Correspondence: Present address: Galina A. Eroshenko, Russian Anti-Plague Research Institute ‘Microbe’, Universitetskaya, 46, Saratov, 410005, Russian Federation. Tel.: +84 52 262 131; fax: +84 52 515 211; e-mail: firstname.lastname@example.org
The biofilm-forming phenotype of 14 isolates from four ‘nonmain’ subspecies of Yersinia pestis was compared with eight isolates from the more commonly studied ‘main’ or epidemic subspecies of Y. pestis in this study. The four nonmain subspecies are more geographically limited, and are associated with certain mammalian hosts and regions of the Caucasus and Central Asia, whereas the main subspecies spread worldwide during the historic plague pandemics. With the main subspecies pestis, pigmentation on Congo red medium (CR+) correlated with biofilm formation on both abiotic and biotic surfaces. Main subspecies pestis strains that do not produce pigmentation on Congo red medium (CR−) have a deletion that includes the hmsF and hmsS genes known to be required for biofilm formation. CR− strains of the nonmain subspecies, altaica and ulegeica, differed however from pestis and, while defective for biofilms on the two surfaces, both had intact hmsF and hmsS genes. The presence of rcsA was also investigated and results showed that it occurred with a 30-bp insertion in all forms of the subspecies. These findings suggest that biofilms are regulated differently in altaica and ulegeica than they are in pestis and also indicate that the rcsA pseudogene arose early in Y. pestis evolution, increasing the ability of the strain to form biofilm and thereby increasing its effective transmission.
The plague is a particularly dangerous vector-borne disease that still poses a significant threat to human health due to its circulation in numerous natural foci including those in the Russian Federation and its neighbouring countries. The causal agent of the plague is Yersinia pestis, which has a complex life cycle, involving survival in the mammal (37 °C) and flea vector (28 °C). Under these two different conditions, the plague agent expresses different pathogenic and housekeeping factors, most of which are shared with another pathogenic yersinia, Yersinia pseudotuberculosis. Genome sequencing of several Y. pestis and Y. pseudotuberculosis strains has indicated that the plague microorganism has recently evolved from the pseudotuberculosis agent (Parkhill et al., 2001; Chain et al., 2004). In a relatively short period of time, Y. pestis transformed from a saprophytic enteropathogenic bacterium to an obligate parasite, and systemic disease agent with new ways of transmission, via flea bite or droplet airway (Wren, 2003). During this process, many characteristics in Y. pseudotuberculosis changed, including the ability to form biofilms on various surfaces.
A bacterial biofilm is a complex, compact community of cells enclosed in an extracellular matrix, often attached to a surface (Costerton et al., 1995). Yersinia pseudotuberculosis uses a biofilm mode of growth to survive in the environment and to avoid ingestion by invertebrates (Darby et al., 2002; Joshua et al., 2003; Erickson et al., 2006). Its descendant, Y. pestis, further transformed this characteristic into the ability to form a bacterial block, a massive biofilm in the foregut of the flea vector. A blocked flea ensures effective transmission of the plague agent, as the flea regurgitates pieces of the bacterial biofilm into the bite site, injecting the bacteria into or under animal skin, thereby causing development of the bubonic plague and subsequent bacteremia (Hinnebusch et al., 1996; Jarrett et al., 2004). Although it has been demonstrated that early-phase transmission of Y. pestis by unblocked fleas could play an important role in the rapid spread of plague epizootics, transmission by blocked fleas nevertheless remains one of the basic mechanisms for epizootic development as well as for the long-term maintenance of the plague agent in animal reservoirs (Voronova, 1989; Bazanova et al., 1991; Eisen et al., 2006).
Ancient forms of the plague microorganism have survived in nature and are known as the Y. pestis nonmain subspecies. They occupy an intermediate position between Y. pseudotuberculosis and the highly virulent Y. pestis ssp. pestis. The nonmain subspecies –caucasica, altaica, hissarica and ulegeica– circulate in different natural plague foci in Russia and its neighbouring countries (Anisimov et al., 2004). They are selectively virulent for laboratory animals and have a low epidemic potential. The coexistence in nature of bacteria related to the evolutionary ancestor, intermediate virulence subspecies and highly virulent Y. pestis ssp. pestis gives an opportunity to retrace the steps of genome reorganization that led to the rapid evolution of a particularly dangerous infectious disease agent.
Twenty-two Y. pestis strains and four Y. pseudotuberculosis strains were used in this study (Table 1). Of the 22 Y. pestis strains used, eight were from the main subspecies pestis and 14 from the nonmain subspecies (caucasica, altaica, hissarica and ulegeica). Most of the strains were originally isolated from natural plague hotspots in the Russian Federation and surrounding areas, and two main subspecies Y. pestis strains, EV76 and 55-801, were isolated from Madagascar and Vietnam, respectively. All strains were received in lyophilized form from the state collection of pathogenic bacteria (Russian Anti-Plague Research Institute, Saratov). Bacteria were grown in Luria–Bertani (LB) broth and agar (pH 7.2) at 28 °C for 24–48 h. The ability of the strains to form pigmented colonies was studied by growing them for 4 days at 28 °C on a solid synthetic medium designed for pigmentation analysis [g L−1: 21.64 agar base, 2.16 d (−) galactose, 0.06 dl-β-phenyl-α-alanine, 0.02 dl-valine, 0.12 l-arginine HCl, 0.03 dl-methionine, 0.05 l-cysteine HCl, 0.03 dl-isoleucine, 0.0 glycine, 0.75 nonhydrated sodium sulphite, 0.03 Congo red, (pH 7.1±0.2)], produced by the Russian Anti-Plague Research Institute ‘Microbe’. To ensure reliability of the results, all experiments were repeated three times.
Table 1. Characteristics of Yersinia pestis and Yersinia pseudotuberculosis strains used in this study
Species, subspecies, strain, source, date of isolation
Formation of pigmented colonies
Presence of genes hmsF, hms S
Presence of hmsP, hmsT; gmhA; speA, speC
Presence of 30-bp insertion in rcsA
Biofilm formation on abiotic surface
Biofilm formation on cuticle of nematode
Values are averages of three or more independent experiments, confidence intervals are indicated.
Quantification of biofilms was performed by the crystal violet assay, with minor modifications of a previously reported method (O'Toole et al., 1999). Strains of Y. pestis and Y. pseudotuberculosis were grown overnight at 28 °C, then diluted 1 : 100 up to OD600 nm of 0.1 and 3 mL put in sterile polystyrol dishes. These were incubated at 28 °C for 48 h; then the broth was decanted. Bacteria were taken from dish walls for analysis by electron microscopy. Cells attached to dish walls were stained with 0.01% crystal violet in 96% ethanol at room temperature for 45 min. Dishes were washed three times with water, and biofilms were solubilized with a mixture of 80% ethanol and 20% acetone. A570 nm was measured on a Spectronic 5000 spectrophotometer.
Biofilm formation on nematode cuticles
Studies of biofilm formation by Y. pestis strains on biotic surfaces were performed on the cuticle of the Caenorhabditis elegans nematode, wild-type strain N2 Bristol, obtained from Caenorhabditis Genetics Center (University of Minnesota). Twenty C. elegans adult worms were placed on a bacterial lawn (grown on NGM agar at room temperature for 24 h) for egg laying and then removed (Joshua et al., 2003). Dishes were incubated at 20 °C for 48 h. The ability of Y. pestis strains to form biofilms on biotic surfaces was quantified by calculating the percentage of blocked adult nematodes on a dish.
Samples for electron microscopy were prepared by standard procedure using 1.5% OsO4 fixation followed by staining with 1% Rutenium red solution. The samples were then examined using transmission electron microscope Hitachi HU-12A (Japan) at accelerated voltage 50 kV.
Specific primers were used for PCR detection of the following genes: gmhA, primer design by Darby et al. (2005); hmsS, design by Tong et al. (2005); hmsP, design by Bobrov et al. (2005); speA and speC genes, design by Patel et al. (2006). For detection of hmsF, hmsT and rcsA, new primer pairs were constructed: hmsF-S – AAGACAGCACAGGGCGGAC and hmsF-R – TCCGTGGCCCACAGGTAA; hmsT-S – TCTACTGACAGCACGATATT and hmsT-As – TATCCAGGCCTAAAACAC; and rcsA-S – TATTGTCGCTATGGTGGT and rcsA-As – TAGGCATCTCTGTCATCC.
Results and discussion
Pigmentation and biofilm formation by the main subspecies Y. pestis strains
We explored biofilm formation by eight main subspecies Y. pestis strains. Six of them, namely A-161, I-1996, 231, I-3223, A-1836 and A-1793 were isolated from different natural plague foci (high-mountain, mountain and steppe) in the Russian Federation and neighbouring countries. One strain, vaccine EV76, was isolated in Madagascar and another, 55-801, in Vietnam (Table 1). Jackson & Burrows (1956) first discovered and defined the pigmentation phenotype (due to dye adsorption on the outer membrane of cells) in Y. pestis, and Surgalla & Beesley (1969) first correlated hemin binding to Congo red binding. More recently, it has been established that biofilm production by Y. pestis KIM correlates with its ability to form pigmented colonies on a solid medium with hemin or Congo red (Darby et al., 2002; Kirillina et al., 2004; Bobrov et al., 2005, 2008; Erickson et al., 2006; Forman et al., 2006). To confirm this correlation with the main subspecies Y. pestis strains under investigation, we explored their ability to form pigmented colonies (Table 1). Bacteria were grown for 4 days at 28 °C on solid medium with Congo red. Of the eight main subspecies strains, five (231, I-1996, I-3223, A-1836, A-1793) were capable of forming pigmented (red) colonies (CR+ phenotype, 100%). The three other strains (EV76, A-161 and 55-801) did not form any pigmented colonies (CR− phenotype).
Biofilm formation by these strains on abiotic surfaces was subsequently investigated. The bacteria were grown in polystyrol dishes in LB broth, and then cells attached to dish walls were stained with Crystal violet. A570 nm of the main subspecies CR+ strains (I-1996, I-3223, A-1836, A-1793) ranged from 1.0 to 1.2 OD, while that of the CR− strains (EV76, A-161 and 55-801) ranged from 0.1 to 0.26 OD. The absorbance of the CR− strains was therefore considered to be insignificant, while that of the CR+ strains was significantly higher due to a large number of cells in biofilms, attached to the dish walls. Electron microscopy analysis of the biofilms from the CR+ strains, showed that they consisted of cell aggregates embedded in an extracellular matrix (Fig. 1a). The CR− isolates did not form biofilms under these conditions (Fig. 1b). Thus a distinct correlation between pigmentation and biofilm formation on abiotic surfaces was observed in natural main subspecies Y. pestis strains.
In Y. pestis, products of the chromosomal hmsHFRS operon play a key role in pigmentation (Perry & Fetherston, 1997; Bobrov et al., 2005). These genes have also been shown to be necessary for biofilm formation (Joshua et al., 2003; Bobrov et al., 2008). The hms operon is part of the chromosomal pgm region that is often lost in main subspecies strains due to a large deletion (Kutyrev et al., 1992; Fetherston & Perry, 1994). To determine the reason for the absence of pigmentation and biofilm production in CR− strains, we looked for their hms genes by PCR with primers specific for hmsF and hmsS. The results showed that these strains indeed lacked hmsF and hmsS (and probably the whole hms operon). The presence of these genes was confirmed in the five main subspecies (CR+) strains (Table 1).
Furthermore, we looked for the presence of other genes known to take part in biofilm formation in the genomes of the main subspecies strains. Using PCR, we were able to detect the regulatory genes hmsT and hmsP, along with the heptose biosynthesis gene, gmhA, and the polyamine putrescine biosynthesis genes, speA and speC (Table 1) (Kirillina et al., 2004; Bobrov et al., 2005; Darby et al., 2005; Patel et al., 2006) in all eight Y. pestis strains.
The correlation between pigmentation and biofilm production on a biotic surface was then investigated using C. elegans as a model system. The three CR− strains did not produce biofilms on nematode cuticles and worms moved freely on the lawns of these strains (Fig. 2a). Conversely, a large number (23.3±1.25%) of C. elegans adult nematodes with well-formed biofilm fragments on their head and neck sections were seen on the lawns of the main subspecies CR+ strains. The worms were blocked by cohesive bacterial aggregates and made aberrant movements to escape (Fig. 2b). These results show that most (five out of eight) of the main subspecies Y. pestis strains produced biofilms on both abiotic and biotic surfaces and that these characteristics distinctly correlate with their ability to form pigmented colonies and with the presence of the hmsHFRS operon (Table 1).
Pigmentation and biofilm production by the nonmain subspecies Y. pestis strains
Studying biofilm production by the nonmain subspecies Y. pestis strains was of significant interest, as this property had not been previously explored. Three strains from subspecies caucasica (1146, 818 and 3544 Arm, isolated in the Caucasus); three strains from subspecies altaica (I-2998, I-2359 and I-2183 from Altai); four strains from subspecies hissarica (A-1249, A-1725, A-1723 and A-1728 from Tadjikistan) and four strains from subspecies ulegeica (I-3069, I-3131, I-3130 and I-2422 from Mongolia) were used (Table 1). In these strains, a close correlation between pigmentation and biofilm production on abiotic surfaces was also observed. Three of the caucasica ssp. strains (818, 1146 and 3544 Arm), one altaica ssp. strain (I-2183), one hissarica ssp. strain (A-1728) and two ulegeica ssp. strains (I-3130, I-2422) formed pigmented colonies (100%) on Congo red medium. These nonmain CR+ subspecies strains also produced visible biofilms on the walls of polystyrol dishes. Absorbance of the stained cells attached to dishes and solubilized in ethanol/acetone mix was high, ranging from 1.0 to1.3 OD, almost equal to that observed in the main subspecies CR+ strains.
In the case of the other nonmain subspecies strains, not all colonies formed on Congo red medium were pigmented. For the subspecies altaica strain I-2359, only 30% of the colonies were pigmented and for subspecies ulegeica strain I-3131 and subspecies hissarica strains A-1723, A-1725, A-1249) 50% of the colonies were pigmented. This percentage was preserved in the subcultured populations of CR+ as well as of CR− colonies. This result indicates the necessity to further investigate the regulatory mechanism responsible for the lack of pigmented colonies formed in the altaica, ulegeica and hissarica ssp. Interestingly, the differences in percentage of pigmented colonies between strains of the nonmain subspecies had no significant effect on biofilm production on abiotic surfaces. The strains forming 30% and 50% pigmented colonies also produced visible biofilms in polystyrol dishes. Absorbance of their solubilized biofilms was 0.9–1.0 OD, only a little lower than that observed in the strains demonstrating 100% pigmentation.
In our experiments, two strains of the nonmain subspecies, ulegeica (I-3069) and altaica (I-2998), reproducibly demonstrated the CR− phenotype. These strains failed to produce biofilms on abiotic surfaces (absorbance <0.1 OD).
In all nonmain subspecies strains studied, positive signals for the hmsF and hmsS genes were observed by PCR independent from their CR phenotypes, confirming the presence of these genes in these strains. Earlier, in our laboratory, whole hmsHFRS operons of some of the nonmain subspecies strains including strains I-3069 (CR−, ulegeica) and I-2998 (CR−, altaica) were sequenced. The nucleotide sequences of these nonmain subspecies CR− strains and of the main subspecies CR+ strains, were identical with the exception of the caucasica strains. The caucasica strains contained two single nucleotide substitutions, one in hmsF and the other in hmsH (Sukhonosov & Krasnov, 2007; Sukhonosov et al., 2007). These results showed that the CR− phenotype of the two altaica and ulegeica strains was not caused by deletions of or point mutations within the hms genes. PCR analysis also demonstrated the presence of the heptose biosynthesis gene (gmhA), the polyamine putrescine biosynthesis genes (speA and speC) as well as the regulatory genes (hmsP and hmsT) in genomes of all nonmain subspecies strains studied (Table 1). Thus the inability to form biofilms on abiotic surfaces by some of the nonmain subspecies Y. pestis strains is not related to the absence of the genes associated with these properties, the hmsHFRS operon and the regulatory genes hmsP, hmsT, as well as the gmhA, speA, speC genes. Preliminary data obtained in our laboratory show that the nucleotide sequences of hmsT and hmsP in the two CR− strains of the altaica and ulegeica ssp. do not differ from those in other strains of these subspecies (not shown).
It therefore seemed interesting for us to study the structure of another gene, rcsA, in the nonmain subspecies strains as rcsA has recently been shown to be a negative regulator of biofilm formation in fleas and a pseudogene in Y. pestis (Sun et al., 2008). We performed comparative sequence analysis of the rcsA gene in Y. pestis KIM, CO92, Pestoides F and Y. pseudotuberculosis IP32953 and YPIII present in the NCBI GenBank database. This analysis revealed that the Y. pestis rcsA gene contained a 30-bp insertion when compared with Y. pseudotuberculosis. We then examined the gene structure in natural Y. pestis strains using primers constructed for a variable gene locus. Results showed that the 30-bp insertion sequence was not only present in the rcsA gene in the main subspecies strains, but in all strains of caucasica, altaica, hissarica and ulegeica ssp. (Fig. 3). These data indicate the importance of switching off the rcsA gene function during the early stages of Y. pestis evolution (in the nonmain subspecies) and show that destruction of rcsA was a result of negative selection.
Close correlation between pigmentation and biofilm production in the nonmain subspecies strains was also confirmed using the C. elegans model. CR− strains of the nonmain subspecies Y. pestis strains (ulegeica I-3069, altaica I-2998) did not form biofilms on nematode cuticles and did not block the worms (Fig. 4a). In the strains that showed a reproducible difference in the percentage of colony pigmentation (30% observed in altaica I-2359 and 50% in ulegeica I-3131 and hissarica A-1723, A-1725, A-1249), biofilm formation was still observed, but blocked fewer nematodes (5.3±1.15%), although most of the worms retained their motility. The same amount of blocked nematodes was seen on the lawns seeded with the 100% CR+caucasica strains (818, 1146 and 3544 Arm).
On the lawns of 100% CR+ (ulegeica I-2422, I-3130, altaica I-2183, hissarica A-1728) strains, blocked C. elegans nematodes were also seen among freely moving nematodes (Fig. 4b). The number of blocked nematodes was higher (8±1.04%) than that observed for the strains forming 30% and 50% pigmented colonies, but smaller than that for the main subspecies CR+ strains.
In summary, most of the nonmain subspecies Y. pestis strains were able to produce biofilms on both abiotic and biotic surfaces and this ability correlated with the formation of pigmented colonies on solid Congo red medium. Evidently, these strains have another regulatory mechanism responsible for the formation of CR− colonies, differing from that of the main subspecies strains. It is important to note that the difference in ability to form pigmented colonies has no influence on the virulence of the nonmain subspecies strains, as they are reportedly all virulent in mice, but avirulent in guinea pigs. Selective virulence for laboratory animals is a characteristic feature of the nonmain subspecies strains (Martinevskii, 1969).
Pigmentation and biofilm formation by Y. pseudotuberculosis strains
To compare biofilm production in Y. pestis and its precursor, Y. pseudotuberculosis, we examined four strains of the latter. In this case, we failed to find a distinct correlation between pigmentation and biofilm production on abiotic surfaces. The strains used were Y. pseudotuberculosis 2600 and 417 (from Turkmenia) and 50–73, 312 (from the Far East of the Russian Federation). These strains did not produce pigmented colonies on Congo red medium, but they did produce biofilms on polystyrol dishes. A570 nm of solubilized biofilms was equal to 1.1–1.3 OD and comparable to that of the CR+Y. pestis strains. Meanwhile, a correlation between pigmentation and biofilm production was observed on biotic surfaces. All four CR−Y. pseudotuberculosis strains failed to produce biofilms on C. elegans nematode cuticles. Worms freely moved on their lawns in the usual sinusoidal way. These results are in accordance with the literature, which reports that most Y. pseudotuberculosis strains are not able to block the nematode C. elegans under laboratory conditions. Joshua et al. (2003) showed that of 41 Y. pseudotuberculosis strains belonging to 21 serovars, 76% did not produce biofilms on nematode cuticles. Evidently, the four isolates used in our study belong to the group of Y. pseudotuberculosis strains that do not produce biofilms on biotic surfaces under laboratory conditions.
Genes of the hms operon, the regulatory genes hmsP and hmsT, as well as the genes gmhA, speA and speC, were detected in the four Y. pseudotuberculosis strains by PCR; however, the rcsA gene of these strains did not contain a 30-bp insertion (Fig. 3). Thus, in the Y. pseudotuberculosis strains studied, the CR− phenotype was neither caused by deletion of the hmsHFRS operon nor by mutation of any of the other genes examined. It is likely that mechanisms for biofilm formation on abiotic and biotic surfaces in Y. pseudotuberculosis are different. They also differ from those of Y. pestis strains.
To conclude, biofilm formation on abiotic and biotic surfaces was studied in natural main and nonmain subspecies Y. pestis strains. Strains of the highly virulent main subspecies pestis (plague agent) formed biofilms on polystyrol dishes and on C. elegans nematode cuticles. The number of blocked nematodes on the lawns of the main subspecies strains was higher than in the nonmain subspecies strains. Biofilm formation in the main subspecies strains strictly correlated with the presence of the hms operon and the CR+ phenotype. Deletion of the hms genes was the reason for the lack of biofilm production in the main subspecies CR− strains.
The majority of the nonmain subspecies strains also produced biofilms on both abiotic and biotic surfaces and these properties also correlated with the CR+ phenotype. Two CR− strains failed to produce biofilms on abiotic surfaces and on nematode cuticles, although these strains contained genes from the hms operon as well as the hmsT, hmsP and other related genes. This indicates that the CR− phenotype of these strains was caused not by a deletion of the hms operon but rather by other, as yet unknown, regulatory mechanisms. The nonmain subspecies Y. pestis strains capable of forming 30%, 50% and 100% pigmented colonies, produced biofilms on nematode cuticles, but in smaller amounts than those produced by the main subspecies CR+ strains. It is likely that the evolution of Y. pestis towards the highly virulent main subspecies pestis required the enhanced ability to form biofilms on biotic surfaces, which was necessary to ensure effective transmission of the plague agent by the flea vector.
This work was supported by the grant of RFFR N 08-04-00731.
G.A.E, N.A.V. and V.V.K. contributed equally to this work.