Yersinia

The pathogenic Yersinia species, Y. pseudotuberculosis, Y. enterocolitica, and Y. pestis, are

zoonotic agents that cause disease in humans ranging from mild gastroenteritis to lifethreatening

plague. Human clinical infections caused by Y. pseudotuberculosis and Y.

enterocolitica most frequently occur after the ingestion of contaminated food or water,

whereas the etiologic agent of plague, Y. pestis, is naturally vectored to humans by fleas.

These species are joined by 11 lesser-known Yersinia species which are largely considered

environmental species, nonpathogenic to humans (Table 1). Pathogenic Yersinia species

share a highly conserved virulence plasmid and a chromosomal high pathogenicity island

(HPI) and show tropism for lymphoid tissue, where their ability to evade host innate

immunity enables extracellular proliferation. Popular awareness of the tens of millions of

human deaths caused by the plague bacillus, Y. pestis, in each of three pandemics is

contrasted by relative ignorance of the current global reach of plague, its firm entrenchment

in a number of recent rodent-flea cycles covering much of the western United States and

other parts of the world, and its reemergence as a significant human pathogen in parts of

eastern Africa and Madagascar. This agent was weaponized by the United States, Japan, and

former USSR during and after World War II and remains a high-level biothreat agent.

Increased biomedical research, in response to these concerns, has led to a number of recent

advances in our understanding of invasiveness and immune system evasion by Y. pestis and

the closely related Y. pseudotuberculosis and Y. enterocolitica and exciting new laboratory

diagnostic and vaccine development approaches (10, 13, 59, 68).

TAXONOMY AND HISTORY OF THE GENUS Back to top

The agent of plague, Y. pestis (previously Bacterium pestis, Bacillus pestis, and Pasteurella

pestis), was first identified in lymph node aspirates from plague victims in 1894 by Alexander

Yersin during his investigation of the leading edge of the third pandemic as it swept quickly

from mainland China to Hong Kong and beyond. Further, he demonstrated that laboratory

mice and rats inoculated with pure cultures of the gram-negative coccobacillus produced

symptoms of plague and death. Four years later, while investigating the advancing epidemic

in India, Paul-Louis Simond and Masanori Ogata independently implicated the role of fleas as

vectors of disease transmission between rodents and humans. By the early 1900s, plague

was introduced along steamship routes to every human-inhabited continent and was

responsible for more than 12 million deaths, most of them in India and China (59). Today,

plague remains enzootic in parts of Africa, North and South America, and much of Asia.

By 1944, cumulative phenotypic and genotypic peculiarities of Pasteurella pestis and P.

pseudotuberculosisprompted their transfer to the new genus Yersinia. Ten years

later, Yersinia species were included in the familyEnterobacteriaceae, and in 1964 Y.

enterocolitica was added to the genus. The diverse group members of Y. enterocolitica were

further divided into a subgroup initially designated Y. enterocolitica-like organisms and later

into an additional four species (Y. intermedia, Y. frederiksenii, Y. kristensenii, and Y.

aldovae) based on sugar fermentation and DNA relatedness. Subsequent species

designations include Y. ruckeri (serogroup 01, the agent of enteric red mouth disease in

rainbow trout), Y. rohdei, Y. mollaretii, and Y. bercovieri. Three additional species have been

described since 2005, bringing the total to 14 (www.bacterio.cict.fr/xz/yersinia.html).

Yersinia species have a G+C content of 46 to 50%. By DNA-DNA

hybridization, Yersinia species are related to other members of the

family Enterobacteriaceae by 10 to 32%. Intraspecies relatedness is variable, ranging from

55 to 74%, with the notable exception of Y. pestis and Y. pseudotuberculosis, which

demonstrate more than 90% relatedness.

The sequenced and annotated genomes of Y. pestis biovar Orientalis and biovar Medievalis

were published in 2001 and 2002, respectively (22, 57). Since then the genomes of at least

a dozen additional Y. pestis strains have been released (13). In 2004, the Y.

pseudotuberculosis genome was published along with a comparative genomic analysis of Y.

pestis (16). Completion and analysis of these genomes have solidified a large body of prior

genetic evidence suggesting that Y. pestis evolved from Y. pseudotuberculosis prior to the

first plague pandemic (1, 2, 13, 57, 81, 82). Were it not for the profoundly distinct

mechanisms of pathogenesis and natural maintenance, taxonomists would likely merge Y.

pseudotuberculosis and Y. pestis. The genome of Y. enterocolitica was published in 2006 (76)

and has enabled further understanding of the common and unique genetic lineages and

functions of the pathogenic Yersinia species.

DESCRIPTION OF THE AGENTS Back to top

Yersinia species, as members of the family Enterobacteriaceae, are gram-negative, nonspore-

forming bacilli that exhibit bipolar staining particularly when seen in primary

specimens stained with Giemsa or Wayson’s dye. The bacilli are smaller (0.5 to 0.8 μm in

diameter and 1 to 3 μm in length) than other members of their family and tend to grow more

slowly as well. With the exception of Y. pestis, Yersinia species are motile at 25°C due to

peritrichous or paripolar flagella (Table 1). Interestingly, the pathogenic yersiniae either

repress(Y. pseudotuberculosis and Y. enterocolitica) or, through mutation, have lost the

capability to express (Y. pestis)the flagellar apparatus; by either mechanism they effectively

avoid stimulation of innate immune responses to these potent inducers (52).

Yersinia species are facultative anaerobes that grow at temperatures ranging from 4 to 43°C.

Although optimal doubling times are observed between 25 and 28°C, the ability to replicate

at 4°C (psychrophile) has important ramifications for blood banking among asymptomatic Y.

enterocolitica-carrying donors (see “Clinical Significance” below). Yersinia species ferment

glucose with the production of acid and no gas and are catalase positive and oxidase

negative. Most strains grow on MacConkey, blood, and chocolate agars but may be

outcompeted by other bacteria in clinical and especially environmental samples (see

“Isolation Procedures” below). Yersiniae exhibit poor growth in liquid media and do not form

a turbid suspension (79).

The cell walls of Yersinia species are very similar to those of other members of

the Enterobacteriaceae family, and lipopolysaccharide (LPS) is a major component of their

outer membrane. The LPS of Y. pseudotuberculosisand Y. enterocolitica (smooth forms) is

complete (lipid A–oligosaccharide core–O antigen polysaccharide), and O-chain variation

within these species has enabled serodiscrimination of close to 100 LPS types (6). In

contrast, through a genetic defect in the biosynthesis of complete LPS, Y. pestis (rough form)

lacks the O-specific polysaccharide chain (9, 67).

Several dozen virulence genes, their environment- dependent expression control, and the

complex mechanisms of their product action and coordination, which enable immune system

evasion and disease progression, have been actively investigated and described. Only a few

of the key features will be presented here, and the reader is referred to several excellent

reviews for more detailed and comprehensive information (35, 46, 68, 78).

Many Yersinia virulence genes are conserved across species lines, while others are species

specific. Of intrigue, a number of the enteropathogenic Yersinia virulence genes have been

rendered dysfunctional in the Y. pestis lineage, through mutation or insertion sequence

interruption, but have been maintained within the species in the absence of apparent

function. One of the key features of the pathogenicYersinia species is the ability to scavenge

iron from the host by a siderophore called yersiniabactin (Ybt). Theybt genes are

chromosomally located and clustered on an HPI located in a 102-kb chromosomal region

termed the pgm (pigmentation) locus. An additional operon within the pgm region, and

unique to Y. pestis, is the hmslocus. The hemin storage proteins encoded by this locus

enable biofilm formation and proventricular blockage required for efficient transmission of Y.

pestis from vector fleas to mammals (25, 57).

Several other chromosomal virulence genes include yst, which encodes a heat-stable toxin

unique to the pathogenic members of Y. enterocolitica, and invA, an epithelial cell adhesin

gene common to all three pathogenic Yersinia species. In Y. enterocolitica and Y.

pseudotuberculosis, invasin facilitates efficient binding to intestinal mucosal cells and

translocation from the lumen to Peyer’s patches. An insertion element within theinvA gene

of Y. pestis renders it dysfunctional in spite of otherwise >99% nucleic acid identity with the

gene ofY. pseudotuberculosis.

Although each of the pathogenic Yersinia species is associated with different clinical entities,

they possess a common and genetically conserved 68- to 75-kb virulence plasmid (termed

pCD1 in Y. pestis and pYV in Y. enterocolitica and Y. pseudotuberculosis). This plasmid

carries the low calcium response genes, components of the type three secretion system

(TTSS), and the associated effectors or Yersinia outer membrane proteins (Yops) (65). The

TTSS forms a needle structure on the surface of pathogenic Yersinia species which interacts

with target cells (macrophages, dendritic cells, and granulocytes/neutrophils) and enables

injection of six different Yops which effectively interfere with phagocytosis and other innate

host cell responses as well as the adaptive inflammatory cascade, ultimately resulting in

target cell apoptosis (19). Another virulence plasmid gene product, V antigen, is assembled

on the tip of the needle structure and required for injection of Yop effectors (45). Recent

experiments also indicate that the anti-inflammatory properties of V antigen enable Y. pestisinfected

rodent survival while bacterial loads increase to levels beyond 108/ml of blood (23).

Such levels are required for effective infection of blood-feeding fleas (49).

YadA is an adhesion protein also encoded on the Yersinia virulence plasmid. In Y.

enterocolitica and Y. pseudotuberculosis, it too facilitates pathogen binding to M cells of the

intestinal mucosa, signal-induced internalization, and subsequent translocation to the Peyer’s

patches and mesenteric lymph nodes. Y. pestis,which does not enter the epithelial cells of

mammalian or flea hosts, carries the 70-kb plasmid-encoded yadAgene, which is nearly

identical to that of Y. pseudotuberculosis; but like invasin, Y. pestis YadA is dysfunctional, in

this case through a base deletion resulting in a frameshift and a truncated product.

Two Y. pestis unique plasmids (#x007E;110 and 10 kb), acquired after its divergence

from Y. pseudotuberculosis, carry a variety of genes responsible for several differential

attributes of this agent: its ability to utilize a flea vector and its ability to cause acute disease

in the infected mammal. The large plasmid encodes the murine toxin (Ymt) and the fraction

1 antigen (F1). Ymt, which is expressed only at temperatures below those of mammalian

systems, is a phospholipase whose activity is necessary for pathogen survival within the

harsh environment of the flea digestive tract during blood meal digestion (33). In contrast to

Ymt, F1 is expressed at temperatures >30°C and during mammalian infection. Upon flea

inoculation of Y. pestis,some of the bacteria are engulfed and transported to regional lymph

nodes by macrophages. Here, during intracellular growth, F1 is expressed and forms a

capsule-like structure on the bacterial surface. Subsequent to release of Y. pestis from its

intracellular location, the antiphagocytic properties of the new F1 surface enable widespread

dissemination and replication and result in host sepsis. Plasminogen activator (Pla), one of

the small plasmid products, is a surface protease that activates mammalian plasminogen and

degrades complement; it has been shown to be essential for dissemination by subcutaneous

or intradermal inoculation, experimental proxies for flea-bite transmission, in mouse models

of bubonic disease. These are but a few of the Y. pestis unique, plasmid-encoded proteins

differentially expressed within the temperature and environmental constraints of its

arthropod and mammalian hosts which enable pathogen survival, replication, and

dissemination (57).

EPIDEMIOLOGY AND TRANSMISSION Back to top

Plague is an acute, often fatal disease caused by Y. pestis. It exists in natural enzootic cycles

between wild rodents and their fleas. The most common mode of transmission to humans is

by the bite of infected fleas. Less frequently, infection is the result of handling infected

animals, direct contact with infectious body fluids or tissues, or inhaling infectious respiratory

droplets or other materials (e.g., laboratory-acquired cases). Epidemics occur occasionally

when the disease spreads from wild rodents into populations of rats (genusRattus) that live

near human populations. Human risk is greatest when epizootics cause high mortality in

these commensal rat populations, thereby forcing fleas to seek alternate hosts, including

humans. Although many different species of fleas parasitize rodents and can transmit

infection to them, transmission of plague is classically associated with the rat

flea (Xenopsylla cheopis) (15). Y. pestis can survive and multiply in the midgut (stomach)

and proventriculus (a valve that connects the esophagus to the midgut) of this flea. Bacterial

replication results in blockage of the proventriculus and causes the infected flea to bite

multiple hosts in repeated attempts to acquire a blood meal, thereby increasing chances for

disease transmission. Interestingly, when the environmental temperature is above 27°C, Y.

pestis does not produce coagulase and blockage of the proventriculus is unlikely to occur,

thus reducing transmission to humans (77).

Approximately 2,500 human cases of plague are identified every year around the world.

Although routine reporting of cases to the World Health Organization was discontinued in

2003, between 1989 and 2003, 38,359 human cases with 2,845 deaths were reported from

25 countries. Approximately 80% of these cases were reported from Africa, 15% from Asia,

and the remainder from the Americas. Between 1960 and 2006, 447 plague cases (~9

cases/year) were reported in the United States. Although plague occurs among wild animal

populations in all of the 17 contiguous states west of the 100th meridian, more than 80% of

human cases occur in New Mexico, Arizona, and Colorado, and approximately 10% occur in

California (20, 29, 51).

Domestic dogs and cats may serve as carriers of Y. pestis-infected fleas into human

dwellings. A recent study in New Mexico found that plague victims were significantly more

likely than controls to have allowed pets to sleep on their beds (32). Cats are also highly

susceptible to plague, acquiring infection by flea bite or ingestion of infected rodents, and

develop all of the clinical forms of disease, including pneumonic plague. Infected cats serve

as direct sources of human infection via aerosol transmission of organisms, leading to

primary pneumonic plague, or by inoculation of organisms through scratches or bites (29).

Dogs are most often resistant to clinical infection and develop antibody titers upon exposure.

Hunting in areas where plague is endemic is associated with occasional human infections.

Infected animals may not display external or internal signs of disease, and as such,

transmission may occur during skinning and handling of animals or ingestion of undercooked

meat. Although many animal species are susceptible to plague infection and thus pose a risk

factor associated with hunting in areas of endemicity, documented cases of transmission are

most frequently associated with rabbits, squirrels, prairie dogs, and bobcats. Direct

inoculation into the bloodstream is associated with an increased risk of septicemia and a high

fatality rate.

Human-to-human transmission is very rare and occurs only with the pneumonic form of

disease; most transmissions are to unprotected/untrained family members and health care

providers through close-contact inhalation of particles from infected persons who are

coughing copious amounts of bloody sputum. In a number of reports, simple measures such

as avoidance of close (<2 meters) contact and/or use of surgical masks and gloves are

reported as protective (36, 42). These findings are consistent with limited-distance spread

of Y. pestis-contaminated respiratory droplets from pneumonic plague patients as opposed to

greater dispersal distances of droplet nuclei among other respiratory pathogens, such

as Mycobacterium tuberculosisfrom patients with pulmonary tuberculosis. In resource-poor

regions, protection of health care providers is often limited to patient contact in wellventilated

wards, minimal time visits, and during such visits, requiring patients to turn their

heads away from the provider or examining them from behind (4, 36, 42, 62). In over 60

separate cases of pneumonic plague in the United States since 1924, there have been no

human-to-human transmissions (36).

In recorded history there have been three plague pandemics, each originating in different

parts of the known world. The earliest pandemic, known as the Justinian plague, spanned the

sixth through eighth centuries, began in northeast Africa, and spread through the Middle

East and the Mediterranean basin. Methodical and accurate mortality records were not

attempted in most affected areas, and competing epidemics such as smallpox have

complicated estimates of plague cases and deaths. Nonetheless, it is ventured that from A.D.

541 to 700 there was an overall population loss of 50 to 60% in many locales of affected

regions. Historical records of the second pandemic are comparatively numerous and in many

cases detailed. Beginning in the Himalaya region of central Asia in the mid 1300s and

spreading westward along overland trade routes, it entered Sicily in 1347. The ensuing

epidemic, later known as the Black Death, spread quickly by sea to Italy, Greece, and

France, and later by land throughout Europe. Between 1347 and 1351, an estimated 17 to

28 million Europeans died of plague. Following epidemics occurred in periodic cycles for over

300 years. Institution of “quarantine,” the 40-day isolation of ships and people while waiting

to see if symptoms of plague would develop prior to allowing city entrance, originated in

Venice in 1377. This and other attempts to slow the spread of plague were futile; the

pandemic killed up to one-third of the European population between the 14th and 15th

centuries and encompassed the Near and Middle East, Europe, and the British Isles. The

cause of plague was still unknown. The third, or modern, plague pandemic originated in

southern China in the mid- to late 1800s and spread globally, reintroducing the disease to

many regions and establishing new foci in North and South America as well as Madagascar

(10, 59, 84).

Four biovars of Y. pestis, three of which cause human disease, persist today and are

differentiated by their ability to ferment glycerol or arabinose and reduce nitrate (see

“Identification” below). Human cases are associated with Antiqua strains found in Asia and

Africa, Medievalis strains found in Asia (Near East through the former Soviet Union to China

and Mongolia), and Orientalis strains distributed globally. A fourth biovar, Microtus (also

known as Pestoides), is pathogenic to the rodent genus Microtus (voles), mice, and other

small rodents but not to large mammals, including humans (39a, 79a, 83). Y. pestis lineage

analysis of available strains from the four biovars using multiple-locus variable-number

tandem repeat analysis, insertion sequence elements, and single-nucleotide polymorphism

genotyping suggest that biovar Microtus is the closest ancestral derivative of Y.

pseudotuberculosis (2).

Y. enterocolitica is a heterogeneous species with worldwide distribution and both pathogenic

(human and animal) as well as nonpathogenic members. Organisms are found in the

gastrointestinal tracts of many animal species, most commonly swine, rodents, and dogs.

Due to their enhanced growth in cold temperatures, geographic distribution is mostly in

subtropical and temperate regions. Food products, particularly raw and undercooked meats,

are frequently found to contain these organisms, although the majority of them are

nonpathogenic. The genus is divided into six biogroups—1A, 1B, 2, 3, 4, and 5—that are

biochemically differentiated (Table 2). Group 1A (which lacks the 68- to 75-kb virulence

plasmid) is nonpathogenic, groups 2 through 5 (which lack the chromosomal HPI) are weakly

pathogenic in mice, and group 1B is highly pathogenic and lethal to mice. Group 1B has been

frequently isolated in North America, whereas groups 2 through 5 are predominantly isolated

in Europe and Japan. Serogroup typing, based on reactivity to O-antigen polysaccharides,

has also been developed, and over 70 serotypes are characterized. Many serotypes are

geographically focused, and only a few are associated with disease in animals or humans.

CLINICAL SIGNIFICANCE Back to top

Plague, one of about ten internationally quarantinable diseases, is a severe febrile illness

characterized by acute onset, headache, myalgia, malaise, shaking chills, prostration, and

gastrointestinal symptoms, and without prompt and appropriate antibiotic treatment, it is

often fatal. A formalin-killed whole-cell plague vaccine which demonstrated efficacy for

bubonic but not primary pneumonic disease was licensed and available in the United States

until 1999. Recent vaccine efforts targeting pneumonic protection have utilized recombinant

F1 and V antigens in cocktail or chimeric (rF1V) formulations. Both formulations have

demonstrated good protection among immunized mice against pulmonary challenge, and

they both appear to be safe, well tolerated, and immunogenic in human trials. Potential

licensure by the U.S. FDA will be in accordance with the “animal rule” which requires safety

and immunogenicity data from humans along with efficacy in animal models that mimic

human disease (68).

There are three major forms of plague: bubonic, septicemic, and pneumonic. Bubonic plague

is the most common clinical presentation, accounting for 80 to 90% of cases and

characterized by acute development of regional lymphadenopathy. During the bite of an

infected flea, up to 104 organisms may be inoculated into the skin. Organisms evading

neutrophil uptake and other innate surveillance and killing are transported to regional lymph

nodes, mainly in macrophages. After a 2- to 8-day incubation period, the patient develops

fever and a painful lymph node swelling (bubo). The case fatality rate for untreated bubonic

plague cases is 50 to 60%.

Septicemic plague can occur when the organisms inoculated by the infected flea spread to

the bloodstream without localizing in regional lymph nodes or when the bacteria are directly

introduced into the bloodstream via a cut or wound. This form of the disease is more

common in children and is rapidly fatal. Septicemic plague can also occur secondary to

bubonic plague that is not adequately treated. In the United States between 1947 and 1977,

approximately 10% of plague cases presented with septicemia and approximately 50% of

these cases were fatal (30).

Pneumonic plague can be a rare secondary complication of bubonic or septicemic plague or

can be the primary infection following direct inhalation of aerosolized organisms from other

pneumonic cases (human or animal), infected tissues, or cultured organisms. After an

incubation period of 1 to 6 days, symptoms include high fever and cough with hemoptysis

and chest pain (36). Mortality among untreated pneumonic cases is essentially 100%, and

even among treated cases mortality often exceeds 50%. Sporadic person-to-person

pneumonic outbreaks continue to occur in Madagascar, Uganda, the Democratic Republic of

the Congo, and Tibet (4, 62). In 2007, a fatal case of primary pneumonic plague occurred in

a Grand Canyon National Park wildlife biologist who had necropsied an infected mountain lion

carcass 7 days earlier (80). Between 1925 and 2006, 13 cases of primary and 52 cases of

secondary pneumonic plague were described in the United States. Among the primary cases,

nine had known exposures: six (67%) were from face-to-face contact with infected pets, and

three (33%) were acquired in laboratory settings. Person-to-person transmission of

pneumonic plague in the United States has not been reported since 1924.

The most common form of disease due to Y. enterocolitica is gastroenteritis associated with

consumption of contaminated food or water. It is not unusual to isolate Y. enterocolitica from

raw meats, including beef, lamb, pork, and chicken. The organism has also been found as a

contaminant of cooked, prepackaged deli meat. The majority of strains isolated from human

food sources are of the nonpathogenic serotypes. Carriage of the pathogenic serotypes of Y.

enterocolitica is more common in swine; therefore, consumption of raw or undercooked pork,

such as chitterlings, is the main risk factor for gastroenteritis (12, 28, 39). Severity of

disease is related to the serotype and can range from self-limited gastroenteritis to terminal

ileitis and mesenteric lymphadenitis, often misdiagnosed as appendicitis. Young children

most commonly develop gastroenteritis and present with fever, watery diarrhea (occasionally

bloody and severe), and abdominal pain following consumption of food contaminated

by Yersinia species. The heat-stable toxin Yst has been indentified in all enteropathogenic Y.

enterocolitica strains but is absent in Y. pseudotuberculosis (81). Although symptoms

typically resolve within approximately 7 days, patients can carry the organism in their

gastrointestinal tracts for as long as several months. Organisms can migrate out of the gut

via the lymphatics into local lymph nodes. An uncommon complication of gastroenteritis is

septicemia. Persons at high risk for septicemia include the elderly and immunocompromised

patients, particularly those with underlying metabolic diseases that are associated with iron

overload, cancer, liver disease, and steroid therapy.

The production of urease allows Y. enterocolitica to survive in the stomach and colonize the

small intestine of the human host. Pathogenic strains contain Yops which enable them to

resist the normal phagocytic and complement killing process that takes place in Peyer’s

patches (73). Y. enterocolitica is the most common cause of transfusion-related infections

due to contaminated red blood cells. Since the organism is able to survive and multiply at

refrigeration temperatures, donated blood contaminated with small numbers of organisms

from an asymptomatic person can transmit infection to the transfused patient (16, 44).

Reactive arthritis is an uncommon sequela of diarrhea due to Y. enterocolitica. Patients at

increased risk include those who are carriers of the HLA-B27 allele and those with

immunologic disorders. Symptoms appear several days to months after the onset of diarrhea

and may persist for months. Other less common diseases associated withY.

enterocolitica infection include inflammatory bowel disease, most commonly associated with

serotype O:3 (63), and autoimmune thyroid disorders, such as Graves’ disease and

Hashimoto ’s thyroiditis (15). Both Y. enterocolitica and Y. pseudotuberculosis have been

isolated from patients with Crohn’s disease, although a causal relationship has not been

proven (34).

Y. pseudotuberculosis usually produces a self-limiting disease, particularly in children and

young adults. Rarely,Y. pseudotuberculosis can cause mesenteric lymphadenitis that clinically

mimics appendicitis and septicemia and generally occurs in immunocompromised patients

(diabetics and those with liver cirrhosis or iron overload) (21). Long-term sequelae of Y.

pseudotuberculosis infection include erythema nodosum, Reiter’s syndrome, and nephritis. Y.

pseudotuberculosis has also recently been implicated in outbreaks of gastroenteritispseudoappendicitis

associated with consumption of contaminated lettuce and carrots

(40, 55).

The lack of classic virulence markers (see “Description of the Agents” above) in the

other Yersinia species (Table 1) has led to their general classification as nonpathogenic.

Nonetheless, several of them (Y. intermedia,Y. frederiksenii, and Y. kristensenii) have been

isolated from stool specimens of up to 20% of diarrhea patients for whom etiologic agents

were not determined (48). Alternate virulence factors have been suggested for some species,

but their proof requires further study. Possible predisposing correlates of infection include

corticosteroid, acid suppressant, and antibiotic use and an immunocompromised host status

(48, 73).

COLLECTION, TRANSPORT, AND STORAGE OF

SPECIMENS Back to top

In the United States, Y. pestis is classified as a select agent. To transfer, receive, or

possess Y. pestis strains, laboratories must be registered with the Centers for Disease

Control and Prevention (CDC). The registration process includes a U.S. Department of Justice

investigation of all personnel having access to select agents. Clinical laboratories are exempt

from the registration requirement provided that within seven calendar days of identifying one

of these agents, they transfer it to a registered entity and/or destroy the agent on site.

Laboratories identifying an organism as Y. pestis are required to report this finding

immediately to the CDC. Report forms, contact information, laboratory registration

information, and pertinent citations of the U.S. Federal Code may be found at

www.cdc.gov/od/sap. All routine procedures performed on Y. pestis should be done in a

facility with a biosafety level of at least 2 (BSL-2), and clinical laboratories should be aware

of the sentinel-level clinical microbiology laboratory guidelines as outlined by the American

Society for Microbiology (http://www.asm.org). Processes which increase risk for creating an

aerosol, such as liquid culture manipulation, should be performed under BSL-3 conditions. Y.

pestis strains lacking either the pgm locus (a 102-kb chromosome region encoding

yersiniabactin iron transport, the Hms biofilm, and other systems) or the 60-to-85-kb

virulence plasmid (which encodes the TTSS and its effector Yop proteins) are exempt from

select agent regulations in the United States and can be handled under BSL-2 conditions.

However, pgm mutants are still capable of causing disease; by intravenous injection they are

fully virulent, and by peripheral inoculation they require only several log orders increased

inocula or iron supplementation for mortality (in mice) comparable to wild-type infections.

Strains lacking the virulence plasmid are considered avirulent (25,38, 58).

Laboratory confirmation of suspected plague diagnosis can be made by detection or growth

of Y. pestis.Preferred samples and tissues are dependent on the clinical presentation; lymph

node aspirates and blood are recommended in bubonic presentation, blood for septicemic

presentation, and respiratory samples and blood for pneumonic presentation. Patients may

shed organisms into the blood intermittently, so obtaining multiple sets of blood cultures

over a 24-h period increases the sensitivity of detection from this sample source. Blood

cultures should be incubated at both 28 and 35°C to increase chances of recovery of the

organism (3, 36). Cary-Blair medium and swabs offer an excellent transport medium for

preservation of viable organisms if samples cannot be cultured immediately. Other culture

sources for pneumonic cases include throat swabs and throat washing specimens; however,

due to contamination of these specimens with normal biota (see “Isolation Procedures”

below), a bronchoalveolar washing or lavage specimen is preferable (3). Tissue samples from

autopsy specimens, lymph node, spleen, liver, and lung can also be utilized for testing. Blood

remnants or tissue specimens can also be collected from animals suspected to have died

from Y. pestisinfection. Convalescent-phase sera can be collected from animals and humans

to test for antibody to Y. pestis. Flea triturates can also be tested. Specimens should be sent

to the laboratory immediately, and if a delay in transit of more than 2 h is expected, the

sample should be transported at 2 to 8°C.

The appropriate specimens for culture of Y. enterocolitica and Y. pseudotuberculosis as well

as other Yersiniaspecies are stool, blood, or lymph nodes, depending on the disease form

suspected. If food is suspected as the source of an outbreak, the local health department

should be involved in the processing of such specimens. Maintain food at 4°C, and transport

it as soon as possible. Swabs should be transported to the laboratory at 4°C in Cary-Blair,

Amies, or Stuart’s medium. Stool specimens can also be placed in transport media and

should be maintained at 4°C if transport is expected to take longer than 2 to 4 h. Enrichment

broths for recovery of Y. enterocolitica from surface waters have been evaluated, but the

ability to recover clinically important strains from this source is uncertain (17).

ISOLATION PROCEDURES Back to top

Yersinia species grow on most routine media, including blood, chocolate, and MacConkey

agars incubated at 35°C in ambient air. Eosin-methylene blue, xylose-lysine- deoxycholate

agar, and Hektoen enteric agars do not provide any advantage in the isolation of Y.

enterocolitica and the differentiation of Yersinia species from other organisms in the normal

stool biota. Due to their ability to ferment sucrose and the fact that Yersiniaspecies grow

more slowly than most Enterobacteriaceae, a selective medium is recommended for

specifically culturing Yersinia species from nonsterile sites. There are various selective media

for the recovery of Y. enterocolitica, including cefsulodin-Irgasan-novobiocin (CIN) agar,

which inhibits the growth of many otherEnterobacteriaceae, and salmonella-shigelladeoxycholate

calcium chloride agar (27). CIN agar has been found to provide better recovery

rates for Yersinia than either MacConkey or salmonella-shigella agar incubated at room

temperature. Growth of many strains of Y. pseudotuberculosis can be inhibited on CIN agar,

and therefore MacConkey agar is preferred for isolation (27).

Recovery of Y. enterocolitica from food is more difficult than recovery from human clinical

specimens, and samples are usually referred to a public health laboratory. Food must be

enriched with saline (or a selective broth, such as modified Rappaport broth containing

magnesium chloride, malachite green, and carbenicillin [MRB]) at cold temperatures for

approximately 21 days (2 to 4 days in MRB) (43).

For isolation of Y. pestis from nonsterile sources, MacConkey, CIN, and research formulations

(5) are useful, although growth is slower on these than on nonselective agar. As Y.

pestis also grows at 25°C, incubation of cultures at this lower temperature can aid in

isolation from contaminated specimens. Cultures from suspected plague patients should be

incubated for 5 days and up to 7 days if the patient has been treated for more than a few

days with an appropriate antimicrobial.

Y. pestis colonies are slow growing and are only 1 to 2 mm in diameter after 48 h of

incubation, with irregular edges. After 48 to 72 hours, a fried-egg appearance is observed

(Fig. 2). No hemolysis is seen on blood agar media. Viewed with a dissecting microscope, the

colonies are raised with irregular edges, with a “hammered copper” appearance. Organisms

growing in broth appear in clumps along the side of the tube in flocculent or stalactite-like

formations if the tube is not shaken. After 24 h of incubation, the clumps settle to the

bottom of the tube.

IDENTIFICATION Back to top

Yersinia species are catalase positive and oxidase negative and ferment glucose, as do all

other members of the family Enterobacteriaceae. Y. enterocolitica and Y.

pseudotuberculosis can be presumptively identified by reactions on triple sugar iron (TSI)

and lysine iron agar slants. Y. enterocolitica produces a yellow color in the entire TSI tube

without gas production, and Y. pseudotuberculosis produces an alkaline slant and an acid

butt, similar to Shigella. Both species are lysine decarboxylase negative and therefore

produce a yellow butt in lysine iron agar slants. Yersinia species are included in the

databases of some automated systems; however, most databases were established with only

a few Yersinia isolates tested. Automated systems may not adequately

identify Yersinia species (particularly Y. pestis) due in part to their slow growth and

biochemical inactivity. In addition, Y. pestis has been misidentified by automated systems

as Y. pseudotuberculosis and asShigella, Salmonella, and Acinetobacter species. API 20E was

shown to have the highest sensitivity and specificity for the identification of Y.

enterocolitica and Y. pseudotuberculosis (54, 56).

Identification of suspected Y. pestis in the clinical laboratory is based on the identification of

small gram-negative bacilli which are catalase positive and indole, oxidase, and urease

negative as well as characteristic growth on agar, including pinpoint colonies after 24 h on

blood and non-lactose-fermenting colonies on MacConkey agar generally after 48 h. See

“Sentinel Level Clinical Microbiology Laboratory Guidelines” on the American Society for

Microbiology website (http://www.asm.org.) for Y. pestis-specific information, pictures, and

flowcharts. Following suspected identification of Y. pestis, routine clinical laboratories should

notify the local public health laboratory and refer the isolate for confirmatory

testing. Y. pseudotuberculosis and Y. enterocolitica can be differentiated from Y. pestis by

urease activities, which are positive, positive, and negative, respectively. Y.

enterocolitica can also be differentiated from Y. pestis by indole activities, which are positive

and negative, respectively. Identification of Y. enterocolitica can be made based on typical

morphology on CIN agar, reactivity on TSI agar, and urease positivity. Identification of the

other Yersinia species can be performed by biochemical analysis (Table 1; 53).

Y. enterocolitica has six biogroups which can be differentiated based on reactivity to esculin,

indole, D-xylose, trehalose, pyrazinamidase, β-D-glucosidase, and lipase. Although the issue

is controversial, biogroup 1A is thought to be nonpathogenic and biotypes 1B and 2 through

5 are pathogenic. Strains belonging to biotype 1A can be differentiated from the others by

salicin and pyrazinamidase positivity (Table 2; 37). Serotyping could also help determine the

pathogenicity of the isolate, since only a small number of the >70 known serotypes are

pathogenic; however, antisera are not readily available. Other methods that have been

evaluated to determine the pathogenicity of Y. enterocolitica are based on the presence of

the virulence plasmid and include autoagglutination, calcium-dependent growth at 37°C, and

pigmentation on Congo red. Selective media containing Congo red as well as PCR assays

have been evaluated for differentiation of virulent from avirulent strains but are currently

being used only in research laboratories (75).

Several Yersinia species can be differentiated by a number of phenotypic methods, and the

four biovars of Y. pestis can be separated based on differential reactivity with glycerol,

nitrate, and arabinose (Table 3; 47, 83). Recent evidence suggests that biovars based on

phenotypic methods do not show a strict correlation to groupings as determined by

genotyping methods (1).

TYPING SYSTEMS Back to top

Methods used for the evaluation of the relatedness of Yersinia species include a number of

different phenotypic methods, including serotyping, biotyping, antibiogram analysis, and

bacteriophage typing. Y. enterocolitica can be divided into six biogroups—1A, 1B, 2, 3, 4,

and 5—by using biochemical analysis. These biogroups vary in geographic locations and

pathogenic potentials (66). Y. enterocolitica also contains more than 70 serotypes, although

serotyping is not often performed in routine clinical laboratories, since the antisera are not

readily available.

Genotyping methods include pulsed-field gel electrophoresis (PFGE), which has long been

considered the gold standard for typing of Yersinia species. PFGE was found to be a more

useful tool than ribotyping for typing of pathogenic isolates of Y. enterocolitica (74) and has

been applied to epidemiological tracing of all three pathogenic species. A PFGE method for

typing of Y. pestis isolates is available through the CDC PulseNet website and has been used

in case investigations to identify the source of isolates (81).

SEROLOGIC TESTS Back to top

The gold standard for the diagnosis of plague is isolation of the organism; however, serology

can play a role, particularly when a culture is not recovered or for retrospective diagnosis or

epidemiologic studies in areas where plague is endemic. Most patients with plague

seroconvert 1 to 2 weeks following the onset of symptoms. The most commonly used antigen

in serologic assays for Y. pestis is the capsular F1 antigen. F1 antigen, which is highly

immunogenic and stable, is typically present in high concentrations in sera and bubo fluids of

plague patients even after several days of appropriate antimicrobial therapy. Serologic

diagnosis of plague can be made based only on a fourfold rise in antibody titers between

acute and convalescent serum samples. Serologic assay methods include passive

hemagglutination, the method recommended by the World Health Organization due to its low

cost and ease of performance. The reagents for passive hemagglutination are available only

in reference laboratories.

Serology can be used as an adjunct in the diagnosis of disease due to Y. enterocolitica or Y.

pseudotuberculosis. Antibody is detectable within the first week of illness and returns to

normal levels 3 to 6 months later. The specificity of serologic assays ranges from 82 to 95%

due to cross-reactivity between the two species and also

with Brucella, Francisella, and Vibrio species, as well as Borrelia burgdorferi, Chlamydia

pneumoniae, and some Escherichia coli serogroups. Another disadvantage of using serology

for diagnosis is that antibodies to Y. enterocolitica O antigens are often found in healthy

subjects due to the frequency of exposure to nonpathogenic serotypes. Most human

infections with Y. enterocolitica involve serotypes O:3, O:5, 27, O:8, and O:9. Serotype O:3

is the most common cause of gastroenteritis. However, as mentioned above, antisera are

available only to public health and research laboratories.

Antibody to outer membrane proteins (Yops) that are present only in virulent strains of Y.

enterocolitica may be more helpful. In a small study of healthy blood donors,

immunoglobulin M (IgM) antibody to Yops was 97% specific for acute infection (72). Testing

of blood donors for anti-Yop IgA in New Zealand, which has a high incidence of Y.

enterocolitica gastroenteritis, showed promise in preventing transfusion-related infections

(41). The presence of IgG and IgA antibodies to Y. enterocolitica Yops is also used as an aid

in the diagnosis of autoimmune disorders that occur postinfection, such as reactive arthritis,

erythema nodosum, Graves’ disease, and Hashimoto ’s thyroiditis (15). IgM-, IgA-, and IgGspecific

antibody reactivity against Yops in Western immunoblot formats has been correlated

with clinical presentation and sequelae (61).

ANTIMICROBIAL SUSCEPTIBILITIES Back to top

Pneumonic plague is nearly 100% fatal if not treated within the first 24 h of development of

symptoms. The drug of choice for the treatment of plague, pneumonic, septicemic, or

bubonic, is streptomycin. However, due to the lack of availability of streptomycin and

negative side effects, other agents have been evaluated in vitro and in animal models. The

only other antibiotic currently approved for treatment of plague is doxycycline; however,

other alternatives would be gentamicin and a fluoroquinolone such as ciprofloxacin or

levofloxacin (7, 36). Steward et al. documented the efficacy of fluoroquinolones in a mouse

model of systemic and pneumonic plague (70). The treatment of choice for plague meningitis

is chloramphenicol. Antibiotic resistance among isolates of Y. pestis has only rarely been

documented and never in the United States (26). Treatment failure due to antibiotic

resistance has never been documented. An isolate of Y. pestis from a 1995 plague case in

Madagascar was found to be multidrug resistant, with resistance to streptomycin,

sulfonamides, tetracycline, and chloramphenicol (31). Routine susceptibility testing of patient

isolates in this region of endemicity has failed to identify any further evidence of multidrug

resistance during the last 14 years. Antimicrobial susceptibility testing of Y. pestis is not

usually performed in clinical microbiology laboratories because of safety concerns in working

with this organism. The Clinical and Laboratory Standards Institute (http://www.clsi.org) has

published interpretative criteria and quality control limits for broth microdilution of Y.

pestis using Mueller-Hinton medium (18).

Most cases of Y. enterocolitica gastroenteritis do not require treatment; however, treatment

is necessary in cases of systemic disease, especially in immunosuppressed patients.

Treatment options include trimethoprim-sulfamethoxazole and a fluoroquinolone. Y.

enterocolitica produces two different β-lactamases, one of which is a class A constitutive

enzyme and the other of which is an inducible class C enzyme that is not inhibited by β-

lactamase inhibitors. The presence of one or both of these enzymes varies depending on the

biogroup (11). Although the β-lactamase confers resistance to penicillin on Y.

enterocolitica, the organism remains uniformly susceptible to the extended-spectrum

cephalosporins (60). Resistance to fluoroquinolones is due to either a mutation in

the gyrA gene or efflux mechanisms. In a study conducted in Spain, 23% of Y.

enterocolitica strains isolated from patients with gastroenteritis were nalidixic acid resistant.

All resistant isolates had a mutation ingyrA, and some were resistant based on an efflux

mechanism as well (11). Y. enterocolitica strains are susceptible in vitro to aminoglycosides,

chloramphenicol, tetracycline, trimethoprim-sulfamethoxazole, and extended-spectrum

cephalosporins.

Y. pseudotuberculosis is susceptible to ampicillin, tetracycline, chloramphenicol,

cephalosporins, and aminoglycosides. Although infections due to Y. pseudotuberculosis are

not usually treated, patients with septicemia should be treated with ampicillin, streptomycin,

or tetracycline. Y. aldovae and Y. ruckeri are also susceptible to penicillin. Y. frederiksenii, Y.

intermedia, and Y. rhodei produce a β-lactamase similar to that of Y. enterocolitica, which is

expressed at different levels in different strains (71).

EVALUATION, INTERPRETATION, AND REPORTING OF

RESULTS Back to top

Y. pestis, Y. enterocolitica, and Y. pseudotuberculosis are the primary pathogens in the

genus Yersinia. Isolation of Y. pestis from any body site warrants further investigation.

Isolation of Y. enterocolitica or Y. pseudotuberculosis from stool culture is not sufficient for

causal evidence of disease, since nonpathogenic serotypes may be normal stool biota.

However, no readily available methods except those using routine biochemicals, which are

not usually maintained in routine clinical laboratories, are available for differentiation of

pathogenic serotypes. Isolation of either species in pure culture from a symptomatic patient

with no other diagnosis should be considered suspect. Isolation of either species from blood

or other normally sterile sites should also be considered significant.

It has not been shown to be cost-effective to screen all stools for Y. enterocolitica by using

CIN agar. Isolation rates vary based on geographic locations, with the highest incidence in

temperate regions, so the decision to routinely rule out these organisms in stool cultures

should be evaluated in individual laboratories after consultation with the infectious disease

physicians.

Although the other Yersinia species besides Y. pestis, Y. enterocolitica, and Y.

pseudotuberculosis are not considered human pathogens, they have been isolated from the

gastrointestinal tracts of symptomatic patients with no other diagnosis. It has been

recommended that the presence of these Yersinia species in pure culture be reported. These

organisms may be underrecognized pathogens (24, 48).

Due to the lack of accuracy of commercial systems for the identification of Y. pestis, the

potential use of Y. pestis as a bioweapon, and the seriousness of the disease, all

suspected Y. pestis isolates should be sent to a local public health laboratory for

firmation.