|Home | About | Journals | Submit | Contact Us | Français|
According to the Kauffmann-White scheme, 39 pairs of serovars in Salmonella serogroup C2 differ only by the minor antigen O:61. We characterized strains from 10 serovars representing five Salmonella serogroup C2 pairs. All strains demonstrated variable expression of O:61. These results indicate that these pairs are not distinct serovars.
Salmonella serovars are designated on the basis of the immunologic variability of two surface structures, O antigens and H antigens. O antigens are divided into two classes, major O antigens and minor O antigens. Major O antigens determine serogroup; minor O antigens are often variably expressed within a serovar and are not necessary for the determination of serogroup (4). For over 70 years, the Kauffmann-White scheme has served as the internationally recognized compendium of Salmonella serovars and has been integral to Salmonella surveillance. Designating serovars according to the conventions of the scheme ensures the historical continuity of data. The scheme also facilitates Salmonella surveillance and data comparison by providing an internationally accepted language for the designation of Salmonella serovars. However, the scheme was developed prior to the development of genotypic subtyping methods. In early attempts to provide epidemiologic information, the scheme relied heavily on minor antigenic variation (6). Once such example is the O:61 (also called O:6) antigen in serogroup C2. Salmonella serogroup C2 is defined by one major antigen (O:8) and two minor antigens O:61 and O:20 (5). There are currently 39 serovar pairs in the scheme which differ only by the presence or absence of O:6. Studies conducted in 1945 indicated that the O:6 antigen is commonly subjected to “form variation”; that is, pure cultures of a group C2 strain commonly contain mixtures of both O:6+ and O:6− colonies (3). Salmonella antiserum production guidelines draw attention to this phenomenon. Guidelines for the production of O:6 antiserum recommend testing multiple colonies and preparing immunizing antigen from colonies which react strongly with O:6 antiserum (7). Despite recognition of this variable expression, group C2 serovars continued to be designated on the presence or absence of O:6 (5).
Serotype- and pulsed-field gel electrophoresis (PFGE)-based surveillance via PulseNet has become the cornerstone of epidemiologic tracking in the United States (1). As the PulseNet USA database grew, it was noted that several isolates of Salmonella enterica serovars Istanbul and Hadar (antigenic formulas I 8:z10:e,n,x and I 6,8:z10:e,n,x, respectively) (5) uploaded to the PulseNet USA database shared indistinguishable XbaI PFGE patterns (CDC, unpublished data). To determine the extent of this phenomenon and its implications for Salmonella surveillance, we characterized the five most common Salmonella group C2 pairs reported in the United States: Salmonella serovars Hadar/Istanbul, Newport/Bardo, Blockley/Haardt, Muenchen/Virginia, and Litchfield/Pakistan.
Isolates were obtained from the Reference Collections of the Enteric Diseases Laboratory Branch (CDC). Serotyping was performed using previously described methods (2, 5). Single colonies from each strain were tested with CDC O:6 monoclonal antibody (CDC Division Of Scientific Resources, Biologics Branch). Eleven to 33 single colonies were tested for each isolate. PFGE of O:6+ and O:6− colony picks of serovars Newport, Litchfield, Hadar, Muenchen, and Blockley was performed at the CDC using the PulseNet standardized protocol for Salmonella with XbaI digestion (8). The resultant patterns were analyzed using BioNumerics version 3.5 (Applied Maths BVBA, Sint-Martens-Latem, Belgium). MLST data on previously characterized strains was obtained from publically available databases (http://mlst.ucc.ie/mlst/dbs/Senterica/).
A total of 396 single colonies were tested from 25 group C2 strains, representing five group C2 serovar pairs. Expression of O:6 varied by serovar pair (Table 1), but, with the exception of Salmonella Haardt, O:6+ and O:6− isolates were recovered from all strains tested. O:6 expression was highest among isolates of Blockley/Haardt (71% of single colonies) and lowest among isolates of Hadar/Istanbul (35% of single colonies). Across all five serovar pairs, 58% (230/396) of single colonies were O:6 positive.
O:6+ and O:6− colonies from the same strain yielded an indistinguishable XbaI PFGE pattern (Fig. 1). A query of the PulseNet USA database identified indistinguishable patterns for isolates of serovar Haardt (from Idaho) and Blockley (from Hawaii) as well as several indistinguishable patterns for a number of serovar Hadar/Istanbul and Newport/Bardo isolates. These findings supported our earlier observations of indistinguishable patterns among these pairs in the PulseNet database. A query of publically available MLST data identified an isolate of Bardo and several Newports with identical sequence types (STs) (Newport ST-46).
Our serotyping results support earlier studies indicating that the O:6 antigen is variably expressed in many of the commonly encountered O group C2 serotypes. When multiple colonies were tested from pure cultures, a mix of O:6+ and O:6− colonies was often encountered, clearly indicating that the designation of serovars on the basis of O:6 represents an inaccuracy in the Kauffmann-White scheme. For routine testing when only a single colony is tested, the distinction between group C2 pairs ultimately falls to random chance; this practice likely contributed to the early observation of variable expression of O:6 being overlooked for so long. Genetic methods are now the gold standard for Salmonella subtyping. Serotyping is the cornerstone of U.S. national Salmonella surveillance and serves as the primary subtyping method for surveillance networks such as PulseNet. Ensuring accurate data collection for these networks is crucial. Inaccuracies such as this associate two different serotype names with a single PFGE pattern and complicates Salmonella surveillance. Similarly, two serotype names associated with a single MLST may be misleading to users that do not understand the subtle difference between the serotype designations. Correcting these inaccuracies will help to ensure accurate data collection and improve Salmonella surveillance.
Published ahead of print 26 September 2012