Letters in Applied Microbiology 1997, 25, 54±57

Subspecies differentiation of Salmonella by PCR-

RFLP of the ribosomal operon using universal
primers

S .A . SHAH AND T. L. ROMICK . 1997.

S.A. Shah and T. L. Romick

Hunt-Wesson Inc., Technical Center, Fullerton, California, USA
1342/96: received 10 October 1996 and accepted 13 December 1996

A polymerase chain reaction (PCR)-based method was

developed to aid identi®cation of bacteria to subspecies level. The method used primers that
annealed to highly conserved regions of the bacterial rRNA operon, which are proposed
to be universal for all bacteria. The resulting PCR products gave unique electrophoretic
patterns due to restriction fragment length polymorphisms (RFLP) within the rRNA
operon, allowing differentiation to the subspecies level. Six serotypes of
Salmonella choleraesuis are presented to demonstrate the speci®city of PCR-
RFLP patterns for building an identi®cation database. As the database continues to
accumulate, the method proves to be speci®c and rapid for identifying bacteria
based on stable genetic characteristics.
INTRODUCTION
Salmonella species, as well as other foodborne bacterial patho-
gens, need to be identi®ed quickly and accurately. Traditional
methods for detecting and identifying this pathogen involve
cultural, biochemical and immunological assays that rely on
phenotypic characterization. These methods require selective
enrichment and plating, which often take several days to complete
and may alter phenotypic expression (Andrews et al. 1992).
Nucleic acid-based methods, such as the poly-merase chain
reaction (PCR), avoid such problems. PCR is a rapid and speci®c
method for detection of bacterial patho-gens. Additionally, PCR
sensitivity is such that theoretically only a single intact nucleic
acid template is needed to amplify the target sequence
suf®ciently to visualize by electrophoresis (Mullis et al. 1986 ;
Mullis 1990).
Many PCR methods have used primers that target viru-
lence genes that make them species-speci®c (Hill 1996). The
present authors have developed a PCR-based identi®cation
method that uses universal primers derived from highly con-
served regions in the rRNA operon (Fox and Woese 1974 ;
Gutell and Fox 1988 ; Jensen et al. 1993 ; Van Lith and Aarts
1994). Multiple rRNA operons are found in the bacterial
genome with hypervariable spacer regions that vary in length
and sequence but are considered stable genetic traits (Fox et
al. 1980 ; Barry et al. 1991). The method described here
differentiates to the subspecies level based on sequence poly-
Correspondence to: Dr Thomas Romick, Hunt-Wesson, Technical
Center, 1645 W. Valencia Drive, Fullerton, CA 92833, USA.
morphism in the 23S rRNA encoding region plus the two
¯anking hypervariable spacer regions. To date, PCR-RFLP
patterns have been generated for foodborne pathogens and
spoilage bacteria and a database is being built for aid in
identifying unknown bacterial strains. The method has
dem-onstrated an immediate use for matching unknown
strains of suspect origin to reference strains.
MATERIALS AND METHODS
Bacterial strains
Salmonella choleraesuis sero arizonae (ATCC 13314), Salm.
choleraesuis sero choleraesuis (ATCC 13312), Salm. choleraesuis
sero enteritidis (ATCC 13076), Salm. choleraesuis sero pul-lorum
(ATCC 19945), Salm. choleraesuis sero typhimurium (ATCC 14028)
and Salm. choleraesuis sero montevideo G4639 (M. Doyle, University
of Georgia, USA) were used.
DNA extraction
Pure cultures were grown overnight at 37°C in tryptic soy
broth (Difco, Detroit, MI). One millilitre of the bacterial
suspension was centrifuged for 5 min at 14 000 rev. min 1 in
an Eppendorf bench top centrifuge (type 5415C) to pellet
cells. The supernatant ¯uid was discarded and the cells were
resuspended in 400 ml LETS buffer (100 mmol l 1 LiCl, 10
mmol l 1 EDTA, 10 mmol l 1 Tris, pH 7´8, 1% SDS) plus 400
ml phenol : chloroform : isoamyl alcohol (50 : 48 : 2). Acid-
© 1997 The Society for Applied
Bacteriology
PCR -RFLP OF THE RIBOSOMAL OPERON
55
washed glass beads (106 mm diam. ; Sigma Chemical, St
Louis, MO) were added (approx. 0´1 volume) and then homo-
genized for 1 min at room temperature in a bead beater
(Biospec Products, Bartlesville, OK). The top (aqueous) phase
was transferred to a new tube and extracted with 400 ml of
chloroform : isoamyl alcohol (48 : 2). After phase separ-ation,
the aqueous phase was transferred to a new tube and DNA
precipitated by adding 1 ml 95% ethanol and chilled at 20°C
for 10 min. The DNA was pelleted by centrifugation (14 000
rev. min 1 for 10 min), the supernatant ¯uid aspi-rated, and the
pellet dried in a speed-vac (Savant, SC 110A, Farmingdale,
NY) for 10 min. The pellet was dissolved in 50 ml ultrapure
water (double glass distilled, 0´2 micron ®ltered, autoclaved
60 min, 121°C) and the DNA concentration was determined
spectrophotometrically (1 O.D. 50 ng/ml at 260 nm) and
checked for purity (260 nm/280 nm 1´8). The DNA template
was diluted to a working concentration of 40 ng/ml with
ultrapure water and used immediately for the reaction or
aliquoted and frozen ( 20°C).
PCR reaction conditions
A 14-mer oligonucleotide primer (5S) was synthesized
(Genosys Biotechnologies, The Woodlands, TX) based on a
conserved sequence mid-way in the 5S encoding region of the
rRNA operon (Fox and Woese 1974). The 5S primer sequence
used was 5?-AACTACCATCGGCG-3?. The other primer was
a 15-mer synthesized (Genosys) according to the G1
oligonucleotide design by Jensen et al. (1993). The G1 primer
was based on a conserved sequence in the 16S enco-ding
region of the rRNA operon speci®ed as 5?-
GAAGTCGTAACAAGG-3?. The total reaction mixture of 50
ml contained 10 mmol l 1 Tris (pH 8´3), 50 mmol l 1 KCl, 2
mmol l 1 MgCl2, 200 mmol l 1 of each deoxynucleotide
triphosphate (dATP, dCTP, dGTP, dTTP), 62´5 ng of each
oligonucleotide primer and 80 ng of bacterial genomic DNA.
The mixture was heated at 95°C for 5 min, cooled and held at
75°C until 1 U of DNA polymerase was added (Amplitaq
polymerase ; Perkin Elmer, Foster City, CA). Thirty-®ve
cycles were performed with an automated thermocycler (PTC-
100 ; MJ Research, Watertown, MA) as follows : 1 min
denaturation time at 94°C ; 2 min ramp to 55°C ; annealing
time of 2 min at 55°C ; 2 min ramp to 72°C ; 2 min poly-
merization time at 72°C with a ®nal polymerization time of 7
min at 72°C after the ®nal cycle. Tubes were held at 4°C until
PCR products were subjected to restriction analysis and gel
electrophoresis.
Restriction analysis
PCR products were digested by HinfI restriction endon-
uclease (Sigma) by premixing 0´5 ml of enzyme (5 U) with 5
´5 ml of 10X buffer (palette buffer blue ; Sigma) and adding
6 ml of mix directly to the 50 ml of completed PCR reaction.
This reaction was gently mixed by vortexing and incubated for
4 h at 30°C. After incubation, the entire 56-ml sample was
mixed with 3 ml of loading buffer (40% sucrose, 0´05%
bromphenol blue, 10 mmol l 1 EDTA, 0´5% SDS, pH 8´0) and
electrophoresed on a 3% TAE±agarose gel (40 mmol l 1 Tris
acetate, 1 mmol l 1 EDTA, pH 8´5), until complete separation
of the molecular weight marker (PCR Marker ; Sigma) was
achieved over the length of the gel. The gel was stained with
ethidium bromide (0´5 mg/ml) and photo-graphed under
ultraviolet transillumination. After scanning the photograph,
the patterns were analysed using image analysis software
(RFLP Proscan, Nashville, TN) and molec-ular weights
calculated.
RESULTS AND DISCUSSION
Many studies have been done to identify bacteria by PCR
(Hill 1996). Using universal primers that anneal to
conserved sequences of the rRNA operon is valuable if
results are reliable for all bacteria under identical reaction
conditions. Jensen et al. (1993) designed such primers for
the 16S±23S hypervariable spacer region. Apparently, the
length poly-morphism associated with the 16S±23S spacer
region is suf-®cient for generic characterization but
provides limited species and subspecies differential
information (Jensen et al. 1993 ; Van Lith and Aarts 1994).
The G1/5S primer set was designed to generate an expected
PCR product (approx. 3 kbp depending on variable lengths of
spacer regions in multiple operons) representing the 23S
region and both ¯anking spacer regions (Fox and Woese
1974 ; Dams et al. 1988 ; Gutell and Fox 1988). Dif-ferences
in the nucleotide sequence provided variation in HinfI
recognition sites and allowed the restriction analysis to reveal
sequence polymorphism in the 3 kbp PCR product. Figure 1
(top panel) shows the 3 kbp products for the six Salmonella
strains before digestion with HinfI restriction endonuclease.
Some smaller bands were produced occasion-ally at much less
intensity, perhaps indicating non-speci®c PCR products. The
purity of the DNA template preparation was critical for
reproducible results, especially when ampli-fying a relatively
large DNA fragment with expected sec-ondary structure. Since
sequence variation existed in the PCR product, especially in
the hypervariable spacer regions, HinfI digestion revealed
measurable differentiation among the strains (Fig. 1, bottom
panel). Calculated molecular weights for the patterns are
summarized in Table 1.
The PCR-RFLP patterns of the six Salmonella serotypes
had some similar features (Fig. 1, bottom panel). Restriction
fragments of approximately 670 and 470 bp were present in all
strains that were similar in size to PCR products reported
when the 16S±23S spacer region was ampli®ed (Jensen et al.
1993 ; Van Lith and Aarts 1994). This suggests that HinfI
© 1997 The Society for Applied Bacteriology,Letters in Applied
Microbiology 25, 54±57
56 S .A . SHAH AND T. L.
ROMICK
 Table 1 Calculated molecular weights (bp) of PCR-RFLP
fragments generated by HinfI digestion of 3 kbp PCR
products depicted in Fig. 1
б±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±± ±±
Lane 1 Lane 2 Lane 3 Lane 4 Lane 5 Lane 6 Lane 7 Lane 8
б±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±± ±±
2000 2000
1732
1500 1500
1217 1261
1107
1000 962 972 1000
944
874 883 883 874
787 802 802 802
750 750
661 659 670 669 681 679
572
500 513 513 513 500
472 472 471 471 478 465
362 363 365 371 364
318
300 287 300
272 276 273
234 235 243 233
225
194 203 197
191 192 187
165 170
157 158 157
150 149 150
115 121
72 72 72 71
68 68
Fig. 1 PCR products of approximately 3 kbp (not calculated) 50 50
б±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±±± ±±
generated by universal primers (top panel) and restriction fragments
generated by digestion of the PCR products by HinfI endonuclease Lanes 1 and 8, molecular weight marker; lane 2, Salmonella
(bottom panel). See Table 1 for lane descriptions and molecular choleraesuis sero montevideo G4639; lane 3, Salm. choleraesuis
weights. Top panel: molecular weight markers are the same as the top sero typhimurium ATCC 14028; lane 4, Salm. choleraesuis sero
four bands in the bottom panel enteritidis ATCC 13076; lane 5, Salm. choleraesuis sero
pullorum ATCC 19945; lane 6, Salm. choleraesuis sero
choleraesuis ATCC 13312; lane 7, Salm. choleraesuis sero
arizonae ATCC 13314.
restriction sites reside in or near the 16S±23S spacer region
and the length of the spacer regions are similar in closely
related species. Previous reports indicate that this is the used to identify bacteria such as Listeria monocytogenes and
case (Jensen et al. 1993 ; Van Lith and Aarts 1994). By Bacillus thuringiensis, based on virulence genes (Ericsson et
amplifying a larger section of the rRNA operon, RFLP al. 1995 ; Kuo and Chak 1996). Recently, Iriarte and Owen
analysis reveals more differential information than PCR (1996) performed PCR-RFLP analysis on a 2´65 kbp amplicon
product length poly-morphisms alone. representing the 23S rRNA genes of Campylobacter jejuni
The conventional method for identi®cation by restriction strains and concluded that not enough polymorphism existed
analysis has been Southern blot hybridization. This has the for subspecies differentiation. Including the hyp-ervariable
disadvantages that considerable time and resources are spacer regions in the PCR product with subsequent RFLP
required for preparation of the DNA hybridization mem-brane analysis may provide better differentiation comparable to the
and probes. Sensitivity is also affected by poor signal to noise results included in this report.
ratio (Vilgalys and Hester 1990). PCR-RFLP over-comes To date, this universal PCR-RFLP technique has been
these problems. The approach of PCR-RFLP has been applied to over 50 strains of foodborne pathogens and spoilage
© 1997 The Society for Applied Bacteriology,Letters in Applied
Microbiology 25, 54±57
PCR -RFLP OF THE RIBOSOMAL OPERON
57
bacteria and their unique patterns have been stored in a
growing, custom database. Differentiation of Salmonella ser-
otypes is presented here to demonstrate the potential of the
method for subspecies discrimination of other bacteria as well.
The method has been used to match unknown isolates to
reference strains for con®rming suspected species not readily
differentiated by biochemical and immunological tests. It also
has been valuable in tracing a contamination source in the
production environment. This could be used for epi-
demiological studies in the food chain or production environ-
ment so that rapid identi®cation would permit removal of the
source, thus preventing new cases (Couto et al. 1995 ;
Samadpour 1995).
REFERENCES
Andrews, W.H., Bruce, V.R., June, G., Satchell, F. and Sherrod, P.
(1992) Salmonella. In FDA Bacteriological Analytical Manual
ed. Tomlinson, L. pp. 51±69. Arlington, VA : AOAC Inter-
national.
Barry, T., Colleran, G., Glennon, M., Dunican, L.K. and Gannon,
F. (1991) The 16S/23S ribosomal spacer region as a target for
DNA probes to identify eubacteria. PCR Methods Applications
1, 51±56.
Couto, M.M.B., Vogels, J.T.W.E., Hofstra, H., Huis in't Veld, J.H.J.
and van der Vossen, J.M.B.M. (1995) Random ampli®ed
polymorphic DNA and restriction enzyme analysis of PCR ampli-
®ed rDNA in taxonomy : two identi®cation techniques for food-
borne yeasts. Journal of Applied Bacteriology 79, 525±536.
Dams, E., Hendriks, L., Van De Peer, Y., Neefs, J., Smits, G.,
Vandenbempt, I. et al. (1988) Compilation of small ribosomal
subunit RNA sequences. Nucleic Acids Research 16(Suppl),
87± 171.
Ericsson, H., Stalhandske, P., Danielsson-Tham, M.-L., Banner-
man, E., Bille, J., Jacquet, C. et al. (1995) Division of Listeria
monocytogenes serovar 4b strains into two groups by PCR and
restriction enzyme analysis. Applied and Environmental Micro-
biology 61, 3872±3874.
Fox, G. and Woese, C.R. (1974) The architecture of 5S rRNA and its
relation to function. Journal of Molecular Evolution 6, 61±76.
Fox, G.E., Stackebrandt, E., Hespell, R.B., Gibson, J., Maniloff,
J., Dyer, T.A. et al. (1980) The phylogeny of prokaryotes.
Science 209, 457±463.
Gutell, R. and Fox, G. (1988) A compilation of large subunit RNA
sequences presented in a structural format. Nucleic Acids
Research 16(Suppl), 175±201.
Hill, W. (1996) the polymerase chain reaction : application for the
detection of foodborne pathogens. Critical Reviews in Food
Science and Nutrition 36, 123±173.
Iriarte, P. and Owen, R.J. (1996) PCR-RFLP analysis of the large
subunit (23S) ribosomal RNA genes of Campylobacter jejuni.
Let-ters in Applied Microbiology 23, 163±166.
Jensen, M.A., Webster, J.A. and Straus, N. (1993) Rapid identi-
®cation of bacteria on the basis of polymerase chain reaction-
ampli®ed ribosomal DNA spacer polymorphisms. Applied and
Environmental Microbiology 59, 945±952.
Kuo, W.-S. and Chak, K.-F. (1996) Identi®cation of novel cry-type
genes from Bacillus thuringiensis strains on the basis of restriction
fragment length polymorphism of the PCR-ampli®ed DNA. Jour-
nal of Applied Microbiology 62, 1369±1377.
Mullis, K. (1990) The unusual origin of the polymerase chain
reac-tion. Scienti®c American 4, 56±65.
Mullis, K., Faloona, F., Scharf, S., Saiki, R., Horn, G. and Erlich,
H. (1986) Speci®c enzymatic ampli®cation of DNA in vitro :
the polymerase chain reaction. Cold Spring Harbor Symposia
on Quantitative Biology 51, 263±273.
Samadpour, M. (1995) Molecular epidemiology of Escherichia
coli 0157 : H7 by restriction fragment length polymorphism
using shiga-like toxin genes. Journal of Clinical Microbiology
33, 2150± 2154.
Van Lith, L.A.J.T. and Aarts, H.J.M. (1994) Polymerase chain
reaction identi®cation of Salmonella serotypes. Letters in
Applied Microbiology 19, 273±276.
Vilgalys, R. and Hester, M. (1990) Rapid genetic identi®cation and
mapping of enzymatically ampli®ed ribosomal DNA from several
Cryptococcus species. Journal of Bacteriology 172, 4238±4246.
© 1997 The Society for Applied Bacteriology,Letters in Applied
Microbiology 25, 54±57