Leptospira and E. coli caught in the act

I found this web photo of E. coli mating with the spirochete Leptospira biflexa in a process called conjugation (image source, Mathieu Picardeau, Pasteur Institute).  The donor E. coli cell is transferring a copy of a plasmid bearing antibiotic resistance genes to the recipient spirochete.  The DNA is most likely pushed through a pore that forms between the mating pair where the outer membranes come together.

There are many types of plasmids, but only self-transmissible plasmids are capable of transferring copies of themselves to other bacteria by conjugation. These plasmids carry a set of at least 20 genes collectively called tra (transfer), which encode all of the proteins necessary to carry out conjugation.  The conjugational proteins assemble into several structures, including the sex pili, which bring the mating pair together, the relaxosome, which processes the DNA for transfer, and the poorly characterized pore through which the DNA traverses.  The plasmids can also harbor additional genes that have no role in conjugation, including genes encoding resistance to antibiotics.  RP4 is one example of a self-transmissible plasmid that can transfer itself to a wide range of bacteria species.  Self-transmissible plasmids have been found in many different bacteria, yet none have been discovered in spirochetes.

Transformation is the microbiologist's favorite genetic tool for delivering DNA of their choosing into bacteria. Unfortunately for those interested in leptospirosis, transformation of disease-causing species of Leptospira such as L. interrogans is difficult.  Conjugation employing a laboratory strain of E. coli as a donor provides scientists another route for delivering DNA into Leptospira.  For example, the plasmid illustrated below (Figure 1 from Picardeau, 2008) has been used to ferry the Himar1 transposon into Leptospira for random insertional mutagenesis.  Many readers may be most familiar with the F conjugational plasmid of E. coli, but the conjugational machinery found on the RP4 self-transmissible plasmid is used here since it is able to deliver DNA to a wide range of bacteria species.

Figure 1 from Picardeau, 2008.  The Himar1 transposon consists of the arrowheads and everything in between, including the kanamycin-resistance gene (Km^R).  The RP4 oriT element and the genes encoding the C9 tranposase and spectinomycin resistance (Spc^R) lie outside of the transposon.


The critical element of the plasmid is the RP4 oriT sequence where relaxase, a component of the relaxosome, nicks the DNA to initiate the transfer process.  The tra genes were removed to permit easy manipulation of the plasmid.  To perform conjugation, the plasmid was transformed into a special E. coli strain that encodes the RP4 tra genes on its chromosome.  The E. coli cells were then mixed with Leptospira and concentrated onto a filter to facilitate mating.  After allowing them to mate for 20 hours, the mating mixture was plated onto Leptospira medium agar plates containing the antibiotic kanamycin to recover Leptospira mutants with the transposon on one of its two chromosomes.  The plasmid itself is unable to replicate in Leptospira, so the transposon must hop onto a chromosome following plasmid transfer to enable growth of kanamycin-resistant Leptospira into colonies.  The donor E. coli bacteria had been genetically modified to require the nutrient diaminopimelate (DAP) to counterselect the donor on the agar plates, which were lacking DAP.

It is not feasible to screen L. interrogans insertion mutants for a desired phenotype (trait) following a single mating experiment since only a few hundred kanamycin-resistant colonies can be recovered.  Tens of thousands of mutants would be necessary to ensure coverage of (almost) all L. interrogans genes.

One application of this genetic tool is to perform multiple mating experiments to generate a library of mutants with insertions of Himar1 in different L. interrogans genes.  The sequence of the insertion site of the transposon in the chromosome can be obtained easily with today's sequencing technology.  Further experiments can be performed to examine any mutants with insertions in genes that hold the investigator's interest.  Several labs have teamed up to embark on a similar approach by delivering the Himar1 transposon into L. interrogans by transformation (see the Murray 2009 paper), which does not yield as many colonies as conjugation.

References

Picardeau, M. (2008).  Conjugative transfer between Escherichia coli and Leptospira spp. as a new genetic tool.  Applied and Environmental Microbiology 74(1):319-322.  DOI: 10.1128/AEM.02172-07

Murray G.L., Morel, V., Cerqueira G.M., Croda, J., Srikram, A.,  Henry, R., Ko, A.I., Dellagostin, O.A., Bulach, D.M., Sermswan, R.W., Adler, B., and Picardeau, M. (2009).  Genome-wide transposon mutagenesis in pathogenic Leptospira species.  Infection and Immunity 77(2):810-816.  DOI: 10.1128/IAI.01293-08

The genetics of both host and pathogen matter in antibiotic-refractory Lyme arthritis

The arthritic form of Lyme disease was first reported in the 1970s by Allen Steere, who described the condition in a group of children (and a few adults) residing in and around the town of Lyme, Connecticut. Lyme arthritis can strike when Borrelia burgdorferi introduced into the skin by an Ixodes tick burrows into deeper tissues and ends up in the joints, usually the knee. Swelling results from an inflammatory response to B. burgdorferi residing in the joint. Lyme arthritis is treated with antibiotics, which destroy the bacteria driving inflammation. Unfortunately, arthritic symptoms endure in ~10% of treated patients despite the complete or almost complete eradication of the infection, as determined by negative PCR tests for B. burgdorferi DNA in joint fluid. Such cases are called antibiotic-refractory Lyme arthritis, which can persist for months or sometimes years. In severe cases cartilage and bone erode. Although the pathogenesis of antibiotic-refractory Lyme arthritis could involve persistence of small numbers of B. burgdorferi (or their antigens) in the joints, investigators have been seeking an autoimmune mechanism to explain the prolonged attack on joint tissue by the immune system after the spirochetes have been cleared.

Many autoimmune diseases are linked to variants of HLA (immunity) genes such as those encoding the MHC class II complex. Antibiotic-refractory Lyme arthritis is associated with MHC class II variants that are able to bind to fragments of the B. burgdorferi protein OspA (outer surface protein A) encompassing amino acid residues 165 through 173. Antigen-presenting cells whose MHC class II molecules display OspA_165-173 peptides on their surface stimulate T cells that recognize the OspA peptide. How OspA_165-173-reactive T cells cause autoimmunity has been an area of intensive research, yet a clear answer has not emerged.

One potential pathway to autoimmunity is molecular mimicry, in which a cross-reactive host protein in the joint continues to stimulate OspA_165-173-specific T cells even after the eradication of B. burgdorferi by antibiotics. Although the simplicity of the molecular mimicry model is appealing, exhaustive efforts to find a cross-reactive autoantigen that stimulates OspA_165-173-specific T cells have failed. Moreover, levels of OspA_165-173-reactive T cells decline soon after initiation of antibiotic therapy despite continuing arthritis following treatment. Thus, chronic arthritis does not seem to involve molecular mimicry driven by a cross reaction between the OspA_165-173 epitope and a self-antigen in the joint. It is possible that molecular mimicry involves another B. burgdorferi antigen that is able to bind the MHC class II variants found in genetically susceptible individuals.

Other potential routes to autoimmunity in antibiotic-refractory Lyme arthritis patients emphasize the role of the high levels of key proinflammatory cytokines and chemokines found in their joint fluid, levels even higher than those found in treatment-responsive patients prior to initiation of antibiotic therapy:

• In a model known as bystander activation, the immune response to OspA_165-173 (or another B. burgdorferi antigen) causes an excessive inflammatory response that activates other T cells that react to autoantigens in the joint.
• The immune system is unable to turn off the intense inflammatory response associated with OspA_165-173 after the spirochetes are cleared from the joint.

Although much attention has been focused on the role of host genetics, a recent study indicates that the genetics of the pathogen could also influence the course of Lyme arthritis. In the July 2009 issue of Arthritis and Rheumatism, Allen Steere and his collaborators showed that antibiotic-refractory Lyme arthritis is associated with different strains of B. burgdorferi. The strains were typed from joint fluid samples collected before or during antibiotic treatment. Among the methods available to group B. burgdorferi isolates, they used the 16S-23S ribosomal RNA intergenic spacer type (RST), of which there are three. Antibiotic-refractory arthritis was defined as joint swelling lasting for at least 3 months after the start of antibiotic treatment. Antibiotic treatment consisted of 8 weeks of oral antibiotics or up to 4 weeks of antibiotics administered intravenously. Joint fluid from all 17 patients in the study tested positive by PCR for B. burgdorferi DNA prior to or during antibiotic treatment.

The authors found that all 7 Lyme arthritis patients infected with RST1 strains had the antibiotic-refractory form. Joint fluid was obtained after antibiotic treatment from 5 of the 7 patients; all 5 samples tested negative for B. burgdorferi DNA by PCR. In contrast, 2 of 6 and 3 of 4 infected with RST2 and RST3 strains, respectively, were successfully treated with antibiotics (see the table below from the Jones et al. 2009 article). A larger number of samples is needed to demonstrate that the difference observed between RST1 and RST2 strains is statistically significant, but there is a clear trend towards RST1 infections having the greatest association with antibiotic treatment failure and RST3 having the least, with RST2 having an intermediate effect. The duration of arthritis also depended on the infecting RST strain.


How do RST1 strains cause arthritis to persist even after the apparent eradication of the spirochetes by the recommended course of antibiotics? The investigators proposed that RST1 strains provoke a stronger inflammatory response in the joint than RST2 or RST3 strains. Coupled with an immune response to OspA_165-173 in genetically susceptible patients, this could cause inflammation to continue at high levels even after elimination of the spirochetes from the joints. RST1 strains may be more likely than the other genotypes to spark intense joint inflammation even in patients who are not genetically prone to antibiotic-refractory arthritis.

In future studies, it would be interesting to see if proinflammatory cytokine levels are related to the RST type that infects the joint. Ultimately, researchers need to identify the B. burgdorferi gene or genes whose variation among the RSTs causes the different treatment outcomes of Lyme arthritis.

Featured paper

Jones, K.L., McHugh, G.A., Glickstein, L.J., & Steere, A.C. (2009). Analysis of Borrelia burgdorferi genotypes in patients with Lyme arthritis: High frequency of ribosomal RNA intergenic spacer type 1 strains in antibiotic-refractory arthritis
Arthritis & Rheumatism, 60 (7), 2174-2182 DOI: 10.1002/art.24812

Other references

Drouin E.E., Glickstein, L., Kwok, W.W., Nepom, G.T., and Steere, A.C. (2008). Human homologues of a Borrelia T cell epitope associated with antibiotic-refractory Lyme arthritis. Molecular Immunology 45(1):180-189. DOI: 10.1016/j.molimm.2007.04.017

Kannian, P., Drouin, E.E., Glickstein, L., Kwok, W.W., Nepom, G.T., and Steere A.C. (2007). Decline in the frequencies of Borrelia burgdorferi OspA_161-175-specific T cells after antibiotic therapy in HLA-DRB1*0401-positive patients with antibiotic-responsive or antibiotic-refractory Lyme arthritis. The Journal of Immunology 179(9):6336-6342.

Shin J.J., Glickstein, L.J., and Steere, A.C. (2007). High levels of inflammatory chemokines and cytokines in joint fluid and synovial tissue throughout the course of antibiotic-refractory Lyme arthritis. Arthritis & Rheumatism 56(4):1325-1335. DOI: 10.1002/art.2241

Steere, A.C., Klitz, W., Drouin, E.E., Falk, B.A., Kwok, W.W., Nepom, G.T., and Baxter-Lowe, L.A. (2006). Antibiotic-refractory Lyme arthritis is associated with HLA-DR molecules that bind a Borrelia burgdorferi peptide. The Journal of Experimental Medicine 203(4):961-971. DOI: 10.1084/jem.20052471

Telomeres without telomerase in Borrelia spirochetes

You've all heard by now that the 2009 Nobel Prize in Physiology or Medicine will be awarded to Elizabeth Blackburn, Carol Greider, and Jack Szostak. They're the ones who figured out that an enzyme called telomerase combats the shortening that occurs at the ends of the linear chromosomes of eukaryotes (including ours) during each round of DNA replication. Telomerase sticks copies of a short string of nucleotides to the 3' ends of the chromosomal DNA. On the other hand, bacteria do not need telomerase because their chromosomes are circular; they do not have ends that can be shortened.

Spirochetes of the genus Borrelia, which include the agents of Lyme disease and relapsing fever, are an oddity in the bacterial world in that their chromosomes are linear. They also have a large set of linear plasmids. For example, the genome of the Lyme disease agent Borrelia burgdorferi consists of one linear chromosome and 12 linear plasmids, along with 11 circular plasmids. I will refer to the chromosomes and plasmids collectively as replicons. Despite having linear replicons, telomerase is nowhere to be found in Borrelia.

If they lack telomerase, how do Borrelia avoid having the ends of their linear replicons getting pruned during DNA replication? The key lies in the covalently closed hairpin ends of their linear replicons, something that's not found in eukaryotic telomeres. As illustrated in the figure below (figure 1 from Tourand 2003), the hairpin loops allow the replication machinery to copy the nucleotides at the very ends of the telomeres.

The replicated structure shown at the bottom of the figure illustrates that this method of replication creates a new problem. Following DNA replication, the two copies of the replicon end up fused at their ends (L'L and RR'). The fused sequences (telomere junctions) must somehow separate so that one copy of the replicon ends up in each daughter cell as the spirochete lengthens and splits in two.

Fortunately, Borrelia possesses an enzyme designated ResT, a telomere resolvase that cleaves the DNA where the two copies of the replicon are fused and reforms the hairpins at the ends of the new telomeres (see figure below). ResT was discovered by George Chaconas' laboratory in Canada. All of the work with ResT that I describe here was carried out by his group.


Chaconas' group was able to demonstrate the telomere resolvase reaction in vitro by simply mixing purified ResT with DNA containing a telomere junction. The products of the reaction were then analyzed following agarose gel electrophoresis. An example of the assay is shown below. A 4.6 kb piece of DNA containing the telomere junction formed by the left end of the B. burgdorferi linear plasmid lp17 ("L'L") is converted by ResT into the expected 2.6 and 2.0 kb products over a 2 minute time period.

Although the genome sequence of B. burgdorferi was published years ago, the sequence of the telomeres could not be determined back then because of the difficulty in cloning DNA with closed hairpin ends. The sequence of most of the telomeres finally appeared in the literature this year. The telomere sequences are aligned in the table below (figure 6 of Tourand 2009). The end of each telomere is the first nucleotide in each sequence.

The alignment shows that all of the telomeres have a "box 3" sequence. Box 3 is the recognition site for ResT binding. The evidence for this is that ResT binds to box 3 in vitro and changing the nucleotides in box 3 inhibits resolution of the telomere junction by ResT.

The telomeres were grouped based on the box 1 sequence. Type 1 telomeres carry the box 1 sequence TATAAT, and Type 2 telomeres harbor the modified box 1 sequence TATTAT. Type 3 telomeres lack the box 1 motif. When the rates of ResT resolving the different telomere junctions were measured in vitro (last column in alignment above), telomeres that lacked box 1 (Type 3) exhibited the slowest rates, with three telomeres failing to react with ResT. Since these three telomeres are obviously resolved in vivo, their resolution may require additional factors yet to be identified.

References

Tourand, Y., Deneke, J., Moriarty, T.J., & Chaconas, G. (2009). Characterization and in vitro reaction properties of 19 unique hairpin telomeres from the linear plasmids of the Lyme disease spirochete Journal of Biological Chemistry, 284 (11), 7264-7272 DOI: 10.1074/jbc.M808918200

Kobryn, K., & Chaconas, G. (2002). ResT, a telomere resolvase encoded by the Lyme disease spirochete Molecular Cell, 9 (1), 195-201 DOI: 10.1016/S1097-2765(01)00433-6

Tourand, Y., Kobryn, K., & Chaconas, G. (2003). Sequence-specific recognition but position-dependent cleavage of two distinct telomeres by the Borrelia burgdorferi telomere resolvase, ResT Molecular Microbiology, 48 (4), 901-911 DOI: 10.1046/j.1365-2958.2003.03485.x

Tourand, Y., Lee, L., & Chaconas, G. (2007). Telomere resolution by Borrelia burgdorferi ResT through the collaborative efforts of tethered DNA binding domains Molecular Microbiology, 64 (3), 580-590 DOI: 10.1111/j.1365-2958.2007.05691.x

Baby steps towards unraveling transcriptional regulation in the unculturable syphilis spirochete

I would never select Treponema pallidum as my experimental model if I had to study gene regulation in a spirochete. The main problem is that no one has figured out how to grow T. pallidum in any type of culture medium. T. pallidum can be propagated only by growing the spirochete in the testes of rabbits. Consequently, investigators have not even begun to develop the genetic tools (e.g., gene knock outs, shuttle plasmids) necessary to unravel the regulatory mechanisms that control T. pallidum gene expression.

Despite the limitations imposed by T. pallidum upon those who wish to study gene regulation, a group of syphilis researchers at the University of Washington in Seattle have started to dissect the regulation of several members of the 12-gene tpr (Treponema pallidum repeat) family. No one has figured out what the Tpr proteins do, but syphilis researchers are interested in them in part because they show how the immune response battles T. pallidum infections. For example, antibodies generated against TprK during infection bind to TprK exposed on the surface of T. pallidum and mark them for destruction by macrophages. More recent studies suggest that TprK undergoes antigenic variation (a topic of a future post), which may allow T. pallidum to persist in the host.

The Seattle group's studies on gene regulation have focused on the Subfamily II tpr genes tprE, tprG, and tprJ, as reported in the journal Molecular Microbiology. The sequences upstream of their transcription start sites contain a sequence that closely matches the consensus binding sequence for the E. coli global transcriptional regulator CRP (cAMP regulatory protein), also known as CAP (catabolite activator protein). The T. pallidum genome encodes a CRP homolog designated TP0262. In E. coli and a few other Gram negatives, CRP is an integral component of the complex network of transporter, regulatory, and enzymatic proteins that allow bacteria to selectively metabolize the preferred sugar, usually glucose, from those available in the environment. When glucose is absent, the enzyme adenylate cyclase is activated and synthesizes the second messenger cAMP (cyclic AMP), which turns on CRP by allosteric activation. (Here's a nice description of the allosteric activation of CRP.) The cAMP-CRP complex then binds upstream of various promoters and activates transcription by recruiting RNA polymerase to the promoter. Additional layers of regulation ensure that the genes are transcribed only when the sugar that is to be broken down by the gene products is present.

Because it's not possible to examine gene regulation in T. pallidum, the Seattle group transferred the tpr genes to E. coli, a genetically pliable bacterium. They fused each tpr gene, including the upstream sequences containing the proposed CRP binding site and the promoter, to a gene whose product is easily measurable, green fluorescent protein (gfp). They then introduced the plasmid carrying the gene fusion into an E. coli strain missing its crp gene so that they could measure tpr-driven GFP levels in the presence and absence of a second plasmid expressing TP0262. They found that TP0262 increased tprE'-gfp and tprJ'-gfp fusion expression while decreasing trpG'-gfp expression. The ability of TP0262 to control tpr'-gfp expression was lost when the CRP binding site was removed from the fusion constructions. They also showed that control of the tprJ'-gfp fusion by TP0262 was lost when the adenylate cyclase gene in E. coli was removed, indicating that cAMP was needed to activate TP0262 (data for tprE and tprG were not presented). Their in vitro experiments demonstrated binding of purified recombinant TP0262 to the proposed CRP binding site upstream of the three tpr genes by DNase I protection and gel shift assays.

What was missing from the study, as acknowledged by the authors, were experiments to demonstrate that TP0262 does the same thing in T. pallidum. For future studies, they plan to show that TP0262 is bound upstream of the Subfamily II tpr genes in T. pallidum by chromatin immunoprecipitation, which entails determining the sequence of the segment of DNA that is bound when TP0262 is immunoprecipitated from a T. pallidum extract. Such experiments would not require genetic manipulation or the ability to cultivate T. pallidum. It would only require harvesting a large number of T. pallidum spirochetes from infected rabbits.

What signal does TP0262 respond to? Does it respond to the glucose found in the host? The insightful Commentary by Radolf and Desrosiers sheds some light on the question. They note that T. pallidum is missing the special transporter genes that in E. coli encode the components necessary to link sugar availability to cAMP and CRP. They surmise that TP0262 has thus been freed to regulate genes not related to sugar metabolism, such as the tpr genes. Since CRP is a global transcriptional regulator in other bacteria, it is likely to regulate expression of not only the Subfamily II tpr genes but also additional genes in T. pallidum.

Near the end of their commentary, Radolf and Desrosiers made one comment that stood out:

One of the most important outcomes of the present study is that it will help put to rest the pregenomic view of the syphilis spirochaete as a transcriptionally invariant organism.

Maybe I'm too young to appreciate their point, but I can't believe that there ever was a time when syphilis researchers believed that T. pallidum genes were not regulated!

Featured articles

Giacani, L., Godornes, C., Puray-Chavez, M., Guerra-Giraldez, C., Tompa, M., Lukehart, S.A., & Centurion-Lara, A. (2009). TP0262 is a modulator of promoter activity of tpr Subfamily II genes of Treponema pallidum ssp. pallidum
Molecular Microbiology, 72 (5), 1087-1099 DOI: 10.1111/j.1365-2958.2009.06712.x

Radolf, J.D., & Desrosiers, D.C. (2009). Treponema pallidum, the stealth pathogen, changes, but how?
Molecular Microbiology, 72 (5), 1081-1086 DOI: 10.1111/j.1365-2958.2009.06711.x

Protein census of Leptospira interrogans

A census of proteins in a bacterial cell was conducted for the first time ever. By "census," I don't mean merely identifying all cellular proteins (which can be accomplished by shotgun tandem mass spectrometry). What I mean is counting the number of copies of every protein. The bacterium targeted for the census was the spirochete Leptospira interrogans. Like the census conducted here in the U.S. every ten years, some proteins were missed. The strategy developed by Malmström and colleagues, as described in the August 6 issue of Nature, allowed them to determine the abundance of 1,864 (or 83%) of the 2,221 proteins that were detectable by tandem mass spectrometry (MS) in Leptospira interrogans that had been grown in standard Leptospira culture medium.

The results of the protein census are compiled in the bar graph below. Proteins with related biological functions were grouped together and are color coded. The "Proteome" bar tabulates the number of different proteins in each group.
The next bar, "Copies per cell (control)," gives you an idea of how much of the protein expression machinery in L. interrogans is directed towards the synthesis of proteins in each functional category. The percentage reflects the amounts and size of the proteins in each category. For example, proteins of unknown function (hypothetical proteins) represent only 12.7% (blue) of the total protein synthesis capacity even though they constitute 30% of the identified proteins and over 40% of all genes in L. interrogans. I would surmise that these hypothetical proteins would account for a more sizable fraction of total protein synthesis under some other condition that L. interrogans would encounter during its life cycle (e.g., during infection).

The other observation noted by the authors is that L. interrogans gears 15% of its protein synthesis effort to make a small number of proteins deemed to be components of the "external encapsulating structure" (green), which is a fancy Gene Ontology term encompassing abundant Leptospira proteins that have been demonstrated to be in the inner or outer membrane. Most of the 15% is accounted for by five outer membrane proteins: LipL32, Loa22, LipL41, LipL21, and LipL36, the functions of which are not entirely clear. The five proteins are among the 10 most abundant proteins in L. interrogans.

The last bar shows the effect of the antibiotic ciprofloxacin ("cipro") on global protein levels in L. interrogans. The most striking change is the massive increase in 15 proteins of unknown function (light blue) leading them to constitute ~20% of the total protein content. As ciprofloxacin is an inhibitor of DNA gyrase, transcription of the genes encoding the 15 proteins may be extremely sensitive to DNA topology.

Did the enormous increase in the copy number of the 15 proteins following ciprofloxacin treatment increase the total number of protein molecules in L. interrogans? The authors found little change in the total cellular protein content:

Interestingly, this large redistribution of the proteome did not significantly change the total cellular protein concentration. Therefore, the large increase in the abundance of [the 15 proteins of unknown function] after ciprofloxacin exposure was compensated by a slight reduction of other high abundant protein classes.... This indicates that in L. interrogans, the cells strive to maintain a certain total number of protein components, that is, a constant cellular proteome concentration.

Featured paper

Malmström, J., Beck M., Schmidt, A., Lange, V., Deutsch, E.W., and Aebersold, R. (2009). Proteome-wide cellular protein concentrations of the human pathogen Leptospira interrogans. Nature 460(7256):762-765. DOI: 10.1038/nature08184

Zebrafish model of leptospirosis: Where's the relevance?

Scrutinized for the past several decades as a model of embryonic development, the zebrafish has recently been promoted as a vertebrate model for investigating the pathogenesis of infectious diseases. Zebrafish embryos are transparent, allowing microbiologists to readily view the course of infections in real time. Another advantage of the zebrafish model is that it's amenable, at least in theory, to large-scale genetic screens for mutated host or microbial genes that affect the infection process.


Davis and colleagues observed the early stages of zebrafish embryo infection by the spirochete Leptospira interrogans, as described in a recent issue of PLoS Neglected Tropical Diseases. They injected 10-100 spirochetes into the hindbrain of zebrafish embryos at 30 hours post-fertilization, when the innate immune response is fully functional. Macrophages rushed to the hindbrain and engulfed the invaders within the first 4 hours following inoculation of the spirochetes. L. interrogans was also rapidly phagocytosed when injected into the caudal vein.

Figure 1a from Lesley and Ramakrishnan, 2008. A zebrafish embryo 30 hours following fertilization. The hindbrain and caudal vein are indicated with the bracket and arrow, respectively.

Macrophages typically kill and destroy their prey following phagocytosis. However, spirochetes were still observed in the macrophages 24 hours following inoculation, suggesting that L. interrogans can survive inside macrophages.

Their most striking observation was the accumulation of infected macrophages near the dorsal aorta in a region known as the aorta-gonad-mesonephros (AGM), where hematopoietic stem cells are born. This was not a general phenomenon of bacterial infections as macrophages harboring Pseudomonas aeruginosa failed to accumlate at the AGM following its injection into zebrafish embryos.

Figures 2C and 2D from Davis et al., 2009. The embryo was infected with fluorescently stained L. interrogans. 24 hours following infection, most of the spirochetes were found near the dorsal aorta (brackets), with a few scattered around the ventral tail. Scale bar, 300 µm in panel C, 100 µm in panel D.

Here's what the authors concluded in the final sentence of the paper:

The strikingly specific delivery of leptospires to [the AGM] by phagocytes provides insights into pathogenesis by suggesting a novel mechanism for targeting of organs during leptospiral dissemination.

In other words, L. interrogans may be capable of steering macrophages towards specific organs. Once the macrophages arrive at their destination, the spirochetes may escape from the macrophage and colonize the organ.

The challenge for the authors in future studies will be to demonstrate the relevance of the zebrafish model to leptospirosis. Hamsters and guinea pigs are appropriate models for leptospirosis because the pathology and lethality of Leptospira infection in these rodents is similar to what's observed in human leptospirosis patients. The fate of Leptospira that macrophages capture in these rodents differs from what is seen in zebrafish embryos. Leptospira that are found inside macrophages in tissue sections from infected rodents often appear to be disintegrating. Nevertheless, it's possible that a few Leptospira survive phagocytosis and subsequently guide the macrophage towards the target organs.

Featured paper

Davis, J.M., Haake, D.A., & Ramakrishnan, L. (2009). Leptospira interrogans stably infects zebrafish embryos, altering phagocyte behavior and homing to specific tissues PLoS Neglected Tropical Diseases, 3 (6) DOI: 10.1371/journal.pntd.0000463

Other references

Lesley, R. and Ramakrishnan, L. (2008). Insights into early mycobacterial pathogenesis from the zebrafish. Current Opinions in Microbiology 11(3):277-283. DOI: 10.1016/j.mib.2008.05.013

The 14th species of the Lyme disease group of Borrelia

The cluster of genetically related Borrelia species that includes the Lyme disease spirochete B. burgdorferi is known in the scientific community as Borrelia burgdorferi sensu lato ("in the wider sense"). Three members of B. burgdorferi sensu lato account for most cases of Lyme disease worldwide. They are B. burgdorferi sensu stricto ("in the stricter sense"), B. garinii, and B. afzelii. Several other species of the cluster are suspected of causing a small number of Lyme disease cases in Europe and Asia.

The United States is home to at least four species of B. burgdorferi sensu lato. They are B. burgdorferi sensu stricto (the only species known to cause Lyme disease in the U.S.), B. bissettii, B. andersonii, and B. californiensis. The discovery of a fifth named U.S. species, christened Borrelia carolinensis, was published in the Journal of Clinical Microbiology earlier this year. The new species hails from South Carolina. Most of the isolates were cultured from cotton mice and eastern wood rats, but one isolate was obtained from an Ixodes minor tick feeding on an eastern wood rat. Whether B. carolinensis is capable of inducing Lyme disease is unknown.

The number of named species of B. burgdorferi sensu lato found worldwide now stands at 14:

• B. burgdorferi sensu stricto
• B. garinii
• B. afzelii
• B. andersonii
• B. bissettii
• B. californiensis
• B. carolinensis
• B. japonica
• B. lusitaniae
• B. sinica
• B. spielmanii
• B. tanukii
• B. turdi
• B. valaisiana

If you read my last post, you will notice that Borrelia lonestari, detected in one case of the Lyme-like illness STARI, is missing from the list. Although STARI clinically resembles a mild form of Lyme disease, genetically B. lonestari is more closely related to the set of Borrelia that causes relapsing fever.

Featured article

Rudenko, N., Golovchenko, M., Grubhoffer, L., and Oliver, J.H. (2009). Borrelia carolinensis sp. nov., a new (14th) member of the Borrelia burgdorferi sensu lato complex from the southeastern region of the United States. Journal of Clinical Microbiology 47(1):134-141. DOI: 10.1128/JCM.01183-08

STARI or Masters disease: More like Lyme than Lyme?

A tick-borne illness has been masquerading as Lyme disease in the southern United States over the past two decades. Victims first notice the expanding "bulls-eye" skin rash that is similar in appearance to the erythema migrans (EM) of Lyme disease. However, the tick that feeds on the victim is not the Ixodes tick that causes Lyme disease but the Lone Star tick Amblyomma americanum. Moreover, Borrelia burgdorferi, the Lyme disease spirochete, is not the infectious agent. B. burgdorferi has never been successfully cultured from a southern case of the EM-like rash, and sera from most of these patients test negative for Lyme disease by CDC criteria. Lyme disease itself is uncommon in the south as the resident Ixodes ticks rarely feed on humans; most ticks found attached to humans residing in the south are the Lone Star tick, which is unlikely to harbor or transmit B. burgdorferi.

The name bestowed upon the condition, "southern tick-associated rash illness" or STARI, is misleading because the territory of the Lone Star tick has been creeping into the northeastern and northern U.S., where Lyme disease is hyperendemic. The illness has also been dubbed "Masters disease" to honor Dr. Edwin Masters, who passed away last month. Dr. Masters' observations of skin rash patients in his private practice in Cape Girardeau, Missouri sparked the contentious CDC investigation that led to the first detailed description of STARI 14 years ago. You can read about his battles with the CDC in a series of blog posts by Pamela Weintraub, author of the book Cure Unknown, Inside the Lyme Epidemic (a book I hope to read some day).

Fig. 1 from Masters et al., 2008. Lone Star tick territory in green.

Contrary to popular belief, the erythema migrans of most Lyme disease patients does not present as a bull's-eye. In fact, in one study the EM-like rashes in Masters' STARI patients were much more likely to appear as a bull's-eye than the EM of Lyme disease patients from New York. In addition, the STARI patients were less likely than those with Lyme disease to suffer from accompanying symptoms such as joint and muscle aches, fatigue, headache, and stiff neck. A question that remains unresolved is whether arthritic, neurologic, or cardiac symptoms can crop up later, as they do in those afflicted with Lyme disease.

Figure 2 from Masters et al., 2008. EM-like skin lesions in Missouri STARI patients

The agent of STARI has eluded scientists. Borrelia lonestari was suspected at one time when it was detected by PCR in one patient and the Lone Star tick attached to his skin. (The spirochete could not be cultured since it does not grow in Borrelia culture medium.) However, B. lonestari could not be detected in a later study of a series of Masters' STARI patients. Thus, B. lonestari is unlikely to bring about most cases of STARI. The failure to identify the infectious agent of STARI has led some to question whether STARI has an infectious cause.

Masters was convinced that a spirochete, perhaps one closely related to Borrelia burgdorferi, was the agent of STARI. He has offered the following observations as evidence:

• Spirochetes have been observed in Lone Star ticks.
  
• Forms resembling spirochetes have been observed by silver staining of the EM-like skin lesions from Masters' STARI patients (see figure below).
• Extracts of B. burgdorferi reacted with sera from some STARI patients in ELISA tests, although the sera were Western blot negative according to CDC criteria. This observation indicates that antibodies were elicited against proteins closely related to those found in B. burgdorferi.
  Figure 8 from Masters et al., 1998. Silver stain of skin biopsy of an EM-like rash from a Missouri patient showing an apparent spirochete.
  

Of course these observations are far from definitive proof. Nevertheless, it is possible that a Borrelia spirochete is the culprit of EM-like lesions in the southern U.S. Like B. lonestari, some of these strains may not grow in the standard Borrelia medium used to cultivate B. burgdorferi from the skin rashes of Lyme disease patients. This would account for the inability of investigators to culture the spirochete from the skin rash of STARI patients.

Finally, how is STARI treated? Although the cause of STARI remains unknown, Edwin Masters declared that Lyme-like illness deserves Lyme-like treatment. That is, he recommended that antibiotics be administered to STARI patients according to Lyme treatment guidelines. Establishing whether antibiotics truly help will require a randomized placebo-controlled study.

Featured article

MASTERS, E.J., GRIGERY, C.N., & MASTERS, R.W. (2008). STARI, or Masters Disease: Lone Star Tick–Vectored Lyme-like Illness Infectious Disease Clinics of North America, 22 (2), 361-376 DOI: 10.1016/j.idc.2007.12.010

Other references

Masters, E., Granter, S., Duray, P., and Cordes P. (1998). Physician-diagnosed erythema migrans and erythema migrans-like rashes following Lone Star tick bites. Archives of Dermatology 134(8):955-960.

Wormser G.P., Masters, E., Liveris, D., Nowakowski, J., Nadelman, R. B., Holmgren, D., Bittker, S., Cooper, D., Wang, G., and Schwartz, I. (2005). Microbiologic evaluation of patients from Missouri with erythema migrans. Clinical Infectious Diseases 40(3):423-428. DOI: 10.1086/427289

Wormser G.P., Masters, E., Nowakowski, J., McKenna, D., Holmgren D., Ma, K., Ihde, L., Cavaliere, L.F., and Nadelman, R.B. (2005). Prospective clinical evaluation of patients from Missouri and New York with erythema migrans-like skin lesions. Clinical Infectious Diseases 41(7):958-965. DOI: 10.1086/432935

Role of the second messenger cyclic diguanylate (c-di-GMP) in the Lyme disease spirochete

Second messengers are the intracellular intermediaries that transmit the signals received from the environment (first messenger) to the cellular machinery that generates the appropriate response. Well known examples of second messengers in mammalian cells include cyclic AMP, cyclic GMP, calcium ion, and inositol triphosphate. A second messenger unique to bacteria is cyclic diguanylate, abbreviated c-di-GMP. First described in the 1980s, c-di-GMP is only now attracting wide interest among those who study signal transduction in bacteria.

Cyclic di-GMP is created from two GTP molecules by diguanylate cyclase and destroyed by phosphodiesterases. Genes encoding the opposing enzymatic activities can be identified by the conserved GGDEF motif in diguanylate cyclases and an EAL or HD-GYP motif in phosphodiesterases. Bacteria modulate the intracellular concentration of c-di-GMP by controlling the amounts and activities of the diguanylate cyclases and phosphodiesterases in response to changes in environmental conditions.

Cyclic di-GMP is best known for promoting the formation of biofilms. Biofilm assembly requires the synthesis and secretion of the special polysaccharides that make up the biofilm matrix and the down-regulation of motility. Both polysaccharide synthesis and the inhibition of motility are modulated by c-di-GMP. The molecule can also affect virulence functions. In many cases, the mechanistic details of how c-di-GMP exerts its effects remain unknown. The molecular target of c-di-GMP includes proteins with the "PilZ" domain (see figure). However, not all proteins bound by c-di-GMP possess the PilZ domain. In some bacteria, c-di-GMP can also bind specific sequences found within the 5' untranslated region of several mRNAs to modulate gene expression.

Figure 1 from Tamayo et al., 2007. DGC, diguanylate cyclase; PDEA, phosphodiesterase A; PDE, phosphodiesterase

The Borrelia burgdorferi gene rrp1 encodes the only protein in the Lyme disease spirochete containing the GGDEF motif. The Rrp1 protein consists of a receiver domain and the GGDEF domain, whose diguanylate cyclase activity requires phosphorylation of the receiver domain. A paper in the March issue of Molecular Microbiology revealed the genes whose expression is affected by Rrp1. The authors compared the transcript profiles (transcriptome) of a B. burgdorferi wild-type and an rrp1 deletion mutant by microarray analysis. It turned out that most of the genes affected by the mutation encode what the authors call the "core" cellular functions of Borrelia burgdorferi. The core functions allow the spirochete to seek out and capture nutrients from the environment, synthesize the building blocks necessary for assembling cellular parts, and extract energy from nutrients to fuel its activities. All bacteria, whether or not they cause disease, possess these core functions. Most transcripts from core genes were increased in the wild-type B. burgdorferi strain relative to the rrp1 mutant. The impaired growth of the rrp1 mutant compared to wild type is consistent with the importance of Rrp1 on the expression of the core functions of B. burgdorferi.

The investigators also found that the rrp1 transcript levels increased 6 fold when ticks harboring B. burgdorferi took a blood meal from mice. High levels of rrp1 mRNA were also maintained in B. burgdorferi growing in culture medium. These observations suggest that more rrp1 transcript is made when the spirochete is awash in nutrients, whether in blood or culture medium. Under these conditions, c-di-GMP signals B. burgdorferi to turn on genes necessary to acquire and metabolize the nutrients.

The protein that directly or indirectly senses changes in nutrient availability is likely to be the histidine kinase encoded by hpk1, the gene that lies immediately upstream of rrp1. The Hpk1 and Rrp1 proteins form a phosphorelay in which phosphate groups swiped from ATP molecules are transferred to the target Rrp1 protein in response to some signal in the environment.

In summary, Rrp1 diguanylate cyclase activity is enhanced at two levels when nutrients become abundant. First, rrp1 transcript levels are increased, leading to more Rrp1 protein being made. Second, the Rrp1 diguanylate cyclase activity is activated by phosphorylation. Increased c-di-GMP levels is the result. The c-di-GMP stimulates increased levels of transcripts emanating primarily from genes encoding the core cellular functions of B. burgdorferi. The question that remains unexplored is how c-di-GMP causes transcript levels to increase. One protein in B. burgdorferi harbors the PilZ domain, but as I mentioned earlier, PilZ is not the only protein domain capable of binding c-di-GMP.

Finally, what does this study reveal about the role of Rrp1 and c-di-GMP in Lyme disease? One possibility is that Rrp1 is involved in tick-to-human transmission and the early stages of the infection:

• As I already mentioned, Rrp1 upregulation in B. burgdorferi residing in a tick taking a blood meal may prepare the spirochete to metabolize the nutrients found in the blood as they are transmitted into the skin of the human victim.
• Transcripts expressed from several genes encoding factor H-binding proteins (some of the few non-core genes affected by the rrp1 knock out) were at higher levels when Rrp1 was present. Factor H is an inhibitor of the complement system found in our bloodstream. As such, binding of factor H by the spirochete may protect it from being killed by the host complement system. Indeed, the authors demonstrated that the rrp1 mutant was more sensitive to human serum than the wild-type B. burgdorferi strain. However, the significance of this observation is unclear as Borrelia garinii, another agent of Lyme disease, was just as sensitive as the B. burgdorferi rrp1 mutant to human serum. Additionally, earlier studies have shown that factor H is not necessary for successful B. burgdorferi infections (at least in the mouse model of Lyme disease).
• The ospC gene, which encodes another protein that may impair immune function during the early stages of infection, was also upregulated by Rrp1.
• Several transcripts expressing motility and chemotaxis functions are expressed at higher levels when Rrp1 is present. Motility and chemotaxis are considered to be core functions, but they may also be necessary for B. burgdorferi to establish infection in humans. Note that the proposed effect of c-di-GMP on B. burgdorferi motility is opposite of that found in other bacteria (see figure above).
  

The obvious experiment to perform is to test whether the rrp1 mutant can cause infection in the mouse model of Lyme disease. Unfortunately, the effect of rrp1 on virulence could not be tested as the authors were unable to knock out the rrp1 gene in an infectious strain of B. burgdorferi.

Featured paper

Rogers, E.A., Terekhova, D., Zhang, H.-M., Hovis, K.M., Schwartz, I., & Marconi, R.T. (2009). Rrp1, a cyclic-di-GMP-producing response regulator, is an important regulator of Borrelia burgdorferi core cellular functions Molecular Microbiology, 71 (6), 1551-1573 DOI: 10.1111/j.1365-2958.2009.06621.x

Image source

Tamayo R., Pratt J.T., and Camilli, A. (2007). Roles of cyclic diguanylate in the regulation of bacterial pathogenesis. Annual Review of Microbiology 61:131-148. DOI: 10.1146/annurev.micro.61.080706.093426

Cholera and spirochetes: Introducing Brachyspira!

Cholera results in a severe form of diarrhea that can lead to dehydration, shock, and ultimately death without prompt treatment. The disease afflicts the poor in developing countries lacking clean water sources and sanitation infrastructure. Vibrio cholerae is the causative agent and can be viewed by microscopic examination of the so-called "rice-water" stool samples from cholera patients.

rice-water stool from a cholera patient (Figure 1 from Sack et al., 2004)

As reported in a recent issue of Emerging Infectious Diseases, Nelson and colleagues, while examining a cholera outbreak in Bangladesh back in 2006, found that stool samples in over a third of cholera patients contained spirochetes mingling with V. cholerae. Samples were fluorescently stained to aid identification of bacteria. One example is shown below. V. cholerae were visualized with a FITC-conjugated monoclonal antibody to its lipopolysaccharide (in red), and bacterial DNA was stained wtih DAP I (green). Only the merged image is shown below. V. cholerae are the rods with a slight bend and appear yellow (combination of red and green) with a red edge; the spirochetes are the W-shaped forms stained green.

bar = 10 µm

This wasn't the first time spirochetes were observed in rice-water stool. Over a century ago, Theodor Escherich (the discoverer of E. coli) was the first to witness spiral-shaped microbes in fecal samples from cholera victims.

What was the identity of these spirochetes? They were not any of the "Big 3" of Borrelia, Treponema, and Leptospira, which garner the most attention from spirochete researchers (and from the writer of this blog). Nelson and colleagues guessed that they were members of the genus Brachyspira as they are the only spirochetes known to live in the human intestine. They turned out to be correct. They successfully amplified the gene encoding the 16s rRNA with Brachyspira-specific PCR primers. The sequence of the PCR product revealed the spirochetes to be Brachyspira pilosicoli and Brachyspira aalborgi.

Brachyspira account for most cases of human intestinal spirochetosis, defined as the presence of spirochetes in the colon. Although colonization of the large intestine by spirochetes is uncommon in the Western world, up to half of those in developing nations may harbor intestinal spirochetes. A typical example is shown below (click on image for larger version).

Figure 3 from Esteve et al., 2006

The sectioned tissue, which was stained with H&E, was obtained by colonic biopsy. The left panel reveals a fuzzy layer covering the colonic epithelium. These are Brachyspira attached at one end to the lining of the colon. The density of spirochetes can reach up to 1,700 per square millimeter. The right panel shows a colonic biopsy from the same patient after successful treatment with the antimicrobial agent metronidazole. Note that the fuzzy layer has disappeared.

Whether intestinal spirochetes cause disease in humans is unclear. Many people with intestinal spirochetes do not suffer any ill effects, but others endure chronic diarrhea. The mode of transmission of Brachyspira is unknown, but scientists have surmised that ingestion of contaminated water is involved.

The role of Brachyspira in cholera, if any, is even more of a mystery. In the conclusion to their article, Nelson et al. present the hypothesis that intestinal spirochetes exacerbate the already devastating clinical course of cholera.

Featured paper

Nelson, E.J., Tanudra, A., Chowdhury, A., Kane, A.V., Qadri, F., Calderwood, S.B., Coburn, J., Camilli, A. (2009). High Prevalence of Spirochetosis in Cholera Patients, Bangladesh Emerging Infectious Diseases, 15 (4), 571-573 DOI: 10.3201/eid1504.081214

Other references

Esteve, M., Salas, A., Fernandez-Banares, F., Lloreta, J., Marine, M., Gonzalez, C.I., Forne, M., Casalots, J., Santaolalla, R., Espinos, J.C., Munshi, M.A., Hampson, D.J., and Viver, J.M. (2006). Intestinal spirochetosis and chronic watery diarrhea: Clinical and histological response to treatment and long-term follow up. Journal of Gastroenterology and Hepatology 21(8):1326-1333. DOI: 10.1111/j.1440-1746.2006.04150.x

Sack, D.A., Sack, R.B., Nair, G.B., and Siddique, A.K. (2004). Cholera. Lancet 363(9404):223-233. DOI: 10.1016/S0140-6736(03)15328-7

Leptospira heme oxygenase frees iron from heme

I mentioned in a recent post that iron is an essential trace metal that bacteria must acquire from its surroundings. (From that same post you will also recall that the Lyme disease spirochete B. burgdorferi is a rare exception that doesn't need iron.) Much of the iron in our body is trapped within the center of the heme molecule. Heme itself is not readily accessible as it is bound to host proteins such as hemoglobin. Pathogenic bacteria have evolved sophisticated systems to kidnap heme from host proteins and transport them into the cytoplasm. These complex systems, which include secreted degradative enzymes, heme capturing proteins, and transporter proteins that sit in the membrane, have been examined in numerous bacteria. However, the fate of heme after it is acquired by the bacteria is poorly understood. In some cases, the captured heme may be incorporated into bacterial proteins such as cytochromes, which participate in electron transport. In other cases, bacteria may need to extract the iron trapped in the middle of the heme molecule. Some bacteria possess the enzyme heme oxygenase, which extracts the iron caged within heme by the following reaction (adapted from Scheme 1 in Kikuchi et al., 2005):
Unlike its cousin that causes Lyme disease, the spirochete Leptospira requires iron for growth. Ben Adler's group at the Monash University in Australia isolated a Leptospira interrogans mutant with the transposon TnSC189 inserted into hemO, the gene encoding heme oxygenase. The properties of the hemO mutant is described in two papers in the journal Microbes and Infection. The graph below (figure 1B in Murray et al., 2008) shows that growth of the hemO mutant is impaired, although not completely, when hemoglobin is the sole source of iron in the culture medium. The residual growth of the mutant indicates that L. interrogans may possess another activity that extracts iron from heme.
In their follow-up study, Adler's group demonstrated that the hemO gene was necessary for L. interrogans to fully express its virulence in the hamster model of leptospirosis. For this study they used the hemO mutant and a control L. interrogans strain that had the TnSC189 element inserted in a noncoding region, presumably where gene expression would not be affected. The two strains were injected into the abdominal cavities of separate groups of hamsters, which were then monitored for 14 days. Only 8 of 24 hamsters (33%) survived the challenge with the control strain, whereas 20 of 24 (83%) injected with the hemO mutant survived. The difference in survival rates between the two groups was statistically significant (P = 0.001).

Although the hemO mutant was ineffective at killing hamsters, it was still able to colonize the kidneys of most of the animals. Colonization was assessed by culturing kidney or urine in Leptospira growth medium. The mutant was recovered by culturing of kidney or urine from 17 of 20 hamsters that survived the challenge with the hemO mutant and all 3 that died. These results were similar to what was obtained with hamsters inoculated with the control strain, which was recovered from all 8 animals that survived and all 12 that died. (Not all hamsters were examined for colonization.)

Why was the hemO mutant able to colonize the kidney when it was unable to extract iron from heme? Heme is not the only source of iron in the body. The mutant may have captured one of the other forms of iron present in the host. The genome of L. interrogans encodes several homologs of transporters that the spirochete may use to acquire non-heme sources of iron (Louvel et al., 2006). Since these other iron sources are less abundant than heme, the tissue burden (density of bacteria) of the mutant in the kidneys may have been lower than that of the control strain thereby allowing most of the hamsters challenged with the hemO mutant to survive.

One obvious limitation of the study is that the investigators did not attempt to complement the hemO mutation with a wild-type copy of the gene. However, I should point out that currently no plasmid is available that replicates in L. interrogans, rendering complementation of L. interrogans mutations difficult. The researchers did verify that the gene immediately downstream of hemO was still transcribed in the mutant.

This work is significant for the following reasons. First, although there have been two other studies that have examined the role of Leptospira genes in virulence, this study was the most satisfying to read because it was the first to show that a gene encoding a product of known function has a role in the virulence of Leptospira. Second and perhaps more importantly, it is the first to demonstrate the importance of a bacterial heme oxygenase in virulence.

Featured papers

Murray, G., Ellis, K., Lo, M., & Adler, B. (2008). Leptospira interrogans requires a functional heme oxygenase to scavenge iron from hemoglobin Microbes and Infection, 10 (7), 791-797 DOI: 10.1016/j.micinf.2008.04.010

Murray, G., Srikram, A., Henry, R., Puapairoj, A., Sermswan, R., & Adler, B. (2009). Leptospira interrogans requires heme oxygenase for disease pathogenesis Microbes and Infection, 11 (2), 311-314 DOI: 10.1016/j.micinf.2008.11.014

Other references

Kikuchi, G., Yoshida, T., and Noguchi, M. (2005). Heme oxygenase and heme degradation. Biochemical and Biophysical Research Communications 338(1):558-567. DOI: 10.1016/j.bbrc.2005.08.020

Louvel, H., Bommezzadri S., Zidane, N., Boursaux-Eude, C., Creno, S., Magnier, A., Rouy, Z., Médigue, C., Saint Girons, I., Bouchier, C., and Picardeau, M. (2006). Comparative and functional genomic analyses of iron transport and regulation in Leptospira spp. Journal of Bacteriology 188(22):7893-7904. DOI: 10.1128/JB00711-06