Email updates

Keep up to date with the latest news and content from BMC Evolutionary Biology and BioMed Central.

Open Access Research article

Evolution of the leukotoxin promoter in genus Mannheimia

Jesper Larsen1*, Anders G Pedersen2, Robert L Davies3, Peter Kuhnert4, Joachim Frey4, Henrik Christensen1, Magne Bisgaard1 and John E Olsen1

Author Affiliations

1 Department of Veterinary Pathobiology, Faculty of Life Sciences, University of Copenhagen, Stigbøjlen 4, DK-1870 Frederiksberg C, Denmark

2 Center for Biological Sequence Analysis, BioCentrum-DTU, Technical University of Denmark, Building 208, DK-2800 Lyngby, Denmark

3 Institute of Biomedical and Life Sciences, Glasgow Biomedical Research Centre, University of Glasgow, 120 University Place, Glasgow G12 8TA, UK

4 Institute of Veterinary Bacteriology, University of Berne, Länggass-Strasse 122, CH-3012 Berne, Switzerland

For all author emails, please log on.

BMC Evolutionary Biology 2009, 9:121  doi:10.1186/1471-2148-9-121


The electronic version of this article is the complete one and can be found online at: http://www.biomedcentral.com/1471-2148/9/121


Received:20 November 2008
Accepted:29 May 2009
Published:29 May 2009

© 2009 Larsen et al; licensee BioMed Central Ltd.

This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

Background

The Mannheimia species encompass a wide variety of bacterial lifestyles, including opportunistic pathogens and commensals of the ruminant respiratory tract, commensals of the ovine rumen, and pathogens of the ruminant integument. Here we present a scenario for the evolution of the leukotoxin promoter among representatives of the five species within genus Mannheimia. We also consider how the evolution of the leukotoxin operon fits with the evolution and maintenance of virulence.

Results

The alignment of the intergenic regions upstream of the leukotoxin genes showed significant sequence and positional conservation over a 225-bp stretch immediately proximal to the transcriptional start site of the lktC gene among all Mannheimia strains. However, in the course of the Mannheimia genome evolution, the acquisition of individual noncoding regions upstream of the conserved promoter region has occurred. The rate of evolution estimated branch by branch suggests that the conserved promoter may be affected to different extents by the types of natural selection that potentially operate in regulatory regions. Tandem repeats upstream of the core promoter were confined to M. haemolytica with a strong association between the sequence of the repeat units, the number of repeat units per promoter, and the phylogenetic history of this species.

Conclusion

The mode of evolution of the intergenic regions upstream of the leukotoxin genes appears to be highly dependent on the lifestyle of the bacterium. Transition from avirulence to virulence has occurred at least once in M. haemolytica with some evolutionary success of bovine serotype A1/A6 strains. Our analysis suggests that changes in cis-regulatory systems have contributed to the derived virulence phenotype by allowing phase-variable expression of the leukotoxin protein. We propose models for how phase shifting and the associated virulence could facilitate transmission to the nasopharynx of new hosts.

Background

The genus Mannheimia includes strains previously classified as trehalose-negative [Pasteurella] haemolytica and is one of the most well-defined and robust clusters within the gamma-proteobacterial family of Pasteurellaceae Pohl 1981 [1]. The Mannheimia species have taken divergent paths toward their distinct lifestyles. The majority of strains isolated from pulmonary infection in cattle belongs to M. haemolytica serotype A1/A6 [2]. These strains are sub-dominant to other serotypes (e.g., bovine serotype A2 strains) in the nasopharynx of healthy cattle but dominate when the host defences are at least partly compromised [3-5]. In the immunocompromised host, they have an increased capacity for proliferation and can achieve relatively high total numbers in the nasopharynx, where they are likely to be transmitted to the nasopharynx of new hosts or to spill over or otherwise enter the lungs [6]. However, pulmonary infection caused by M. haemolytica serotype A1 is considered to be non-communicable (i.e. no direct transmission between the lungs) and the continuous circulation of these bacteria in bovine populations seems to depend on their capacity for asymptomatic transmission to the nasopharynx, and not the lungs, of new hosts [7]. By contrast, bovine serotype A2 strains of M. haemolytica have had a long history in bovine populations, where they colonise the nasopharynx asymptomatically and rarely cause disease [2]. Also strains of M. glucosida serotype A11 and M. ruminalis, the sister group to M. haemolytica + M. glucosida, appear to be adapted to a relatively benign lifestyle in the ovine nasopharynx and rumen, respectively [1,2]. In addition, bovine strains of M. varigena biogroup 6, which is the most basal of the Mannheimia species, colonise the nasopharynx asymptomatically, although they have the propensity to cause disease [1]. It therefore seems reasonable to conclude that the genes responsible for the virulence of Mannheimia strains must evolve in response to the demands (selection pressures) associated with a commensal lifestyle rather than any advantages that might arise from causing disease.

Previous works have revealed that the leukotoxin (LktA) protein produced by M. haemolytica serotype A1 provides protection from the circulating constitutive and inducible immune defences during pulmonary infection via interactions with host cells [8-10]. Interestingly, much of the virulence can be blamed on the seemingly misguided overresponse of the immune defences [11]. The LktA protein induces the overwhelming activation of neutrophils leading to production of cytokines that mediate tissue injury [12-15].

The leukotoxin (lkt) operon codes for four proteins: an internal acyltransferase, encoded by lktC [16]; the structural toxin, encoded by lktA, which belongs to the Escherichia coli HlyA-like subfamily of cytotoxic RTX (repeats in toxin) proteins [16]; an inner membrane protein with a cytoplasmic ATP-binding cassette (ABC) domain, encoded by lktB, which pumps out the LktA protein via interaction with the C terminus of the LktA protein [17]; and a membrane fusion protein, encoded by lktD, which forms a bridge between the inner and outer membranes [17]. The genes for these four proteins are physically adjacent on the chromosome and are transcribed as lktCA or lktCABD messages [18,19].

Recent reports provided the first support for the view that the lkt operon was vertically inherited from the last common ancestor of genus Mannheimia to any ancestor of the diverging species, while changes such as chromosomal rearrangements, pseudogene formation, and deletions of various sizes have occurred in the region upstream of the lkt genes in individual subclades/species [20,21]. However, the extent to which these molecular evolutionary forces affect lkt gene regulation is not known.

In this article, we analysed the lkt promoter among representatives of the five Mannheimia species in order to identify changes in cis-regulatory systems that could have catalysed adaptive evolution. We started our study using multiple sequence alignment as a tool for the identification of orthologous sequences in the intergenic region upstream of the lkt genes (called hereafter the conserved promoter, CP). We also surveyed existing data on regulatory sequences in order to delineate the spatial and functional boundaries of the CP region. On the basis of the multiple sequence alignment, we estimated the rate of evolution for each CP sequence in order to infer the selective constraints. We finally consider how the evolution of the lkt operon fits with the evolution and maintenance of virulence.

Results and discussion

Table 1 presents the Mannheimia strains analysed in this study, sorted by their overall similarity to M. haemolytica serotype A1 str. PHL213. This measure of similarity is based on 16S rRNA phylogenies [1,20] and on relationships resolved by multilocus enzyme electrophoresis (MLEE) typing within M. haemolytica + M. glucosida [2].

Table 1. Strains used in this study

Conservation and flexibility in the intergenic region upstream of the lkt genes

The alignment revealed 100% identity over the entire 406-bp intergenic region from M. haemolytica serotype A1 str. PHL101 and str. PHL213 (data not shown) and previous reports [18,22,23] allowed us to infer cis-regulatory sequences. The 406-bp region was shared among M. haemolytica serotype A2 and M. glucosida serotype A11 as reflected in Figure 1. The alignment showed significant sequence and positional conservation over a 225-bp stretch immediately proximal to the transcriptional start site of the lktC gene among all Mannheimia strains (CP region), with the notable exception of the region corresponding to the repeat tract in M. haemolytica serotype A1 [22], which accommodated many sequence variations (Figures 1 and 2).

thumbnailFigure 1. Structure and evolution of the intergenic regions upstream of the lkt genes. Cladogram that describes bifurcation order is based on 16S rRNA phylogenies and MLEE. Greek letters indicate individual branches. The CP region (sequence shared among all Mannheimia strains) is hatched. The site of the repeat sequences in the CP region is indicated by an interruption in hatched box and the number of imperfect and perfect repeat units per promoter is indicated by red and green blocks, respectively. The ancient lapB region is presented by fountain filled box. Individual noncoding regions upstream of the CP region in M. haemolytica + M. glucosida and M. granulomatis are shown by black and white boxes, respectively. Arrows denote the putative transcription start sites. Mh, M. haemolytica: Mgl; M. glucosida; Mr, M. ruminalis; Mgr, M. granulomatis; Mv, M. varigena.

thumbnailFigure 2. Alignment of the CP region. The alignment of the intergenic regions upstream of the lkt genes showed significant sequence and positional conservation over a 225-bp stretch immediately proximal to the transcriptional start site of the lktC gene among all Mannheimia strains (CP region), with the notable exception of the region corresponding to the repeat tract in M. haemolytica serotype A1 [22], which accommodated many sequence variations. Capital letters denote aligned residues. Lower-case letters are not considered to be aligned. Numbers (0–9) below the alignment reflect the relative degree of local similarity among the sequences. Previous works [18,22,23] allowed us to infer cis-regulatory sequences. The footprinted palindromic motifs are indicated by an over line. The repeat sequences are enclosed by a rectangle. The core promoter including the -10 hexamer, the extended TG motifs, and the transcription start site (+1) is highlighted. Mh, M. haemolytica: Mgl; M. glucosida; Mr, M. ruminalis; Mgr, M. granulomatis; Mv, M. varigena.

In M. haemolytica + M. glucosida, the distal 180-bp stretch of sequence was juxtaposed to the transcriptional start site of the artJ gene (Figure 1). No homology was detected between this 180-bp stretch and the noncoding regions upstream of the CP region in the remaining Mannheimia species (data not shown).

In order to gain some insight into the possible evolutionary mechanisms of this differential pattern, we performed for each species analysis of sequence and positional conservation in the noncoding region upstream of the CP region. In M. ruminalis and M. varigena, these regions contained stretches of sequence unique to the lapB genes (data not shown). The region of M. granulomatis was adjacent to the xylB pseudogene and contained a fragment of about 620 bp corresponding to the distal region of the parental xylB gene as described elsewhere [21]. Moreover, this region contained a 50-bp stretch of inverted sequence that was homologous to the hslV-hslU intergenic regions and stretches of sequence unique to the lapB genes (data not shown).

These observations provide support for the view that the artJ operon, including 180-bp of sequence upstream of the transcribed artJ gene, was translocated into the intergenic region upstream of the lkt genes during early evolution of M. haemolytica + M. glucosida and that the hslVU and lapB operons has undergone pseudogene formation and deletions in M. granulomatis after divergence from the remaining Mannheimia species [21].

These observations illustrate how a few evolutionary events in closely related strains can alter dramatically an intergenic region. In the case of M. haemolytica + M. glucosida, the alteration coincided with an evolutionary success. Interestingly, transcriptional coupling has been reported for the divergently transcribed artJ and lkt promoters [23]. However, the role of this regulatory feature during transmission of bacteria through host populations is not known. The chimeric promoter was found in all M. haemolytica + M. glucosida strains in the present study, irregardless of their propensity to cause disease, suggesting that it evolved in response to selection for something else than making their hosts sick.

Evolutionary rates for CP sequences

The role for natural selection in the evolution of regulatory regions is quite unclear [24]. Recognisable features and motifs are often labile, small in size, and may depend on the sequence context [25,26]. Therefore, the complex structure of these regions could be due to alternative evolutionary forces [27-30]: (i) compensatory selection occurring when a pair of mutations at different sites that would be singly deleterious produces normal fitness in combination; (ii) stabilising selection could maintain stable levels of gene expression while allowing mutational turnover of functional important sites; (iii) purifying selection could facilitate maintenance of specific functional sites.

The evolutionary rates for CP sequences were tested for deviations from expectations derived from the null hypothesis of neutral evolution. The overall length of the 16S rRNA tree (obs. T16S) was 0.102024 substitutions per site, while the overall length of the CP tree (obs. TCP) was 1.33986 substitutions per site. The null hypothesis predicts that the expected length of any given branch on the CP tree (exp. vCP) can be calculated by multiplying the observed length of the corresponding branch on the 16S rRNA tree (obs. v16S) by 1.33986/0.102024 = 13.13292. The observed and the expected lengths for any given branch on the CP tree and their pairwise differences are summarised in Table 2.

Table 2. Evolutionary rate differences for CP sequences

For three of four branches ancestral to M. haemolytica serotype A1, the differences between the observed and the expected branch lengths were significantly smaller than zero (the value expected under the null hypothesis of neutral evolution) (Table 2). This is in accord with the operation of stronger levels of purifying selection and possibly reflects an ancient balance between the immune system and these primarily commensal bacteria and their expression of the LktA protein during asymptomatic colonisation of the nasopharynx. The difference between the observed and the expected branch lengths ancestral to M. haemolytica + M. glucosida was not significantly different from zero (Table 2). This coincided with the translocation of the artJ operon, including 180-bp of sequence upstream of the transcribed artJ gene, into the intergenic region upstream of the lkt genes [21]. It is possible that this relatively short window of time was dominated by stabilizing selection on this chimeric region, maintaining stable levels of lkt (and artJ) expression and thus the ancient balance between the immune system and bacteria during asymptomatic colonisation of the nasopharynx.

The difference between the observed and the expected branch lengths ancestral to M. granulomatis was significantly greater than zero (Table 2). Moreover, the CP region appeared to be more variable in M. granulomatis, especially in terms of the overall conservation of cis-regulatory sequences (Figure 2). Most notably, there was a high level of sequence degeneracy in the core promoter, which was highly conserved among the other Mannheimia species. Strains belonging to the bovine taxon of M. granulomatis, including M. granulomatis str. P1135/26T, cause severe skin infection (lechiguana) [1,31,32]. Interestingly, these strains have never been isolated from healthy cattle, including the nasopharynx, and the disease seems to be acquired by contact with the human botfly (Dermatobia hominis), i.e. there is no direct transmission between cattle or transmission is short-lived [33]. Moreover, comparison of lktA genes from different taxa within M. granulomatis have shown that those from bovine strains are under significantly relaxed selective constraints (J. Larsen, unpublished results). We propose that genetic diversity of the CP sequence in this strain reflects functional decay due to relaxed selection outside the nasopharynx, enabling mutations to accumulate in the portions of the genome that encode unused functions in the integumentary system. However, it is also possible that positive Darwinian selection for alternative cis-regulatory sequences has acted as a mechanism for coordinating expression of the LktA protein with the nutritional and environmental conditions of the integument. For example, Larsen et al. [20] have shown that the β-haemolytic phenotype, a marker of the lktA genotype in M. haemolytica [34], is present in this strain, suggesting at least some functional conservation of the lkt operon.

Evolution of repeat sequences

Tandem repeats of 5'-ACAAAAAACA-3' upstream of the core promoter were first identified in M. haemolytica serotype A1 str. PHL101 [22]. Our preliminary analysis revealed three and two repeat units per promoter in M. haemolytica serotype A1 and serotype A2, respectively (Figure 1). In the latter case, the two repeat units varied from the consensus sequence by one nucleotide change each (Figure 2). Because this region accommodated many sequence variations in the remaining Mannheimia species, the unit sequence could not be determined, suggesting that repeat sequences are confined to M. haemolytica.

The relatively recent appearance of three repeat units (at least after the divergence of M. haemolytica serotype A1 and serotype A2 as reflected in Figure 1) may provide us with an opportunity to trace the entire evolutionary process back to its origin. Therefore, we examined the number of repeat units per promoter in 28 additional M. haemolytica strains, which represent the diversity within this species based on MLEE, geographic origin, and host association (Table 3). The distribution of the number of repeat units per promoter showed that there were three or more repeat units in 14 out of 16 M. haemolytica strains belonging to MLEE lineage A (Table 3). In the two other strains belonging to MLEE lineage A as well as strains belonging to MLEE lineages B (n = 7) and C (n = 5), there were only two repeat units per promoter (Table 3).

Table 3. Distribution of repeat tract polymorphisms in clonally distinct M. haemolytica lineages

The phylogenetic tree of the distal two repeat units from M. haemolytica showed a strong association between the sequence of the repeat units, the number of repeat units per promoter, and the phylogenetic history of this species (Figure 3). First, repeat tracts containing three or more repeat units were restricted to MLEE lineage A as described above and belonged to a cluster (I) of perfect repeat units. Second, repeat tracts containing two repeat units were localised in two clusters (II and III) according to the number of repeat units that varied from the consensus sequence (one and two, respectively) and to the MLEE lineage (B and C, respectively) (Figure 3). These observations are consistent with the generally accepted view that tandem arrays of perfect repeats, such as those present in M. haemolytica strains belonging to MLEE lineage A, are hotspots for replication errors, resulting in high rates of expansions/contractions.

thumbnailFigure 3. Evolution of the repeat sequences in M. haemolytica. Maximum-parsimony tree of the two distal repeat units from each promoter (A) and repeat tract polymorphisms (B) using M. haemolytica strains, which represent the diversity within this species based on MLEE, geographic origin, and host association. See the text for clusters I, II, and III. The green and red boxes highlight polymorphic sites that match and differ, respectively, from the consensus sequence. The number of repeat units per promoter is indicated. Sequence names contain the strain ID and abbreviations of the corresponding MLEE lineage (A, B, or C).

Role of repeat sequences

M. haemolytica, like other usually commensal bacteria, must overcome a variety of environmental hurdles to successfully colonise a host, evade the constitutive and inducible defences of this host, migrate from one within-host habitat to another, and be transmitted to another genetically and immunologically distinct host. Presumably to meet these challenges, some bacteria have highly mutable contingency genes that undergo phenotypic switching (phase shifting) [35].

Although the LktA protein plays a role in evading host immunity in the lungs, it may also function as a nasopharyngeal colonization and survival factor. We propose that, in response to the selective pressures associated with the host immune system, expression of the LktA protein is maintained at a baseline level during nasopharyngeal colonisation within individual hosts and asymptomatic nose-to-nose transmission between hosts, while expression is above the baseline level during pulmonary infection.

We further propose that expansions/contractions in the number of perfect repeat units in the lkt promoter are associated with phase-variable expression of the LktA protein. The repeat units contain the characteristic sequences of the so-called A-tracts, which are in phase with the helical repeat (i.e. at 10–11 bp intervals). These upstream A-tracts function together as transcriptional enhancers and appear to exhibit inherent DNA curvature [22,23]. Although the precise mechanism of action of these bacterial enhancers is still unknown, this theory is at least partly supported by previous empirical works [36-38], which demonstrated a strong, but not necessarily direct, relationship between the number of A-tracts per promoter and the level of promoter strength.

The hypothesis that phase shifting is correlated with pulmonary infection rests on the idea that virulence is a consequence of the within-host evolution for proliferation into new tissues. We tested this hypothesis retrospectively by analysing the frequency distribution of genetic variants among 17 M. haemolytica serotype A1 strains isolated from the two within-host habitats (Table 4).

Table 4. Distribution of repeat tract polymorphisms using M. haemolytica serotype A1 strains collected from the nasopharynx of healthy cattle and the lungs of cattle with pulmonary infection

The alignment showed significant sequence and positional conservation over the entire 406-bp stretch among all strains, with the exception that they contained different numbers of repeat units per promoter (data not shown). All strains isolated from the nasopharynx of healthy cattle and representing at least two epidemiological clones based on outer membrane protein (OMP) typing (OMP types 1.2 and 1.4) contained three perfect repeat units per promoter (Table 4). The majority of strains isolated from the lungs of cattle with pulmonary infection and representing at least three epidemiological clones (OMP types 1.1, 1.2, and 1.3) also contained three perfect repeat units per promoter, with the exception of two strains (both OMP type 1.2), which contained four perfect repeat units per promoter (Table 4).

The presence of three repeat units per promoter in strains from both within-host habitats suggests that this pattern is associated with a commensal lifestyle in the nasopharynx and that the occasional movement of these bacteria into the lungs may be due to coincidental spill over from the nasopharynx. By contrast, strains containing four repeat units per promoter were confined to the lungs. This observation supports the view that genetic change(s) is (are) required for successful colonisation of the lungs under certain conditions.

Consider an inoculum of M. haemolytica that has entered the nasopharynx of a host. Furthermore, assume that successful colonisation of the lungs requires (at least) one change (i.e. addition of one repeat unit) to generate the necessary phenotype (i.e. high-level expression of the LktA protein). Whether or not these mutants exist in the potentially colonising population depends on the total number of cells in the nasopharynx, the fraction of the inoculum that enters the lungs, the rates at which the required genetic change is generated by mutation, and the duration of the carrier state (i.e. the number of generations that survive in the nasopharynx before the population becomes extinct).

This analysis has several potential limitations. First, strain pairs collected from the nasopharynx and lungs of individual hosts with pulmonary infection (rather than from healthy and sick hosts, respectively) should be used to infer within-host evolution. Second, we were unable to perform in vivo tests of the predictions generated from this analysis because of the lack of a good animal model for M. haemolytica.

Selection for phase shifting

Whether or not virulence associated with phase shifting will be selected at an epidemiological level depends on the trade-off between the transmission advantage accruing to the cell during short bouts of symptoms (sneezing and coughing) and the loss in long-term asymptomatic nose-to-nose transmission due to either increased clearance, rapid host death, or reduced fitness in the nasopharynx, which is the port-of-entry in new hosts.

If symptoms (sneezing and coughing) and transmission are coupled so the more rapid phase shifting (more virulent) bacteria are transmitted at higher rates than the more slow (benign) ones, phase shifting (virulence) would be favoured in the bacterial population. This epidemiological selection model is supported by the apparent evolutionary success of the high virulence M. haemolytica serotype A1/A6 [2].

However, if the genetic and phenotypic changes associated with enhanced fitness in the lungs are detrimental to the bacteria back in the nasopharynx, they are unlikely to confer any long-term advantage in their ability to spread from the lungs of one host to the nasopharynx of another unless revertant changes occur. In this case, disease is a dead end for these bacteria and is a consequence of short-sighted within-host evolution.

If the asymptomatic and symptomatic sites overlap as in the case of respiratory disease, sneezing and coughing generated by the lung-adapted phenotype could favour transmission of freeriders that remain adapted to the nasopharynx, even when disease is a dead end (i.e. revertant changes do not occur, thus preventing transmission of revertant cells to the nasopharynx of new hosts). In other words, freeriders benefit from the collective action without paying the necessary cost associated with the transmission-enhancing feature.

Finally, we present a second model of the epidemiology and within-host infection dynamics of M. haemolytica not considered above. Our association analysis showed that the nucleotide change leading to the origin of short tandem arrays of perfect repeat units coincided with the evolutionary success of M. haemolytica strains belonging to MLEE lineage A, including relatively benign ones that colonise the ovine nasopharynx asymptomatically and rarely cause disease. To understand the selective forces responsible for phase shifting in these bacteria, which are not invariably pathogenic, it is important to consider its relevance to the commensal existence of these bacteria at an epidemiological level, as a mechanism to facilitate colonisation of genetically and immunologically diverse hosts.

Conclusion

The continuous circulation of Mannheimia bacteria in host populations seems to depend on their capacity for asymptomatic nose-to-nose transmission, although other routes of transmission may exist for some species. Our analysis suggests that the origin of short tandem arrays of perfect repeat units have contributed to the derived virulence phenotype of M. haemolytica serotype A1/A6 by allowing phase-variable expression of the LktA protein. Further studies are required to better understand the epidemiology of these opportunistic pathogens. Areas of study should include testing the predictions generated from this analysis using both prospective (experimental) and retrospective (epidemiological) methods.

Methods

Taxa used

For this study we included M. haemolytica serotype A1 str. PHL213, M. haemolytica serotype A2 str. CCUG 12392T, M. glucosida serotype A11 str. P925T, M. ruminalis str. HPA113, M. granulomatis str. P1135/26T, and M. varigena str. 177T to balance the number of representatives from the five species within genus Mannheimia (Table 1). These strains have been represented in previous studies by Angen et al. [1,39-41] and Larsen et al. [20,21].

Sequence analysis

The intergenic regions upstream of the lkt genes were aligned using Dialign 2 [42] with default settings. This segment-based multiple alignment program is known to produce better alignments for the purpose of phylogenetic footprinting than does the tree-based global multiple alignment program ClustalW [43], because it starts by identifying short conserved regions and then incorporates them into a multiple alignment [44].

Estimation of evolutionary rates for CP sequences

The relationships of the strains in this study were inferred with 16S rRNA; these sequences have been used successfully for systematic studies in genus Mannheimia [1,20]. The 16S rRNA sequences were aligned using Dialign 2 [42] with default settings.

Phylogenetic inference from 16S rRNA and CP sequences (Figure 2) was done under the criterion of maximum-likelihood (ML) using PAUP* version 4.0b10 for UNIX [45]. The best fitting model for both data sets was Hasegawa, Kishino, and Yano's [46,47] model with a discrete four-category gamma distribution [47] of substitution rate among different nucleotide sites (HKY85 + Γ) based on the Akaike Information Criterion (AIC) computed by the programs Modeltest version 3.06 [48] and PAUP* version 4.0b10 for UNIX [45].

We estimated all HKY85 + Γ model parameters (nucleotide frequencies, substitution rates (μ), branch lengths (ν), gamma shape parameter) from the 16S rRNA data set. The 16S rRNA tree topology was also found to be the single best estimate when using the program MrBayes version 3.0B4 [49], having a posterior probability (PP) of 100%. The tree was rooted using the M. varigena str. 177T sequence as outgroup.

We then used the 16S rRNA tree topology as the basis for determining the ML estimates of the HKY85 + Γ model parameters from the CP data set, assuming that the 16S rRNA tree topology is a correct representation of the phylogeny connecting the investigated strains. The evolutionary rates for CP sequences were tested for deviation from expectations derived from a neutral evolution model. The null hypothesis is that the 16S rRNA and CP sequences have evolved under neutral evolution on the same tree topology but at different rates. If this is the case, we would expect the CP tree to be a scaled version of the 16S rRNA tree (with some noise due to the stochasticity in the substitution process). The null hypothesis predicts the following relationship between the expected length of any given branch on the CP tree and the observed length of the corresponding branch on the 16S rRNA tree:

where T is the total tree length.

To test whether the observed branch lengths on the CP tree deviate significantly from what we expect under the null hypothesis of neutral evolution, we performed a so-called parametric bootstrapping analysis [50,51]. The program evolver from the PAML package [52] was used to construct 1,000 synthetic "CP-like" data sets by simulating sequence evolution on the CP tree and using the HKY85 + Γ model parameters estimated from the real CP data set. These synthetic "CP-like" data sets are representatives of CP sequences that have evolved under neutral evolution and they can therefore be used to test whether the observed length of any given branch on the CP tree deviates significantly from what we would expect under the null hypothesis of neutral evolution. We used the program BASEML from the PAML package [52] to perform a ML analysis for each of the 1,000 synthetic "CP-like" data sets using the HKY85 + Γ model. For each branch in the 1,000 synthetic "CP-like" trees, we then computed the difference between the observed and the expected branch lengths and tested whether these differences were significantly different from zero (the value expected under the null hypothesis of neutral evolution) by calculating two-tailed p-values adjusted by the Bonferroni correction for multiple tests (as we have no a priori expectation whether the variation of the difference between the observed and the expected branch lengths should be greater (diversifying evolution) or smaller (unifying evolution) than zero, this test is two-sided).

Phylogenetic analysis of repeat sequences in M. haemolytica

The number of repeat units per promoter was examined using 28 M. haemolytica strains representing the diversity within this species based on MLEE, geographic origin, and host association (Table 3) and 17 M. haemolytica serotype A1 strains collected from the nasopharynx of healthy cattle and the lungs from cattle with pulmonary infection (Table 4). These strains have been represented in previous studies by Davies et al. [2] and McCluskey et al. [53], respectively. To amplify the intergenic regions in these strains, we designed the forward primer lktpro_UP (5'-CCACACACCCGAATAAAAGGGTCAAAAGTG-3') and the reverse primer lktpro_DOWN (5'-GGAGTTCATCCATAGCCAAGTAATGTTTCC-3') from conserved sequences in the 5' and 3' flanking genes between the M. haemolytica + M. glucosida strains. The reaction conditions were 2.5 U Taq polymerase, 16 mM (NH4)2SO4, 67 mM Tris-HCl, 0.01% Tween-20, 2.5 mM Mg2SO4, each primer at 0.5 mM, and each nucleotide at 0.1 mM. The cycling conditions were initial denaturation at 94°C followed by 25 cycles of 94°C for 30 s, 52°C for 30 s, and 72°C for 30 s, finishing with extension at 72°C for 10 min. These PCR products were then directly sequenced.

A phylogenetic tree of the two distal repeat units from 30 M. haemolytica strains (Tables 1 and 3) was reconstructed using the maximum-parsimony method as implemented in the program PAUP* version 4.0b10 for UNIX [45]. The first site of the unit sequence was defined arbitrarily by using the AluI restriction site as a landmark [23].

Authors' contributions

JL carried out the molecular genetic studies, participated in the sequence annotations and in the interpretation of the results, and drafted the manuscript. AGP conducted the phylogenetic analysis of repeat sequences and estimated the evolutionary rates for CP sequences. RLD was responsible for the additional M. haemolytica strain collections and participated in the interpretation of the results. PK and JF participated in the design of the study. HC participated in the analysis of gene order data. MB was responsible for the primary strain collection and participated in the interpretation of the results. JEO conceived of the study, participated in its design and coordination, and helped to draft the manuscript. All authors read and approved the final manuscript.

Acknowledgements

We thank Christel G Buerholt and Yvonne Schlatter for technical assistance. This work was supported by a grant from the Danish Agricultural and Veterinary Research Council (9702797)

References

  1. Angen Ø, Mutters R, Caugant DA, Olsen JE, Bisgaard M: Taxonomic relationships of the [Pasteurella] haemolytica complex as evaluated by DNA-DNA hybridizations and 16S rRNA sequencing with proposal of Mannheimia haemolytica gen. nov., comb. nov., Mannheimia granulomatis comb. nov., Mannheimia glucosida sp. nov., Mannheimia ruminalis sp. nov. and Mannheimia varigena sp. nov.

    Int J Syst Bacteriol 1999, 49(Pt 1):67-86. PubMed Abstract | Publisher Full Text OpenURL

  2. Davies RL, Arkinsaw S, Selander RK: Evolutionary geneticsof Pasteurella haemolytica isolates recovered from cattle and sheep.

    Infect Immun 1997, 65:3585-3593. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  3. Frank GH: Pasteurella haemolytica and respiratory disease in cattle.

    Proc Annu Meet US Anim Health Assoc 1979, 83:153-160. OpenURL

  4. Frank GH: When Pasteurella haemolytica colonizes the nasal passages of cattle.

    Vet Med 1988, 83:1064. OpenURL

  5. Jones CD: Proliferation of Pasteurella haemolytica in the calf respiratory tract after an abrupt change in climate.

    Res Vet Sci 1987, 42:179-186. PubMed Abstract OpenURL

  6. Whiteley LO, Maheswaran SK, Weiss DJ, Ames TR, Kannan MS: Pasteurella haemolytica A1 and bovine respiratory disease: pathogenesis.

    J Vet Intern Med 1992, 6:11-22. PubMed Abstract OpenURL

  7. Briggs RE, Frank GH, Purdy CW, Zehr ES, Loan RW: Rapid spread of a unique strain of Pasteurella haemolytica serotype 1 among transported calves.

    Am J Vet Res 1998, 59:401-405. PubMed Abstract OpenURL

  8. Highlander SK, Fedorova ND, Dusek DM, Panciera R, Alvarez LE, Rinehart C: Inactivation of Pasteurella (Mannheimia) haemolytica leukotoxin causes partial attenuation of virulence in a calf challenge model.

    Infect Immun 2000, 68:3916-3922. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  9. Petras SF, Chidambaram M, Illyes EF, Froshauer S, Weinstock GM, Reese CP: Antigenic and virulence properties of Pasteurella haemolytica leukotoxin mutants.

    Infect Immun 1995, 63:1033-1039. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  10. Tatum FM, Briggs RE, Sreevatsan SS, Zehr ES, Ling HS, Whiteley LO, et al.: Construction of an isogenic leukotoxin deletion mutant of Pasteurella haemolytica serotype 1: characterization and virulence.

    Microb Pathog 1998, 24:37-46. PubMed Abstract | Publisher Full Text OpenURL

  11. Ramirez-Romero R, Brogden KA: The potential role of the Arthus and Shwartzman reactions in the pathogenesis of pneumonic pasteurellosis.

    Inflamm Res 2000, 49:98-101. PubMed Abstract | Publisher Full Text OpenURL

  12. Clarke CR, Lauer AK, Barron SJ, Wyckoff JH III: The role of eicosanoids in the chemotactic response to Pasteurella haemolytica infection.

    Zentralbl Veterinarmed B 1994, 41:483-491. PubMed Abstract OpenURL

  13. Cudd L, Clarke C, Clinkenbeard K: Mannheimia haemolytica leukotoxin-induced increase in leukotriene B4 production by bovine neutrophils is mediated by a sustained and excessive increase in intracellular calcium concentration.

    FEMS Microbiol Lett 2003, 224:85-90. PubMed Abstract | Publisher Full Text OpenURL

  14. Radi ZA, Brogden KA, Dixon RA, Gallup JM, Ackermann MR: A selectin inhibitor decreases neutrophil infiltration during acute Mannheimia haemolytica pneumonia.

    Vet Pathol 2002, 39:697-705. PubMed Abstract | Publisher Full Text OpenURL

  15. Slocombe RF, Malark J, Ingersoll R, Derksen FJ, Robinson NE: Importance of neutrophils in the pathogenesis of acute pneumonic pasteurellosis in calves.

    Am J Vet Res 1985, 46:2253-2258. PubMed Abstract OpenURL

  16. Lo RY, Strathdee CA, Shewen PE: Nucleotide sequence of the leukotoxin genes of Pasteurella haemolytica A1.

    Infect Immun 1987, 55:1987-1996. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  17. Highlander SK, Chidambaram M, Engler MJ, Weinstock GM: DNA sequence of the Pasteurella haemolytica leukotoxin gene cluster.

    DNA 1989, 8:15-28. PubMed Abstract OpenURL

  18. Highlander SK, Engler MJ, Weinstock GM: Secretion and expression of the Pasteurella haemolytica leukotoxin.

    J Bacteriol 1990, 172:2343-2350. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  19. Strathdee CA, Lo RY: Regulation of expression of the Pasteurella haemolytica leukotoxin determinant.

    J Bacteriol 1989, 171:5955-5962. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  20. Larsen J, Pedersen AG, Christensen H, Bisgaard M, Angen Ø, Ahrens P, et al.: Evidence for vertical inheritance and loss of the leukotoxin operon in genus Mannheimia.

    J Mol Evol 2007, 64:423-437. PubMed Abstract | Publisher Full Text OpenURL

  21. Larsen J, Kuhnert P, Frey J, Christensen H, Bisgaard M, Olsen JE: Analysis of gene order data supports vertical inheritance of the leukotoxin operon and genome rearrangements in the 5' flanking region in genus Mannheimia.

    BMC Evol Biol 2007, 7:184. PubMed Abstract | BioMed Central Full Text | PubMed Central Full Text OpenURL

  22. Highlander SK, Weinstock GM: Static DNA bending and protein interactions within the Pasteurella haemolytica leukotoxin promoter region: development of an activation model for leukotoxin transcriptional control.

    DNA Cell Biol 1994, 13:171-181. PubMed Abstract | Publisher Full Text OpenURL

  23. Marciel AM, Highlander SK: Use of operon fusions in Mannheimia haemolytica to identify environmental and cis-acting regulators of leukotoxin transcription.

    Infect Immun 2001, 69:6231-6239. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  24. Tautz D: Evolution of transcriptional regulation.

    Curr Opin Genet Dev 2000, 10:575-579. PubMed Abstract | Publisher Full Text OpenURL

  25. Fessele S, Maier H, Zischek C, Nelson PJ, Werner T: Regulatory context is a crucial part of gene function.

    Trends Genet 2002, 18:60-63. PubMed Abstract | Publisher Full Text OpenURL

  26. Lemon B, Tjian R: Orchestrated response: a symphony of transcription factors for gene control.

    Genes Dev 2000, 14:2551-2569. PubMed Abstract | Publisher Full Text OpenURL

  27. Carter AJ, Wagner GP: Evolution of functionally conserved enhancers can be accelerated in large populations: a population-genetic model.

    Proc Biol Sci 2002, 269:953-960. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  28. Ludwig MZ, Bergman C, Patel NH, Kreitman M: Evidence for stabilizing selection in a eukaryotic enhancer element.

    Nature 2000, 403:564-567. PubMed Abstract | Publisher Full Text OpenURL

  29. Ludwig MZ: Functional evolution of noncoding DNA.

    Curr Opin Genet Dev 2002, 12:634-639. PubMed Abstract | Publisher Full Text OpenURL

  30. Stone JR, Wray GA: Rapid evolution of cis-regulatory sequences via local point mutations.

    Mol Biol Evol 2001, 18:1764-1770. PubMed Abstract | Publisher Full Text OpenURL

  31. Ribeiro GA, Carter GR, Frederiksen W, Riet-Correa F: Pasteurella haemolytica-like bacterium from a progressive granuloma of cattle in Brazil.

    J Clin Microbiol 1989, 27:1401-1402. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  32. Riet-Correa F, Mendez MC, Schild AL, Ribeiro GA, Almeida SM: Bovine focal proliferative fibrogranulomatous panniculitis (Lechiguana) associated with Pasteurella granulomatis.

    Vet Pathol 1992, 29:93-103. PubMed Abstract OpenURL

  33. Ladeira SL, Riet-Correa F, Pereira DB, Carter GR: Role of Pasteurella granulomatis and Dermatobia hominis in the etiology of lechiguana in cattle.

    Ann N Y Acad Sci 1996, 791:359-368. PubMed Abstract | Publisher Full Text OpenURL

  34. Murphy GL, Whitworth LC, Clinkenbeard KD, Clinkenbeard PA: Hemolytic activity of the Pasteurella haemolytica leukotoxin.

    Infect Immun 1995, 63:3209-3212. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  35. Moxon R, Bayliss C, Hood D: Bacterial contingency loci: the role of simple sequence DNA repeats in bacterial adaptation.

    Annu Rev Genet 2006, 40:307-333. PubMed Abstract | Publisher Full Text OpenURL

  36. Aiyar SE, Gourse RL, Ross W: Upstream A-tracts increase bacterial promoter activity through interactions with the RNA polymerase alpha subunit.

    Proc Natl Acad Sci USA 1998, 95:14652-14657. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  37. Cheema AK, Choudhury NR, Das HK: A- and T-tract-mediated intrinsic curvature in native DNA between the binding site of the upstream activator NtrC and the nifLA promoter of Klebsiella pneumoniae facilitates transcription.

    J Bacteriol 1999, 181:5296-5302. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  38. Katayama S, Matsushita O, Jung CM, Minami J, Okabe A: Promoter upstream bent DNA activates the transcription of the Clostridium perfringens phospholipase C gene in a low temperature-dependent manner.

    EMBO J 1999, 18:3442-3450. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  39. Angen Ø, Caugant DA, Olsen JE, Bisgaard M: Genotypic relationships among strains classified under the (Pasteurella) haemolytica-complex as indicated by ribotyping and multilocus enzyme electrophoresis.

    Zentralbl Bakteriol 1997, 286:333-354. PubMed Abstract OpenURL

  40. Angen Ø, Olsen JE, Bisgaard M: Further studies of the relationships among strains classified as taxon 15, taxon 18, taxon 20, (Pasteurella) granulomatis or the (Pasteurella) haemolytica-complex in ruminants using quantitative evaluation of phenotypic data.

    Zentralbl Bakteriol 1997, 286:317-332. PubMed Abstract OpenURL

  41. Angen Ø, Aalbaek B, Falsen E, Olsen JE, Bisgaard M: Relationships among strains classified with the ruminant Pasteurella haemolytica-complex using quantitative evaluation of phenotypic data.

    Zentralbl Bakteriol 1997, 285:459-479. PubMed Abstract OpenURL

  42. Morgenstern B: DIALIGN 2: improvement of the segment-to-segment approach to multiple sequence alignment.

    Bioinformatics 1999, 15:211-218. PubMed Abstract | Publisher Full Text OpenURL

  43. Thompson JD, Higgins DG, Gibson TJ: CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice.

    Nucleic Acids Res 1994, 22:4673-4680. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  44. Blanchette M, Tompa M: Discovery of regulatory elements by a computational method for phylogenetic footprinting.

    Genome Res 2002, 12:739-748. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  45. Swofford DL: PAUP*: phylogenetic analysis using parsimony (*and other methods), version 4.0b10. Sunderland, Massachusetts, Sinauer Associates; 2003. OpenURL

  46. Hasegawa M, Kishino H, Yano T: Dating of the human-ape splitting by a molecular clock of mitochondrial DNA.

    J Mol Evol 1985, 22:160-174. PubMed Abstract | Publisher Full Text OpenURL

  47. Yang Z: Estimating the pattern of nucleotide substitution.

    J Mol Evol 1994, 39:105-111. PubMed Abstract OpenURL

  48. Posada D, Crandall KA: MODELTEST: testing the model of DNA substitution.

    Bioinformatics 1998, 14:817-818. PubMed Abstract | Publisher Full Text OpenURL

  49. Ronquist F, Huelsenbeck JP: MrBayes 3: Bayesian phylogenetic inference under mixed models.

    Bioinformatics 2003, 19:1572-1574. PubMed Abstract | Publisher Full Text OpenURL

  50. Goldman N: Statistical tests of models of DNA substitution.

    J Mol Evol 1993, 36:182-198. PubMed Abstract | Publisher Full Text OpenURL

  51. Huelsenbeck JP, Hillis DM, Jones R: Parametric bootstrapping in molecular phylogenetics: applications and performance. In Molecular zoology: advances, strategies and protocols. Edited by Ferraris JD, Palumbi SR. New York: John Wiley and Sons, Inc; 1996:19-45. OpenURL

  52. Yang Z: PAML: a program package for phylogenetic analysis by maximum likelihood.

    Comput Appl Biosci 1997, 13:555-556. PubMed Abstract OpenURL

  53. McCluskey J, Gibbs HA, Davies RL: Variation in outer-membrane protein and lipopolysaccharide profiles of Pasteurella haemolytica isolates of serotypes A1 and A2 obtained from pneumonic and healthy cattle.

    Microbiology 1994, 140(Pt 4):807-814. PubMed Abstract | Publisher Full Text OpenURL