OUP user menu

Evaluation of invertebrate infection models for pathogenic corynebacteria

Lisa Ott, Ashleigh McKenzie, Maria Teresa Baltazar, Sabine Britting, Andrea Bischof, Andreas Burkovski, Paul A. Hoskisson
DOI: http://dx.doi.org/10.1111/j.1574-695X.2012.00963.x 413-421 First published online: 1 August 2012


For several pathogenic bacteria, model systems for host—pathogen interactions were developed, which provide the possibility of quick and cost-effective high throughput screening of mutant bacteria for genes involved in pathogenesis. A number of different model systems, including amoeba, nematodes, insects, and fish, have been introduced, and it was observed that different bacteria respond in different ways to putative surrogate hosts, and distinct model systems might be more or less suitable for a certain pathogen. The aim of this study was to develop a suitable invertebrate model for the human and animal pathogens Corynebacterium diphtheriae, Corynebacterium pseudotuberculosis, and Corynebacterium ulcerans. The results obtained in this study indicate that Acanthamoeba polyphaga is not optimal as surrogate host, while both Caenorhabtitis elegans and Galleria larvae seem to offer tractable models for rapid assessment of virulence between strains. Caenorhabtitis elegans gives more differentiated results and might be the best model system for pathogenic corynebacteria, given the tractability of bacteria and the range of mutant nematodes available to investigate the host response in combination with bacterial virulence. Nevertheless, Galleria will also be useful in respect to innate immune responses to pathogens because insects offer a more complex cell-based innate immune system compared with the simple innate immune system of C. elegans.

  • Acanthamoeba
  • Caenorhabtitis elegans
  • Corynebacterium diphtheriae
  • Corynebacterium pseudotuberculosis
  • Corynebacterium ulcerans
  • Galleria mellonella


The genus Corynebacterium belongs to the class of Actinobacteria (high G+C gram-positive bacteria) and comprises a collection of morphologically similar, irregular-shaped, or club-shaped microorganisms (Ventura et al., 2007; Zhi et al., 2009). To date, more than 80 species were taxonomically classified into the genus Corynebacterium including industrially important bacteria, commensals of animals and humans as well as pathogens. A well-known member of the pathogenic species is Corynebacterium diphtheriae, which is also the type species of the whole genus.

Corynebacterium diphtheriae is the etiological agent of diphtheria, a localized toxaemic infection of respiratory tract and skin that can be fatal (Hadfield et al., 2000). Owing to its medical importance, C. diphtheriae is the best-investigated pathogenic corynebacterium; however, even for this species, only a few virulence factors have been characterized. Beside the diphtheria toxin, these include mainly pili and other adhesion factors (for reviews, see Collier, 2001; Rogers et al., 2011). Already these data show a high degree of variation in different isolates, for example, in respect to pili formation (Ott et al., 2010). Molecular data on infection are even scarcer for two closely related species, Corynebacterium ulcerans and Corynebacterium pseudotuberculosis, which are separate species within a distinct cluster of the genus Corynebacterium but share the ability to harbor diphtheria toxin-encoding lysogenic corynephage (Riegel et al., 1995).

Corynebacterium ulcerans has been detected as a commensal in domestic and wild animals, which may serve as reservoirs for zoonotic infections. During the last decade, the frequency and severity of human infections associated with C. ulcerans appear to be increasing in various countries. Respiratory diphtheria-like illnesses caused by toxigenic C. ulcerans strains are increasingly reported from various industrialized countries (Tiwari et al., 2008) and have become more common than C. diphtheriae infections in the United Kingdom (Wagner et al., 2010). Infections of humans with toxigenic C. ulcerans can be fatal in unvaccinated patients. They usually occur in adults who have consumed raw milk or had close contact with domestic animals (Bostock et al., 1984; Hart, 1984; Wagner et al., 2010). Beside respiratory diphtheria-like illnesses, C. ulcerans can also cause extrapharyngeal infections in humans, including severe pulmonary infections (Dessau et al., 1995; Nureki et al., 2007; Cosson & Soldati, 2008).

Corynebacterium pseudotuberculosis is the etiological agent of caseous lymphadenitis, which is prevalent in sheep and goat herds throughout the world (Dorella et al., 2006; Baird & Fontaine, 2007). Infections due to C. pseudotuberculosis are rare in humans, but this pathogen is occasionally recovered from cases of suppurative lymphadenitis in patients with a classical risk exposure of close contact with sheep. Corynebacterium pseudotuberculosis is a facultative intracellular pathogen that is able to survive and grow in macrophages, thus escaping the immune response of the host (McKean et al., 2005; Dorella et al., 2006).

As mentioned earlier, the three described species are not very well characterized on a molecular level, despite the fact that genome sequences became available for C. diphtheriae (Cerdeno-Tarraga et al., 2003), C. ulcerans Trost et al., (2011), and C. pseudotuberculosis Trost et al., 2010; Cerdeira et al., 2011a, b; Silva et al., 2011) during the last years. This might be partly due to the lack of a suitable, easy-to-handle infection model, which would allow fast identification of virulence factors in large-scale mutant pools. Besides the classical guinea-pig model already used by Loeffler, a mouse model for septic arthritis was successfully established for C. diphtheriae (Puliti et al., 2006); however, nonmammalian model systems might offer several advantages over mammals, for example, in respect to cost effectiveness, handling, and ethical aspects.

The aim of this study was to identify a nonmammalian infection model for pathogenic corynebacteria with potential for high throughput screening and testing of mutants. A number of different model systems, including amoeba, nematodes, insects, and fish, have been introduced to study host—pathogen interactions (for recent reviews, see Dorer & Isberg, 2006; Kurz & Ewbank, 2007, O'Callaghan & Vergunst, 2010). Here, we present data for three invertebrate models and their suitability as surrogate hosts of pathogenic corynebacteria, the amoeba Acanthamoeba polyphaga, which is especially applied to investigate adhesion and phagocytosis (Mattos-Guaraldi et al., 2008), the nematode Caenorhabtitis elegans, which has been widely used and often serves as an oral infection route model (for review see Dorer & Isberg, 2006; Waterfield et al., 2008; O'Callaghan & Vergunst, 2010), and larvae of the greater wax moth (Galleria mellonella), an established model which represents the more complex insect immune system (Waterfield et al., 2008). Bacterial strains were selected based on their previous characterization and the availability of experimental data and sequence information.

Materials and methods

Strains and growth conditions

Strains used in this study are listed in Table 1. Corynebacterium diphtheriae, C. ulcerans, and C. pseudotuberculosis were grown in heart infusion (HI) broth at 37 °C, Corynebacterium glutamicum was grown in brain heart infusion (BHI) broth at 30 °C, and Escherichia coli OP50 was grown in Luria broth (LB) (Sambrook et al., 1889) at 37 °C. If appropriate, kanamycin was added (25 µg mL−1 for C. glutamicum and C. pseudotuberculosis, 30 µg mL−1 for E. coli, 50 µg mL−1 for C. diphtheriae and C. ulcerans). Caenorhabtitis elegans N2, used as the wild type strain, were maintained and propagated on E. coli OP50 as described (Brenner, 1974).

View this table:
Table 1

Strains, plasmids and cell lines used in this study

Corynebacterium diphtheriae
DSM 43988Nontoxigenic, isolated from throat cultureDSMZ, Braunschweig, Germany
DSM 43989tox+, unknown sourceDSMZ, Braunschweig, Germany
ISS3319C. diphtheriae var. mitis, nontoxigenic, isolated from patient affected by pharyngitis/tonsillitisBertuccini et al. (2004)
Tn5-46ISS 3319 carrying a Tn5 insertion in DIP1546 (encoding a hypothetical protein)Laboratory strain collection
Corynebacterium glutamicum
ATCC 13032Type-strain, nonpathogenicAbe et al. (1967)
Corynebacterium pseudotuberculosis
FRC41Isolated from the inguinal lymph node of a 12-year-old girl with necrotizing lymphadenitisJoin-Lambert et al. (2006)
Corynebacterium ulcerans
809Bronchoalveolar lavage sample from an elderly woman with a fatal pulmonary infectionTrost et al. (2011)
BR-AD22Nasal sample of an asymptomatic female dogTrost et al. (2011)
ELHA1Phospholipase D-deficient BR-AD22 mutant strainE. Hacker (pers. commun.)
Escherichia coli
OP50Uracil auxotrophic E. coli B strain(Brenner, 1974)
pEPR1-p45gfpP45, gfpuv, KmR, rep, per, T1, T2Knoppova et al. (2007)
pXMJ19mCherryori colE1, oricg , ptac, mCherry, CmRM. Höller (pers. commun.)
Cell lines
Detroit 562Human hypopharyngeal carcinoma cellsPeterson et al. (1968)

Infection of Acanthamoeba polyphaga

For a amoebae-bacteria co-culture assay, A. polyphaga cultures were grown to exponential phase (3–5 days) in Peptone Yeast Glucose medium at 21 °C (Rowbotham, 1980). Cells were harvested by centrifugation and re-suspended in phosphate-buffered saline (PBS), and live cell titers (visualized by trypan blue exclusion) were carried out to assess the number of amoebae for each assay (normally 2 × 105 amoebae mL−1). Corynebacteria were grown overnight in HI medium, diluted to an OD600 nm of 0.1 and grown to an OD600 nm of 0.6 and compared with a standard curve of viable numbers for infection assays at a range of multiplicity of infection (MOI). The amoebae and bacteria were combined in a 24-well plate at a MOI of 1 : 1 or 1 : 10. Plates were incubated at 21 °C, with cells harvested for microscopy at the indicated time intervals.

For a phagocytic plaque assay of amoebae killing, the method of Froquet et al. (2009) was adapted for assaying the interaction of A. polyphaga and Corynebacterium species. Aliquots (0.75 mL) of HI agar in varying concentrations (10–100% of the manufacturer's recommended concentration) were placed in each well of a 24-well plate and dried. Diluting the media equalizes the conditions for amoeba and bacterial growth and establishes the appropriate conditions for plaque formation (Froquet et al., 2009). Corynebacterial cultures were prepared as described earlier and 50 µL of the culture was pipetted out onto the surface of the agar and allowed to dry. Amoebal cells were harvested as described earlier and diluted to a concentration of 200 × 104, 20 × 104, 2 × 104 or 0.2 × 104 live cells mL−1 with PBS. Amoebal suspension (5 µL) was pipetted into the centre of each well. Plates were incubated at room temperature for 2–5 days and observed daily for signs of plaque formation.

Infection of C. elegans

Caenorhabtitis elegans N2 were maintained on E. coli OP50 for 6–7 days until the worms become starved, indicated by clumping behaviour (de Bono & Bargmann, 1998). The nematodes were infected with different Corynebacterium wild-type strains, transformed with pEPR1-p45gfp as well as E. coli OP50 transformed with pEPR1-p45gfp. Infection of 20 L4 stage larval worms was carried out with 200 µL of each bacteria strain (from an overnight culture) on NGM plates at 21 °C for 24 h. Subsequently, C. elegans was transferred to plates with 200 µL of unlabeled E. coli OP50 for a further 24 h, to allow the gut to clear of fluorescent organisms. Worms were assessed each day following infection, and the dead worms were counted. Transfer back to E. coli OP50 is essential following infection as corynebacteria are not a preferred prey source for C. elegans and continued feeding results in the nematodes attempting to leave the culture plates. Nematodes were paralyzed with 0.6% 2-phenoxy-2-propanol (Sigma; Wormbook.org), mounted onto agar pads, and photographed using a Leica DMR fluorescence microscope.

Infection of Galleria mellonella larvae

Aliquots (20 mL) of HI broth were inoculated with bacteria from an overnight culture to an OD600 nm of 0.1, and the cultures were grown until an OD600 nm of 0.6 was reached. Bacteria were harvested by centrifugation (10 min, 4500 g) and resuspended in 10 mM MgSO4 to an OD600 nm of 10 (approximately 3 × 109 CFU mL−1). For infection, a 50-µL Hamilton syringe was used to inject 5 µL aliquots into G. mellonella larvae via the hindmost left proleg (Jander et al., 2000). For each biological replicate, five larvae were infected with each strain and incubated at 25 °C for 48 h.

Adhesion assays

Detroit 562 cells were seeded in 24-well plates (bio-one Cellstar; Greiner, Frickenhausen, Germany) at a density of 2 × 105 cells per well 48 h prior to infection. Bacteria were subcultured (OD600 nm of 0.1 from overnight cultures) in HI broth for 3.5 h and adjusted to an OD600 nm of 0.2. A master mix of the inoculum was prepared in DMEM without penicillin/streptomycin at a MOI of 200 (viable counts experiments). The plates were centrifuged for 5 min at 500 g to synchronize infection and subsequently incubated for 1.5 h. The cells were washed with PBS nine times, detached with 500 µL trypsin solution (0.12% trypsin, 0.01% EDTA in PBS) per well (5 min, 37 °C, 5% CO2, 90% humidity), and lysed with 0.025% Tween 20 for 5 min at 37 °C. Serial dilutions were made in prechilled 1× PBS and plated out on blood agar plates. After overnight incubation, the number of colony forming units (CFU) was determined.

Results and discussion

The use of multiple, disparate lower eukaryotic hosts, has been used to model bacterial pathogenesis in mammals and is well established for some organisms (Mahajan-Miklos et al., ). The application and comparison of such models in corynebacteria to indicate strain-specific differences has not been demonstrated before; however, it is clear that some of these models offer utility in the study of pathogenic corynebacteria.

Acanthamoeba polyphaga

Amoebae species are well-established model systems for a number of pathogenic bacteria; however, some bacteria respond in different ways to phagocytosis, and it has been shown that some bacteria, such as Listeria monocytogenes, do not survive ingestion by Acanthamoeba (Akya et al., 2010). Amoebae models have been used extensively for studies on phagocytosis of mycobacteria as surrogate macrophages to study vacuole survival and adaptation (Mattos-Guaraldi et al., 2008). We tested different C. diphtheriae strains using a modified version of the method of Froquet et al. (2009), plaques on C. diphtheriae lawns appeared on average after three to five days of incubation, indicating that the corynebacteria tested were not able to resist predation by A. polyphaga (data not shown). This observation was further supported by co-culture experiments and visualization by fluorescence microscopy, which indicated digestion of the bacteria inside the acid vacuole of A. polyphaga (data not shown). Although the amoebae were able to predate corynebacteria, in co-culture experiments, the number of dead amoeba increased depending on the presence of bacteria (data not shown). The reason for this is unclear; however, these results suggest that the A. polyphaga model was not optimal for characterization of corynebacterial strains, and further experiments with C. diphtheriae, C. pseudotuberculosis, and C. ulcerans strains were abandoned.

Caenorhabtitis elegans

Caenorhabtitis elegans has been extensively used to study bacterial virulence and offers many advantages such as its ease of culture, microscopic tractability, extensive genetic resources, mutant libraries, and ease of manipulation (Mahajan-Miklos et al., 1999). We colonized C. elegans with a range of GFP-labeled corynebacterial strains and subsequently transferred to plates with unlabelled E. coli OP50 before microscopy (Fig. 1) and quantifying survival (Fig. 2). Worms infected with different strains of corynebacteria, exhibiting a range of virulence properties, showed different characteristics in the kill curve reflecting the differences in virulence between strains (Fig. 2), with the clinical C. diphtheriae strain ISS 3319 resulting in faster worm death rates than strain DSM 43988. Infecting C. elegans with GFP-labeled E. coli or C. glutamicum followed by transfer to unlabeled OP50 E. coli resulted in clearance of GFP-expressing strains, suggesting that these bacteria were not colonizing the worms pathogenically. In the case of C. diphtheriae strain ISS 3319, foregut (buccal cavity and pharynx) and midgut were colonized by the bacteria, while C. diphtheriae DSM 43988 colonized solely the midgut. Toxigenic C. diphtheriae DSM 43989 as well as C. ulcerans 809 colonized the hindgut of the worm, while C. ulcerans BR-AD22 and C. pseudotuberculosis FRC41 spread all over the body. These data suggest that different virulence properties and mechanisms used by the wide range of strains we tested are reflected in the localization and the killing kinetics of the experiments. Investigation of the time required to colonize the worms indicated that a minimum of 6 h feeding on pathogenic corynebacteria was required for successful colonization (data not shown). Interestingly, the colonization of the nematodes in the foregut reflects the natural niche of these organisms in their mammalian hosts, rather than the disperse colonization of the gut observed in Pseudomonas or Serratia (Mahajan-Miklos et al., 1999; Kurz et al., 2003). In summary, in contrast to the nonpathogenic C. glutamicum and E. coli, pathogenic corynebacteria were able to colonize C. elegans in a strain-specific manner and were tractable by microscopic studies. The availability of innate immune mutant C. elegans strains would offer attractive possibilities for further studying host—corynebacterium interactions.

Figure 1

Infection of Caenorhabtitis elegans with different corynebacteria. Infection of worms was carried out with 200 µL of each bacterial strain on NGM plates at 21 °C for 24 h. Nematodes were mounted onto agar pads, paralyzed with 0.6% 2-phenoxy-2-propanol (Sigma), and photographed using a Leica DMR fluorescent microscope (1 bright field, 2 fluorescence). In each of at least five independent experiments, approximately 20 worms were infected. Pictures show representative results. Caenorhabtitis elegans N2 fed with (a) Escherichia coli OP50, (b) Corynebacterium glutamicum ATCC 13032, (c) Corynebacterium diphtheriae ISS3319, (d) C. diphtheriae DSM 43988, (e) C. diphtheriae DSM 43989, (f) Corynebacterium ulcerans 809, (g) C. ulcerans BR-AD22, (h) Corynebacterium pseudotuberculosis FRC41. Tails of nematodes are indicated by arrows.

Figure 2

Nematode survival following infection with Corynebacterium diphtheriae. Control — Escherichia coli OP50 (▲); C. diphtheriae ISS3319 (▲); C. diphtheriae DSM 43988 (●); C. diphtheriae DSM 43989 (■). Data are the mean of three experiments, and the error bars represent the standard deviation of the data.

Galleria mellonella

The greater wax moth (Galleria mellonella) model system has also been successfully used to characterize bacterial virulence in a range of strains, such as in Pseudomonas aeruginosa and Burkholderia to illustrate inter-strain variation in pathogenicity (Jander et al., 2000). The larvae show a high sensitivity towards these pathogenic bacteria resulting in melanization in response to infection, and a 50% lethal dose (LD50) of < 10 bacteria was reported when these were injected into the hemolymph (Jander et al., 2000; Wand et al., 2011). In general, distinct differences in the response of the larvae were detected when different corynebacterial strains were injected (Fig. 3). While control larvae showed a pale whitish colour and were highly active, injection of C. glutamicum, a nonpathogenic Corynebacterium species of biotechnological importance, led to the development of small melanized spots. The melanization of larvae is associated with pathogen killing and phagocytosis in lepidopteran larvae (Nappi & Christensen, 2005; Senior et al., 2011). The effects on the larvae became more severe in response to injection of nontoxigenic C. diphtheriae strain DSM 43988, which led to brown larvae, and toxigenic C. diphtheriae strain DSM 43989, which caused even stronger melanization, with all infected larvae being immobile without external stimulation. The strongest effects on the larvae were obtained for C. pseudotuberculosis and C. ulcerans strains, which exhibited high degrees of melanization, immobility, and rapid death. These data suggest that the C. pseudotuberculosis and C. ulcerans strains are more virulent in this model, perhaps reflecting their wider host range when compared to C. diphtheriae; a phenomenon observed in the opportunistic broad host range pathogen P. aeruginosa (Jander et al., 2000).

Figure 3

Infection of Galleria mellonella with different corynebacteria. A 50-µL Hamilton syringe was used to inject 5 µL aliquots of the different bacteria into G. mellonella larvae via the hindmost left proleg. Five larvae were infected with each strain at 25 °C for 48 h. Infection was carried out with (a) buffer control (10 mM MgSO4), (b) Corynebacterium glutamicum ATCC 13032, (c) Corynebacterium diphtheriae ISS3319, (d) C. diphtheriae DSM 43988, (e) C. diphtheriae DSM 43989, (f) Corynebacterium pseudotuberculosis FRC41, (g) Corynebacterium ulcerans 809, (h) C. ulcerans BR-AD22.

Application of model systems

In addition to the already obvious strain-specific differences detectable in the C. elegans and Galleria model systems, two mutant strains were tested as further proof of principle (Fig. 4). When the C. diphtheriae wild-type ISS 3319 was compared with a corresponding transposon mutant Tn5-46, carrying an insertion in a gene encoding a hypothetical protein (DIP1546), the mutant was drastically impaired in gut colonization in the C. elegans model, while in the Galleria system, no difference of wild type and corresponding mutant was observed upon injection, indicating that the DIP1546 protein is influencing adhesion and host colonization. In fact, in cell culture experiments, Tn5-46 showed a decreased adhesion rate to Detroit 562 cells.

Figure 4

Application of model systems. Caenorhabtitis elegans and Galleria larvae were infected with mCherry-labelled corynebacteria (details as described in legends to Figs 1 and 3), and adhesion rates to Detroit 562 cells were determined for the corresponding strains (experiments carried out in at least in triplicate). Upper panel: Corynebacterium diphtheriae strain ISS 3319 and corresponding mutant Tn5-46; lower panel: Corynebacterium ulcerans strain BR-AD22 and phospholipase D-deficient mutant ELHA1. Tails of C. elegans are indicated by arrows.

In contrast, the C. ulcerans phospholipase D-deficient strain ELHA1 showed unaltered colonization behaviour compared with the parental strain BR-AD22 in the C. elegans system, while lack of this virulence factor causes a less severe response in the Galleria model. In this case, the adhesion rates of wild type and mutant were identical.


The results obtained in this study indicate that A. polyphaga is not optimal as surrogate host for the study and characterization of pathogenic corynebacteria. Both C. elegans and Galleria larvae seem to offer attractive models for rapid assessment of virulence differences between strains. While the Galleria larvae are very easy to handle, C. elegans gives the more differentiated results, has extensive genetic tools as a host, is microscopically tractable, and might represent the best model system for pathogenic corynebacteria, especially when factors responsible for colonization are screened. Based on our data, both of these hosts, however, will be useful given the differences detectable in the Galleria system in response to absence or presence of toxins (diphtheria toxin) and other virulence factors (phospholipase D). Furthermore, insects such as Galleria offer the possibility to study a more complex cell-based innate immune system compared with the simple innate immune system of C. elegans (Mahajan-Miklos et al., 2000).


The Burkovski laboratory is supported by the Deutsche Forschungsgemeinschaft in frame of SFB 796 (project B5). The work in the Hoskisson laboratory is supported by Medical Research Scotland (Grant FRG-422) and by a Society for General Microbiology Vacation Studentship grant to A.M. Corynebacterium ulcerans phospholipase D mutant ELHA1 was kindly provided by E. Hacker and plasmid pXMJ19mCherry by M. Höller (both Friedrich-Alexander-Universität Erlangen-Nürnberg), C. elegans N2 was a gift from J. Pettitt (University of Aberdeen). The support of R. Palmisano (Friedrich-Alexander-Universität Erlangen-Nürnberg) in respect to fluorescence microscopy is gratefully acknowledged.


View Abstract