OUP user menu

Yersinia HPI in septicemic Escherichia coli strains isolated from diverse hosts

U. Gophna, T.A. Oelschlaeger, J. Hacker, E.Z. Ron
DOI: http://dx.doi.org/10.1111/j.1574-6968.2001.tb10540.x 57-60 First published online: 1 March 2001


High pathogenicity islands (HPIs), first identified in various Yersinia species, encode an iron uptake system. We have studied the occurrence of HPIs in septicemic strains of Escherichia coli isolated from a variety of hosts. The results presented in this communication indicate that most septicemic strains tested contained HPI sequences even though they already have the aerobactin encoding genes. We have also observed two types of HPI deletions, suggesting genetic instability of this element. Notable exceptions are several strains isolated from septicemia in sheep that lacked both iron acquisition systems.

  • High pathogenity island (HPI)
  • Yersinia
  • Escherichia coli

1 Introduction

Iron is essential for the metabolism of most known bacteria, including pathogens, with the notable exceptions of lactobacteria and Borrelia burgdorferi [8]. In the body fluids of warm-blooded animals iron is almost entirely bound to the iron transport proteins transferrin and lactoferrin. As a result, the free iron concentration under these conditions is insufficient for bacterial survival and multiplication, and the pathogens depend on their ability to scavenge iron.

In Escherichia coli, the pervasive iron uptake mechanism is the production of siderophores, low molecular mass iron chelators. Most known E. coli strains produce the catecholate siderophore enterobactin [9], and many clinical isolates also possess the hydroxamate siderophore aerobactin [13]. Recently, a third siderophore, yersiniabactin, coded by the Yersinia high pathogenicity island (HPI), was found in E. coli strains pathogenic to humans [6,12]. Efficient iron uptake is most essential for bacteria which enter the bloodstream. We determined the prevalence of HPI in E. coli strains causing septicemia in various hosts including mammals and birds. The results provide additional evidence supporting the role of HPI in systemic infection.

2 Materials and methods

2.1 Bacterial strains and media

Wild-type E. coli strains were isolated from sheep, poultry and humans with acute septicemia. Four–nine independent isolates were examined in each experiment. E. coli strain 536 had been described elsewhere [3].

2.2 Detection of HPI and iucD genes by PCR

Single colonies were picked into 1 ml of deionized water. Samples were incubated at 98°C for 10 min and the suspension was used as a template. Amplifications were carried out in a total volume of 50 μl using 5 μl of bacterial suspension, each deoxynucleoside triphosphate at a concentration of 0.25 mM, 10 pmol of each primer, 5 μl of 10-fold PCR buffer (Takara) and 2.5 U of Taq DNA polymerase (Takara). Reaction conditions used were: 10 min denaturation at 94°C, then 30 cycles of 1 min denaturation at 94°C, 1 min annealing at the specific temperature (see Table 1), and 1 min extension at 72°C and after completion of the cycles, an additional 10 min at 72°C. A PTC100 programmable thermal cycler (MJ research Inc.) was used for all reactions. For HPI detection PCR reactions, E. coli strain 536, whose HPI region was previously characterized [11], was used as positive control template and E. coli K-12 strain C600 was used as a negative control. For iucD detection, the same negative control was used along with strain 781 as a positive control, previously shown by Southern hybridization to contain the aerobactin encoding cluster aer [1].

View this table:
Table 1

PCR primers used in this study

Primer designationSequence of primerTarget geneAnnealing temperatureLength of product (bp)Reference
iucDF5′-TGCAAAATCCGCTGTGGCTGGTA-3′iucD561295This study

2.3 DNA sequencing

Automated DNA sequencing was performed on double-stranded DNA templates by the dideoxynucleotide chain-termination method [10] with an Applied Biosystems model 373A sequencer (Foster City, CA, USA), or a LiCor automatic sequencing system (MWG).

3 Results

3.1 Presence of the Yersinia HPI in septicemic E. coli serotype O78 strains from various hosts

E. coli strains of the serotype O78 cause a variety of diseases, among them sepsis, in humans, poultry, and sheep ([1,4]). We examined strains causing avian colisepticemia in chickens and turkeys, strains isolated from septicemic sheep and strains from newborn meningitis in humans for the presence of the Yersinia HPI. PCR was conducted using specific primers for irp2 located in the middle of the HPI and encoding the high molecular mass protein 2 (HMWP2) assumed to be involved in production of the siderophore yersiniabactin, and for the fyuA from the gene encoding the pesticin and yersiniabactin receptor located on the HPI's 3′ extremity. PCR was also performed to test for the presence of the integrase gene int located in the 5′ extremity of the HPI adjacent to the HPI integration site. The results are summarized in Table 2.

View this table:
Table 2

HPI genes in septicemic E. coli serotype O78 strains

HostPositive strains (%)
Region amplified

The results indicate that all isolates from humans and nearly all isolates (89%) from poultry contained the HPI. In contrast, none of the ovine strains tested contained the HPI genes. PCR products obtained from most strains had the predicted lengths. Interestingly, the PCR products obtained from the int gene of 22% of the avian strains were about 350 bp smaller than those of the other strains (850 instead of 1200 bp), indicating a deletion in the int gene as can be seen in Fig. 1. The PCR fragment was sequenced and the deletion was found to be identical to a deletion in int previously identified in Shiga toxin E. coli serotype O26 strain 5720/96 [6].

Figure 1

Agarose gel electrophoresis of PCR products from the int gene. Lane 1, K-12 strain C600; lane 2, human O78 strain from NBM; lanes 3 and 4, O78 isolates from avian colisepticemia. Epicentre size marker (lane #).

3.2 Presence of the Yersinia HPI in avian E. coli septicemic strains of various serotypes

To examine whether high prevalence of HPI was correlated with the O78 serotype or whether it is also common among avian septicemic strains of other serotypes, we screened isolates belonging to serogroups O2, O35 and O111 for the HPI genes as described above. The results, presented in Table 3, indicate that in serotype O2, which along with O78 is the most common in colisepticemia in poultry, all isolates contained the HPI genes. It should be noted that strain oh2 has a unique gene rearrangement: it contains the int gene, but the irp2 and fyua genes were not detected. These findings indicate that the strain has a dysfunctional HPI cluster which does not contain the regions encoding the siderophore yersiniabactin or the pesticin receptor. Nevertheless, additional PCR reactions indicated that this isolate does possess another part of the yersiniabactin biosynthetic gene cluster, namely the ybtS and ybtQ genes. HPI sequences were seldom detected in the other two serotypes examined. These results show that strains of serotypes most frequently associated with avian colisepticemia are also more likely to possess the HPI gene cluster.

View this table:
Table 3

HPI genes in non-O78 avian septicemic E. coli strains

SerotypePositive strains (%)
Region amplified

3.3 Presence of the aerobactin biosynthesis gene cluster in septicemic strains

Many E. coli strains which cause septicemia and meningitis have been shown to possess the aerobactin biosynthesis genes, usually encoded on ColV plasmids in those strains [5,13]. It was therefore interesting to determine whether the aerobactin cluster would be more common in non-HPI strains, as a substitute iron uptake system, or whether the aerobactin system will be just as prevalent in HPI strains, suggesting two independent parallel systems. A screen using PCR for the iucD gene in the aerobactin biosynthetic pathway was used to detect strains potentially carrying the aerobactin iron uptake system.

As can be seen from Table 4, the majority of septicemic strains (22 out of 26) contain the iucD gene, suggesting that they can produce aerobactin. The presence of this iron uptake system did not preclude the existence of the yersiniabactin system – 18 out of 26 strains contained both sytems. It is interesting to note the existence of avian colisepticemic strains that contained only the HPI system and not the aerobactin. Another observation indicates that most ovine strains, all of which were isolated from colisepticemic sheep, did not carry either of the two iron uptake systems associated with virulence.

View this table:
Table 4

Presence of aer gene iucD in septicemic E. coli strains

HostStrains (%)
NoneHPI onlyaer onlyHPI+aer

4 Discussion

The HPI, originally identified and described in Yersinia sp., was recently identified in E. coli strains pathogenic to humans as well as in Citrobacter diversus and various species of Klebsiella [2,12]. In this report we have screened septicemic strains from both human and animals to determine whether distribution of the HPI may be host-specific.

Our comparison the prevalence of the HPI genes in O78 strains from humans, sheep and poultry indicates that even within the same serogroup, presence of the HPI differs between isolates from different hosts. Specifically, none of the strains isolated from sheep was identified by PCR as containing HPI sequences. The absence of HPI genes in ovine strains may be due to different selective pressures than those present in other hosts or simply because acquisition of the HPI by horizontal gene transfer did not take place. The int gene coding a putative integrase is considered as possibly linked to the mobility of HPI [11]. The fact that two of the avian strains contained a deletion in the int gene, formerly identified in Shiga toxin E. coli serotype O26 strain, is interesting because it may indicate these strains have acquired the HPI by a different horizontal gene transfer event than the other avian strains. These two strains were shown previously to be genetically related by pulse field gel electrophoresis [1].

The screening for HPI genes in non-O78 serogroups of avian septicemic E. coli suggests that HPI prevalence is correlated with the serogroup. Interestingly, nearly all the isolates from the serogroups commonly associated with avian colisepticemia, O2 and O78, contained the HPI genes. The finding that one isolate, oh2, contains a truncated HPI region which apparently lacks at least parts of the irp2 and fyuA genes is not surprising since E. coli strains containing a truncated fyuA-irp gene cluster were previously identified [12].

The presence of aerobactin biosynthetic genes in avian septicemic E. coli is in agreement with previous reports which found that most of these strains carry a plasmid-encoded aerobactin iron uptake system [7]. However the lack of both yersiniabactin and aerobactin systems in all but one of the ovine strains tested is surprising, since it has been claimed that aerobactin is much more stable than enterobactin and less pH dependent, and is therefore essential for survival of invasive strains within the host [13]. Since both the usually plasmid-encoded aer gene cluster is and the HPI are considered to be unstable genetic regions, lack of both siderophores in those strains may indicate selective pressure.


This work was supported by the German Israeli Foundation (GIF). We thank Dr. D. Zhang for his kind assistance and advice.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
View Abstract