of the susceptible strain versus those of the resistant strain, showing that the mechanisms essential for resistant host survival may not activate during infection of the susceptible host. This work endeavoured to study the biochemical processes involved in the adaptation of M. avium to genetically different host organisms. We compared sequences transcribed in mice of the I/St and B6 strains in the 13th week of infection. EXPERIMENTAL Standard DNA and RNA procedures were carried out according to ref. [11]. Genomic DNA of the M. avium 724R strain was isolated according to the procedure described in ref. [12]. Infection Mice of inbred strains I/StSnEgYCit (I/St) and C57BL/6YCit (B6) were bred and maintained under conventional, non-specific-pathogen-free conditions at the Animal Facilities of the Central Institute for Tuberculosis (Moscow, Russia) in accordance with guidelines from the Russian Ministry of Health (guideline 755) and the NIH Office of Laboratory Animal Welfare (assurance A5502-06). Female mice, 2.5-3.0 months old, were infected by the respiratory route with 1-2 x 103 viable CFU of M. avium 724R strain, characterized in ref. [13], using an inhalation exposure system (Glas-Col, USA) according to the procedure described in ref. [10]. RNA isolation and cDNA synthesis RNA was isolated from the lungs of mice of both strains in the 13th week after infection, using the RNA Isolation System kit (Promega, USA). RNA samples were treated with DNase I (MBI Fermentas, Lithuania) to remove DNA traces. The first cDNA strand was constructed using oligonucleotide primers BR (5’-AAGC AGTGGTATCAACGCAGAGTAC(N)9) and SMART (5’-AAGCAGTGGTATCAACGCAGAGTACGCrGr GrG). Both primers (at 12 ^M) were annealed with 2 ^g of total RNA in 11 ^l of solution. The mixture was incubated for 2 minutes at 70°C and then placed in ice for 10 minutes. cDNA was synthesised using reverse transcriptase PowerScript II (Clontech, USA). In parallel with reverse transcription (RT+), a reaction used as a control (RT-) without reverse transcriptase was performed. The RT+ and RT- mixtures were incubated at 37oC for 10 minutes, then for 40 minutes at 42°C. cDNA was synthesised in 30 PCR cycles (95oC for 20 sec, 64oC for 20 sec, and 72oC for 2 min) using 5S primers (5’-GT-GGTATCAACGCAGAGT). Then, cDNA was purified using QIAquick PCR Purification kit (Qiagen, USA). Coincidence cloning was carried out following the procedure described in ref. [14]. Genomic DNA of the M. avium 724R strain and total cDNA samples (synthesised using total RNA) were fragmented with restrictases RsaI and AluI. The obtained genomic DNA fragments were ligated with suppressive adaptors 1A for hybridisation with the I/St cDNA sample, and adapters 1B for hybridisation with the B6 cDNA sample (Table 1). Suppressive adapters 2A and 2B were ligated to cDNA fragments from the lung tissues of I/St and B6 mice, respectively. A mixture of 100 ng of the genomic DNA sample and 100 ng of one of the cDNA samples in 2 |^l of the hybridisation buffer (50 mM HEPES, pH 8.3; 0.5 M NaCl; 0.02 mM EDTA, pH 8.0) was incubated at 99oC for 5 minutes (denaturation) and then at 68oC for 18 hours (renaturation). After this procedure, 100 ^l of the hybridisation buffer at 68oC was added to the mixture, and 1 ^l of the resulting solution was used as a template for PCR. The first PCR stage was performed in a 25 |^l reaction volume containing 10 pmol of T7 primer. After incubation for 5 minutes at 72oC (filling-in sticky ends), 20 amplification cycles were carried out (94oC for 30 sec, 66oC for 30 sec, and 72oC for 90 sec). The second stage of amplification was performed with 10 pmole of internal primers pr1A/pr1B and pr2A/pr2B, and it consisted of 25 cycles (94oC for 30 sec, 68oC for 30 sec, and 72oC for 90 sec), using the PCR product of the first stage, diluted ten-fold. The amplification product was purified using the QIAquick PCR Purification kit (Qia-gen) and then used for 454 sequencing. Total cDNA from infected tissue I Ligation of suppressive adapters 454 SEQUENCING Nucleotide sequences of cDNA libraries were determined by massive parallel pyrosequencing using the genetic analyser GS FLX (Roche, Germany) and a 20 x 75 cm picotitration plate. The sequences of 83,000 independent reactions were determined. The sequences were mapped to the genome sequence of the M. avium strain 104, since the M. avium 724R genome has not yet been sequenced. The number of cDNA fragments 1. Denaturation/renaturation 2. Fill-in of sticky ends Homoduplexes Host cDNA Pathogen genomic DNA Pathogen cDNA Homoduplexes y PCR with primers No amplification Fig.1. Coincidence cloning. Suppressive oligonucleotide adapters are ligated to fragments of bacterial genomic DNA and total cDNA. The samples are mixed, denatured, and slowly renatured, which leads to the formation of two types of duplexes. Due to selective suppression of PCR, only heteroduplexes containing fragments of bacterial genomic DNA and bacterial cDNA are amplified. corresponding to each gene was determined using the BLASTn algorithm (http://blast.ncbi.nlm.nih.gov/ Blast.cgi). A sequence was considered to belong to a certain gene if a fragment of that sequence had more than 95% homology with the gene segment longer than 40 nucleotides. The M. avium genes, the expression of which in samples I/St and B6 is significantly different, were determined following the procedure described in ref. [15]. RESULTS AND DISCUSSION The course of the pathology and immune response to M. avium infection in mice from the susceptible (B6) and resistant (I/St) strains are discussed in detail in refs. [10] and [16]. The airborne infected susceptible B6 mice died after 7 months, while the resistant I/St mice survived for longer than 11 months. In the susceptible B6 mice, the lung pathology developed quickly, Table 2. M. avium genes differentially transcribed in the lungs of infected I/St and B6 mice. Gene Coded protein Increased expression in the lungs of the I/St mice MAV_2015 MbtG; mycobactin lysine-N-oxygenase MAV_1696 Glutamate dehydrogenase MAV_1304 NarH; nitrate reductase, p-subunit MAV_2379 MetH; vitamin B12-dependent methionine synthase MAV_2385 Mce protein MAV_2063 Mce protein MAV_2386 Mce protein MAV_0118 PPE protein MAV_3109 RifB; polyketide synthase 7 MAV_0880 3-Ketosteroid-6-1-dehydrogenase MAV_3000 Acyl-CoA dehydrogenase MAV_4019 Assumed acyl-CoA dehydrogenase MAV_4679 Cyclopropane fatty acid synthase Increased expression in the lungs of the B6 mice MAV_2514 PPE protein MAV_2924 PPE protein MAV_2926 PPE protein MAV_2244 GlnA; glutamine synthetase MAV_4011 NO-reductase, p-subunit MAV_1074 SucC; succinyl-CoA-synthase, p-subunit MAV_3303 AcnA; aconitate hydratase MAV_1130 NADH-dehydrogenase I, H-subunit MAV_4040 NADH-dehydrogenase I, H-subunit MAV_1524 ATP-synthase F^O, 6-subunit MAV_5034 Transposase MAV_1059 Transposase accompanied by enhanced infiltration of lung tissue by immune system cells and increased production of pro-inflammatory cytokines IFN-y, TNF-a, IL-6, and IL-12. Two parameters showing the susceptibility to infection correlated well: in the susceptible B6 mice, M. avium grew faster in lungs, and the lung pathology was deeper than in the resistant I/St mice. We studied transcribed sequences in vivo using the coincidence cloning method we had developed previously [14, 17]. From the lungs of infected mice, we isolated total RNA, a mixture of the mice and bacterial RNA, with the amount of bacterial RNA being very small (less than 0.1-0.2% according to [18]). Using the total RNA from the I/St and B6 mice, total cDNA was synthesized. In the coincidence cloning method (Fig. 1), total cDNA and M. avium genomic DNA were denatured and renatured in one mixture. After a two-step selective PCR amplification, a set of fragments enriched with the bacterial cDNA fragments was obtained. Qualitative (determination of nucleotide sequences of specifically expressed genes) and quantitative (the level of their expression) analyses of the sets were performed using parallel pyrosequencing. The sequencing produced two libraries of M. avium cDNA sequences expressed in the lung tissue of infected I/St and B6 mice. We selected a series of genome loci, the expression of which was higher in sample I/ St than in sample B6, and a series of loci, the expression of which was higher in sample B6 than in sample I/St (Table. 2). Locus annotation was performed using the KEGG (www.genome.jp/kegg) database. We anticipated that differential gene expression in the samples could be a manifestation of the microorganism’s environmental adaptation; therefore, the products of the genes we found could be potential virulence factors. We found the differential expression of the PPE gene family (MAV_0118, MAV_2514, MAV_2924, and MAV_2926). These proteins play an important role in the course of the mycobacterial infection because of both their antigen and immune functions. These acidic proteins, rich in glycine, are identified by the specific Pro-Pro-Glu (PPE family) and Pro-Glu (PE family) domains; they often contain polymorphic GC-rich sequences (PGRSs) and multiple copies of basic polymorphic tandem repeats. It is believed that these proteins are expressed on the cell’s surface and are responsible for antigen variability, inducing different immune responses depending on the type of PE/PPE proteins expressed on the cell’s surface [19]. Thus, the MAV_0118 gene is expressed in the resistant mice, while the MAV_2514, MAV_2926, and MAV_2924 genes are expressed in the susceptible mice. Since the mechanism of PPE protein action remains unknown, the abovementioned observation is hard to explain; however, it B6 (susceptible strain) - Anaerobic environment-transition to nitrate respiration Fe2+ I/St (resistant strain) - Frequent cell division - NO-reductase inactivates NO, secreted by macrophages - Decrease of cell wall permeability - Synthesis of mycobactin to compensate for Fe2+ deficiency LATENT INFECTION HOST DEATH Fig. 2. M. avium metabolic state in the lungs of the I/St and B6 mice is possible that the differential PPE expression is due to the differences in the immune responses. The expression of the MAV_2244 locus is increased in the B6 sample. This gene is an ortholog of the glnAl M. tuberculosis gene; it codes for glutamine synthetase, a key enzyme for nitrogen assimilation. It has been shown that this enzyme is important for M. tuberculosis persistence in macrophages. It is possible that in infected B6 mice M. avium enters into an environment auxotrophic for L-glutamine [20]. The MAV_4011 locus that codes for the cytochrome b-containing subunit of NO-reductase is also worth mentioning. This enzyme reduces NO to N2O, and it participates in the denitrification process in some soil microorganisms. However, no denitrification is observed in the M. avium from the susceptible B6 mice. It is speculated that M. avium use NO-reductase to get rid of the NO released by the macrophage into the endosomes and thus avoid the harmful effects of NO [21]. This could be the reason why M. avium are resistant to NO [22]. The expression of NOreductase in M. avium from the lungs of the susceptible mice could be a result of the stronger immune response and increased NO production by macrophages. In sample B6, we observed a more active and diverse expression of genes coding for the Krebs cycle enzymes: MAV_1074 and MAV_3303 coding for succinyl-CoA-synthesase and aconitate hydratase, respectively; and of genes coding for the proteins important for oxidative phosphorylation, as well as the respiratory electron-transport chain proteins: MAV_1130, MAV_4040, and MAV_1524. It is likely that in the susceptible mice, respiration is increased during persistent infection in order to supply the pathogen. The MAV_4040 locus codes for one of the NADH-dehydrogenase I subunits, which is typical for the M. tuberculosis virulent form, during exponential proliferation of the pathogen [23]. The expression of the MAV_5034 and MAV_1059 genes coding for transposases in sample B6 indicates an enhanced level of gene rearrangements. Also, in sample B6, we detected an increased expression of the MAV_5024 and MAV_5027 genes coding for type II re-striction-modification enzymes that protect cells from foreign DNA. The increased expression of the MAV_0382 (subunits of DNA-polymerase III) and MAV_4450 (ribosomal protein) genes in sample B6 indicates an increase in the DNA replication level due to a more frequent mitosis, as well as a higher translation level. In sample I/St, there was a very high level of expression of the MAV_2015 gene that codes for mycobactin lysine-N-oxygenase (MbtG). This enzyme is responsible for one of the last stages of mycobactin synthesis; it is an iron-chelating agent that supplies the microorganism with iron from the environment [24]. It has been shown for M. tuberculosis that the activation of the mbt B-H gene cluster involved in mycobactin synthesis occurs either when the environment is depleted in iron [25] or in an anaerobic environment [26]. The expression of this gene is high in M. avium from the resistant I/St mice, but it is very low in the B6 mice. As mentioned above, these two mice strains differ in the Nrampl gene allele that codes an ionic pump which is assumed to pump out divalent cations from the endosomal region, where M. avium is located [16]. There is the functional allele of this gene in the I/St mice, as opposed to a nonfunctional one in the B6 mice. Apparently, the M. avium endosomes from the resistant mice are iron-deficient, and the microorganism synthesizes vast amounts of mycobactin in order to compensate for the deficiency. In M. avium from the I/St mice, an increased expression of the MAV_1696 gene coding for NAD+-de-pendent glutamate dehydrogenase is observed. It is believed that, in contrast to NADP+-dependent glutamate dehydrogenase, which is responsible for nitrogen assimilation, in microorganisms the former enzyme takes part in the glutamate catabolism, and this gene expression is independent of the NH4+ concentration. On the other hand, it has been shown recently that in M. smegmatis the expression of the msmeg_4699 gene, an ortholog of MAV_1696, increases in response to NH4+ deficiency [27]. In addition, there is no gene coding for NADP+-dependent glutamate dehydrogenase in the M. avium genome [28]. Some researchers speculate that, in mycobacteria, nitrogen assimilation involving NAD+-dependent glutamate dehydrogenase may be more energy-efficient than via the GS/GOGAT pathway; this being important, for example, when the pathogen is in a latent state [27]. The MAV_2379 gene coding for B12-dependent methionine synthase MetH is expressed at a high level in M. avium from the I/St mice. This protein is involved in the final stage of methionine synthesis. In the M. avium genome, this reaction is controlled by MetE-B12-independent methionine synthase, which is not expressed in the presence of vitamin B12 [29]. The regulation of the metH gene expression has not been studied in detail, so the reason for the increase in its expression in the resistant mice is not quite clear. The MAV_1304 locus coding for the P-subunit of nitrate reductase is of particular interest. This gene is orthologous to the narH gene of M. tuberculosis. Its product is a subunit of anaerobic nitrate reductase NarGHJI, an enzyme enabling nitrate respiration in the absence of oxygen. Mutants lacking NarH cannot reduce nitrogen under anaerobic conditions [30]. When this gene was deleted in M. bovis, BCG bacteria demonstrated normal growth in vitro with sufficient oxygen supply; however, they appeared significantly less virulent when used for infecting mice [31]. The expression of the MAV_1304 gene in M. avium from the lungs of the I/St mice might be an indication that, due to the harmful effects of the host’s defence systems, the microorganism is subjected to anaerobic conditions and has to switch to nitrate respiration. The MAV_2063, MAV_2385, and MAV_2386 genes coding for proteins from the Mce family are expressed in the resistant mice. The function of the Mce proteins has not been clarified, although it is known that they enable invasiveness. These proteins supposedly represent a new group of ABC-transporters participating in the remodeling of the cell’s membrane [32]. The MAV_4679 locus coding for an enzyme involved in the synthesis of mycolic acids is expressed at a high level in M. avium from the I/St mice. An ortholog of this gene in M. tuberculosis is important for persistence in mice lungs. Mutants of this gene cannot cause lung infection in mice [33]. The MAV_3109 locus codes for the RifB protein and is an ortholog of the M. tuberculosis pks7 gene. An increased expression of the gene is observed in infected mice of the resistant strain. The protein product of this gene codes for an enzyme involved in the synthesis of phthiocerol dimycocerosate, one of the components of the mycobacterial cell wall, which ensures its impermeability [34]. The MAV_0880 locus codes for 3-ketosteroid-§-1-dehydrogenase, one of the enzymes involved in cholesterol catabolism. During the M. tuberculosis-induced infection, cholesterol provides the pathogen with energy for persistence in macrophages [35]. In M. avium from the I/St mice, we observed an increased expression of the MAV_3000 and MAV_4019 genes coding for enzymes degrading fatty acids: acyl-CoA-dehydrogenase and acyl-CoA-synthase. During persistence in macrophages, the catabolism of fatty acids is the primary energy source for M. tuberculosis [36]. CONCLUSION This paper contains the first description of the M. avium transcriptome during infection in vivo. Until now, only a single publication on the M. avium paratubercu-losis transcriptional response to various factors in vitro [37] has been available. We employed the model of genetic control of susceptibility to M. avium infection and disease severity in mice in order to detect the sequences that are transcribed differently in infected mice from the genetically resistant and genetically susceptible strains, i.e. when the pathogen persists in genetically different microenvironments. We obtained data on the qualitative and quantitative differences in the transcription profiles of genes of bacteria persisting in the resistant and susceptible mice, which indicate some changes in the metabolism of M. avium (Fig. 2). In the course of the infection in the genetically susceptible organism (B6 strain), we found an increased expression of several genes responsible for nitrogen assimilation, NO reduction, the Krebs cycle, and oxidative phosphorylation, as well as replication and translation. The infection proceeds with active division of the mycobacteria and death of the host organism. In the course of infection in the genetically resistant organism (I/St strain), we found an increased expression of several genes responsible for the modification of the cell surface’s properties, switching to anaerobic nitrate respiration, degradation of fatty acids, synthesis of polycyclic derivatives of fatty acids, and biosynthesis of mycobactin and other polyketides. In general, the changes in the M. avium metabolism are an indication that, in the resistant mice, the bacterial pathogen transitions to the latent state, because of the deficit in divalent metal ions.

About the authors

D V Ignatov

Shemyakin and Ovchinnikov Institute of Bioorganic Chemistry, Russian Academy of Sciences

T A Skvortsov

Shemyakin and Ovchinnikov Institute of Bioorganic Chemistry, Russian Academy of Sciences

K B Majorov

Central Research Institute for Tuberculosis, Russian Academy of Medical Sciences

A S Apt

Central Research Institute for Tuberculosis, Russian Academy of Medical Sciences

T L Azhikina

Shemyakin and Ovchinnikov Institute of Bioorganic Chemistry, Russian Academy of Sciences

Email: tatazhik@ibch.ru

References

Supplementary files