Comparison of a Real-Time Multiplex PCR and Sequetyping Assay for Pneumococcal Serotyping
Comparison of a Real-Time Multiplex PCR and Sequetyping Assay for Pneumococcal Serotyping
Felix S. Dube 0 1
Suzan P. van Mens 0 1
Lourens Robberts 0 1
Nicole Wolter 0 1
Paul Nicol 0 1
Joseph Mafofo 0 1
Samantha Africa 0 1
Heather J. Zar 0 1
Mark P. Nicol 0 1
0 1 Division of Medical Microbiology, Faculty of Health Sciences, University of Cape Town , Cape Town , South Africa , 2 Department of Medical Microbiology and Immunology, St Antonius Hospital, Nieuwegein, the Netherlands, 3 National Health Laboratory Service, Groote Schuur Hospital , Cape Town , South Africa , 4 Department of Paediatrics and Child Health, Red Cross War Memorial Children's Hospital , Cape Town, South Africa, 5 MRC Unit on Child and Adolesscent Health , University of Cape Town , Cape Town , South Africa , 6 Centre for Respiratory Diseases and Meningitis (CRDM), National Institute for Communicable Diseases of the National Health Laboratory Service , Johannesburg , South Africa , 7 School of Pathology, Faculty of Health Sciences, University of the Witswatersrand , Johannesburg , South Africa , 8 The State Agricultural Biotechnology Centre, Murdoch University , Murdoch , Australia , 9 Centre for Proteomic and Genomic Research (CPGR) , Cape Town , South Africa
1 Editor: Shamala Devi Sekaran, University of Malaya , MALAYSIA
Funding: This work was funded in part by grants
from the Bill and Melinda Gates Foundation Global
Health Grant (OPP1017641) and H3Africa (1U01
HG006961-01). Felix S. Dube is supported by the
National Research Foundation of South Africa; S.P.
van Mens received funding from the St Antonius
Research Fund. The funders had no role in study
design, data collection and analysis, decision to
publish, or preparation of the manuscript.
Pneumococcal serotype identification is essential to monitor pneumococcal vaccine
effectiveness and serotype replacement. Serotyping by conventional serological methods are
costly, labour-intensive, and require significant technical expertise. We compared two
different molecular methods to serotype pneumococci isolated from the nasopharynx of South
African infants participating in a birth cohort study, the Drakenstein Child Health Study, in
an area with high 13-valent pneumococcal conjugate vaccine (PCV13) coverage.
A real-time multiplex PCR (rmPCR) assay detecting 21 different serotypes/-groups and a
sequetyping assay, based on the sequence of the wzh gene within the pneumococcal
capsular locus, were compared. Forty pneumococcal control isolates, with serotypes
determined by the Quellung reaction, were tested. In addition, 135 pneumococcal isolates
obtained from the nasopharynx of healthy children were tested by both serotyping assays
and confirmed by Quellung testing. Discordant results were further investigated by whole
genome sequencing of four isolates.
Of the 40 control isolates tested, 25 had a serotype covered by the rmPCR assay. These
were all correctly serotyped/-grouped. Sequetyping PCR failed in 7/40 (18%) isolates. For
Competing Interests: The authors have declared
that no competing interests exist.
the remaining isolates, sequetyping assigned the correct serotype/-group to 29/33 (88%)
control isolates. Of the 132/135 (98%) nasopharyngeal pneumococcal isolates that could
be typed, 69/132 (52%) and 112/132 (85%) were assigned the correct serotype/-group by
rmPCR and sequetyping respectively. The serotypes of 63/132 (48%) isolates were not
included in the rmPCR panel. All except three isolates (serotype 25A and 38) were
theoretically amplified and differentiated into the correct serotype/-group with some strains giving
ambigous results (serotype 13/20, 17F/33C, and 11A/D/1818F). Of the pneumococcal
serotypes detected in this study, 69/91 (76%) were not included in the current PCV13. The most
frequently identified serotypes were 11A, 13, 15B/15C, 16F and 10A.
The rmPCR assay performed well for the 21 serotypes/-groups included in the assay.
However, in our study setting, a large proportion of serotypes were not detected by rmPCR. The
sequetyping assay performed well, but did misassign specific serotypes. It may be useful for
regions where vaccine serotypes are less common, however confirmatory testing is advisable.
The pneumococcus (Streptococcus pneumoniae) is a common cause of invasive disease and
respiratory tract infections including bloodstream infections, meningitis, pneumonia and otitis
media [1–3]. Patients at risk include those at the extremes of age and the
immunocompromised, particularly those affected by cell-mediated immune deficiencies. Colonisation of the
nasopharynx with a homologous strain of pneumococci precedes the development of invasive
and respiratory tract disease [2,4,5]. Serotyping of the pneumococcal polysaccharide capsule,
the immunogenic component of current vaccines, remains the cornerstone of strain
characterization. To date, more than 90 capsular serotypes have been described and new ones continue
to be described [6,7]. Multiple pneumococcal serotypes can colonize the nasopharynx
successively over long period of time, or at any one time [8–10]. Invasive disease is commonly
regarded as resulting from a single serotype. Public health programs employ serotype
prevalence data from invasive disease to assist vaccine selection. Regular surveillance is required, and
relies mostly on phenotypic serotyping methods, most notably the Quellung method developed
in 1902 . The antiserum utilised in this assay is costly, methods employed are labour
intensive, and require significant technical expertise and experience.
More practical, higher throughput typing techniques are required for expanding public
health laboratory services in many areas of the world to support growing disease control
programs and epidemiological surveillance. Emerging technologies include alternative
culturebased phenotypic methods such as latex agglutination, dot blot ELISA and microbead assays
[10,12,13]. While the newer phenotypic methods all have their distinct benefits and often
surpass Quellung in terms of rapidity and cost, some of the methods require sophisticated and
Promising genotypic typing methods that target serotype-specific regions of the cps genes
have been developed including multiplex Polymerase Chain Reaction (PCR) with subsequent
agarose gel electrophoresis [14–16]; restriction fragment length polymorphism (PCR-RFLP)
; automated fluorescent capillary electrophoresis (FAF-mPCR) ; electrospray
ionization mass spectrometry (PCR/ESI-MS) ; reverse line blot hybridization assay (mPCR/RLB)
 and real-time multiplex PCR (rmPCR)  including the recently described nanofluidic
rmPCR . PCR with subsequent target detection is prone to amplicon contamination and is
more labour intensive than rmPCR. rmPCR obviates the need for amplicon manipulation, is
highly sensitive, fast and less labour intensive. PCR assays do not require viable isolates and
have the potential to detect multiple serotypes simultaneously [21,23–26]. More recently
sequetyping, a sequence-based typing method, has been described . There are currently no
published head-to-head comparisons of the accuracy of the sequetyping vs. multiplex PCR
approaches. Given the heterogeneity and recombinogenic nature of pneumococci, capsular
typing tools which infer type from DNA sequence, including target enrichment-based next
generation sequencing (NGS) and whole genome sequencing (WGS)  are attractive newer
methods to complement the molecular typing methods discussed above and may also aid in
resolving discrepant phenotypic and genotypic findings.
Materials and Methods
Isolates comprised 40 Quellung-typed control strains, Fig 1, (kindly donated by Dr. Anne von
Gottberg, Centre for Respiratory Diseases and Meningitis (CRDM), National Institute for
Communicable Diseases (NICD), South Africa ). These isolates were transported on
Dorset egg medium , subcultured onto Columbia blood agar base with 2% agar, 5% horse
blood and 4 μg/mL gentamicin media (CAG) upon receipt (Green point Media Laboratory of
the National Health Laboratory Service, Cape Town, South Africa) and incubated at 37°C in
5% CO2 overnight. The resulting colonies were inoculated into in 1 ml skim
milk-tryptone-glucose-glycerol (STGG) transport medium frozen at -80°C for batch processing.
Subsequently, 135 pneumococcal isolates (Fig 1) were cultured from nasopharyngeal (NP)
swabs that were collected from 83 healthy infants by employing nylon flocked swabs (Copan
Italia, Brescia, Italy). Infants were recruited between May 2012 and September 2013 as part the
Drakenstein Child Health Study (DCHS), a South African birth cohort study . NP swabs
were collected employing the World Health Organization protocol for pneumococcal carriage
studies. Briefly, the collected NP swabs were immediately placed into 1 ml STGG, transported
on ice to the laboratory and frozen at -80°C for batch processing. After thawing, STGG samples
were vortexed for 15 s before a 10 μl aliquot was inoculated onto Columbia blood agar base with
2% agar, 5% horse blood (BA) plates and incubated at 37°C in 5% CO2 overnight. Presumptive
pneumococcal isolates were identified by colony morphology, α-hemolysis and
ethylhydrocupreine (optochin) disk susceptibility (Oxoid, Basingstoke, UK) as previously described [32–34].
Nucleic acid extraction
Prior to rmPCR and sequetyping, all isolates were subjected to nucleic acid extraction
employing a heat lysis method as previously described . Briefly, a sweep of pneumococcal colonies
was obtained from primary BA plates that were inoculated with thawed STGG aliquots
containing either pneumococcal control strains or carriage isolates. The colony sweeps were
resuspended in 100 μl of phosphate-buffered saline, pH 7.4 (PBS; Sigma-Aldrich, St. Louis, MI)
thereafter heated at 95°C for 5 min. The supernatant containing genomic DNA (gDNA) was
ten-fold serially diluted in PBS before nucleic acid amplification.
Real-time multiplex PCR
The rmPCR, designed as a 7x3-plex that targets 21 serotypes. We used the multiplex scheme
for an African region, based on a relatively limited sample observed . This assay targets all
Fig 1. Flow chart showing the pneumococcal isolates included in the study. *Of the 40 isolates that were tested by rmPCR, only 25 were included as
part of the rmPCR targets.
the pneumococcal serotypes included in PCV13 and 8 other additional serotypes/-groups
(PCV13 serotypes: 1, 3, 4, 5, 6A/6B, 7F/7A, 9V/9A, 14, 18C/18A/18B/18C, 19A, 19F, 23F, 23A;
The 8 additional serotypes/-groups include: 2, 6C/6D, 11A/11D, 12B/2F/46, 15A/15F, 16F,
22F, 33A/33F). Briefly, the PCR reaction comprised 12.5 μl of 2X SensiFAST Probe No-ROX
One-Step master mix (Bioline, Taunton, MA), primers and probes for serotype as described by
Pimenta et al, 5 μl gDNA (diluted 1:1000) and nuclease/RNase-free water (Applied Biosystems,
Irving, CA) for a final reaction volume of 25 μl (S1 Table). The thermal cycling conditions
consisted of initial denaturation at 95°C for 10 min, followed by 40 cycles of denaturation at 95°C
for 15 s and annealing/extension at 60°C for 1 min employing a CFX96 Touch Real-Time PCR
amplification system (Bio-Rad Laboratories, Hercules, CA)
The assay was performed as previously described  with minor modification: the PCR
reaction comprised 12.5 μl of 2X KAPA Taq Ready Mix (KAPA Biosystems, Boston, MA), 1 μl of
primer mix, 2 μl gDNA (diluted 1:10), 8.5 μl nuclease/RNase-free H2O (Applied Biosystems) in
a final volume of 25 μl (S2 Table). Thermal cycling consisted of an initial denaturation at 95°C
for 5 min, followed by 30 cycles of denaturation at 95°C for 30 s, annealing at 65°C for 30 s,
and extension at 72°C for 90 s employing an Applied Biosystems 2720 Thermal Cycler
(Applied Biosystems). The PCR products were separated by electrophoresis in 1.5% agarose gel
(SeaKem LE Agarose; Lonza, Rockland, ME) for 45 min at 80 V in a 1X Tris-acetate EDTA
buffer. Ethidium bromide-stained DNA products were visualized under UV illumination and
sized by using a 1–kb DNA molecular size marker (HyperLadderv1-kb; Bioline).
PCR products were prepared for sequencing employing Exo-SAP IT (Affymetrix, Maumee,
OH) according to the manufacturer’s instructions. Prepared amplicons were submitted for
cycle sequencing employing the BigDye Sequence Terminator kit V3.1 (Applied Biosystems)
and analysed on an ABI 3500 XL Genetic Analyzer (Appplied Biosystems) by Inqaba Biotech
(Inqaba Biotechnical Industries [Pty] Ltd, Pretoria, South Africa). Sequencing was performed
in both directions using forward (cps1), 5’-GCA ATG CCA GAC AGT AAC CTC TAT-3’, and
reverse (cps2), 5’-CCT GCC TGC AAG TCT TGA TT-3’ primers.
DNA sequences obtained were assembled and edited using DNA Baser Sequence Assembler
v4 (www.DnaBaser.com). The consensus sequences were used to interrogate the GenBank
database (http://www.ncbi.nlm.nih.gov/blast/) and assign a serotype using the criteria as per
protocol . Briefly, the serotype of the wzh nucleotide sequence from GenBank with the highest
BLAST bit score was assigned, provided that sequence identity was >98% with the query
amplicon nucleotide sequence. To automate the above process, a Java-based program, Sequetyper
(available at http://www.gematics.com/sequetyper.html) was developed and validated to
automatically analyse and determine the pneumococcal serotype based on interrogation of GenBank
with the input forward and reverse sequences of the generated wzh amplicon. This application
is suitable for high-throughput analysis of sequetyping data (S1 Fig).
Pneumococcal control and nasopharyngeal isolates were submitted to CRDM for quellung
testing using specific anti-sera (Statens Serum Institut, Copenhagen, Denmark). Serotype 6C was
distinguished from serotype 6A by PCR , while serotype 25A and 38 were
undistinguishable and hence reported as 25A/38 .
Three pneumococcal carriage isolates serotyped as 16F by Quellung and rmPCR but identified
as 9V by sequetyping were subjected to WGS. The 3 discordant isolates as well as a control
strain identified as 9V by Quellung, sequetyping and rmPCR were also included. Briefly,
gDNA was isolated with a Wizard Genomic DNA Purification Kit (Promega Corporation,
Fitchburg, WI) according to the manufacturer's instructions. The gDNA quality was assessed
using the Qubit Fluorometer (Life Technologies, Carlsbad, CA), the NanoDrop ND-1000 (Life
Technologies) and agarose gel electrophoresis used to determine absolute concentration,
polyphenolic/polysaccharide/chaotropic salt contamination and gDNA integrity respectively.
Quantified gDNA was submitted to the Centre for Proteomics and Genomic Research (CPGR)
for WGS. Briefly, sequencing libraries were generated using the Nextera XT DNA Sample Prep
Kit (Illumina, San Diego, CA) and the libraries were indexed according to the dual-bar cording
protocol (with i7 and i5 primers) using the Nextera XT Index Kit (Illumina). Libraries were
then normalized, pooled, and a 5% PhiX control added before sequencing with the Illumina
MiSeq Reagent Kit v2 (500 cycle) on the Illumina MiSeq system.
De novo sequence assembly
The quality of the output sequence data was assessed using FastQC  and sequencing
adapters were trimmed using Trimmomatic . The 3'-end nucleotides with PHRED scores below
20 were trimmed using the fastx_trimmer tool of FASTX toolkit (http://hannonlab.cshl.edu/
fastx_toolkit) . The sequence data was then assembled de novo using SPAdes v3.0.0
assembler . Draft genome assemblies were annotated individually using RAST (Rapid
Annotation using Subsystem Technology) . The contigs containing putative cps regions were
identified through the standalone blastall homology searches against the 16F (Accession:
CR931668) and 9v (Accession: CR931648) annotated reference genomes and then extracted to
a separate file using a shell command based on SAMtools . These contigs were then aligned
and visual representation of the alignments was performed using the Artemis Comparison
Tool (ACT) v6 and WebACT .
Results of the two molecular serotyping assays up to the serogroup level were compared with
serotyping results obtained by Quellung testing. In cases of discordance between the two
molecular serotyping assays, the results were confirmed by Quellung testing. Serotype
distribution was determined based on Quellung results. Where more than one isolate was tested from
the same child, isolates of the same serotype were included only once in the analysis.
Ethical approval was obtained from the Human Research Ethics Committee of the Faculty of
Health Sciences, University of Cape Town (HREC ref: 062/2011) and the Western Cape
Provincial Child Health Research committee. Mothers provided written informed consent at
Of the 40 pneumococcal control isolates subjected to rmPCR, 25 isolates yielded a positive
signal; 15 isolates failed to yield detectable amplification signal. Of the 25 rmPCR positive isolates,
results were all (25/25) concordant with Quellung confirmed serotypes (Table 1).
Of the 40 pneumococcal control isolates subjected to sequetyping, 33 isolates yielded single
amplicons of ~ 1,061 bp; 7 isolates failed to yield detectable amplicons. Sequence analysis
yielded 29/33 (88%) sequetype-Quellung concordant results (Table 1). Of the four discordant
results, Quellung 16F was identified as 9V by sequetyping, Quellung 46 was identified as 12A
by sequetyping while Quellung 18C was identified as 18B. The fourth discordant isolate yielded
no match (>98%) when submitted to GenBank and is considered novel (GenBank submission
accession number: BankIt1792036). The wzh PCR was negative in 7/40 (18%) control isolates
tested, which consequently could not be sequetyped. Detailed results are provided in Table 1.
Of 135 pneumococcal isolates tested, 132 (98%) were assigned a serotype/-group by the
Quellung reaction (Tables 2 and 3). Three (3) isolates could not be typed by either Quellung or
molecular methods. A total of 69 (52%) isolates were assigned a serotype covered by the rmPCR
assay. Of these, the rmPCR assay assigned the correct serotype to all 69 isolates (Tables 2 and 3).
Of the 135 pneumococcal nasopharyngeal isolates that were sequetyped, 125 isolates yielded
single amplicons of ~ 1,061 bp. A correct serotype/-group was determined in 112 (85%) of the 132
nasopharyngeal isolates. The partial wzh sequence of 2/3 Quellung 18C isolates did not match any
2 rmPCR as 22A/22F; 2 sequetyped as 22A/22F
aThe numbers in closed brackets indicate the correct identification of a Quellung-confirmed serotype by the rmPCR and sequetyping assays;
brmPCR: real-time multiplex PCR;
* = serotypes not included in rmPCR assay.
¥ Mixed serotypes detected;
Negd = negative sequetyping PCR result;
cSerotype 25A and 38 were undistinguishable by Quellung and hence reported as 25A/38 .
of the pneumococcal wzh sequences in GenBank with >98% identity while the third was
determined as 18B. The wzh PCR was negative for seven isolates of which three were serotype 25A/38,
two were serotype 11A and two were serotype 19A, as confirmed by Quellung testing. Consistent
misidentifications by sequetyping, occurring in more than one isolate, were observed for serotype
16F (two isolates sequetyped as 9V) and for serotype 17F (two isolates sequetyped as 33C).
Fig 2 shows the serotype distribution of the pneumococcal isolates, excluding duplicate
isolates of the same serotype from the same infant. The most frequently identified serotypes were
11A (9 infants), 13 (8 infants), 15B, 15C (both 7 infants), 16F and 10A (both 6 infants). Of the
91 isolates (the total number of isolates when calculating each serotype only once per child), 22
(24%) were serotypes included in PCV13 while 69 (76%) serotypes were not.
Table 3. Summary of molecular serotyping results of pneumococcal nasopharyngeal isolates from healthy children compared with the serotype
determined by the Quellung reaction.
rmPCRa, correctly serotypedb
Sequetyping, correctly serotypedb
a rmPCR: real-time multiplex PCR.
b isolates typed correctly to the serogroup level compared with phenotypic Quellung reaction results.
c Negative: no amplification.
¥ all serotypes not covered by the rmPCR panel.
A total of 14.3 million paired-end sequence reads (2 x 250) were obtained for the four samples
as shown in Table 4. The quality control steps used preserved the sequence number though
reducing the sequence read length to 230 forward, and 120 reverse (230–120 fr) respectively.
Fig 2. Serotype distribution of nasopharyngeal pneumococcal isolates. The figure includes serotypes detected from the Drakenstein Child Health
Study, determined by Quellung reaction, excluding duplicate serotypes from the same infant. Blue = serotypes included in PCV13; Red = serotypes not
included in PCV13. Green = non-typable isolates.
Draft genome size (Mb)
One of pneumococcal control strains and two DCHS strains were serotyped as 16F in
Quellung, but mistyped as 9V by sequetyping. A comparison of the cps gene loci showed that the
wzh sequence of all the three queried 16F strains was entirely 9V-like (Fig 3). This is in contrast
to the rest of their cps loci: which in terms of structural gene organization as well as specific
sequence of these genes were entirely 16F-like. Comparative genome analysis of the annotated
gene structure showed a marked clustering for the other three queried 16F serotypes and were
all significantly different from the 9V reference (Fig 4). MLST loci of these 16F strains showed
Fig 3. Similarity of 16F-like capsular polysaccharide (cps) gene loci. Sequences from pneumococci serotyped as 16F Quellung but sequetyped as 9V
was compared to reference 9V (CR931648) and 16F (CR931668) cps sequences. Artemis Comparison Tool (ACT) was used to generate and view gene
homology. The top lines represent the forward and reverse strand of a serotype 9v reference, the middle lines represent the queried 16F strain and the
bottom lines shows the 16F reference. The portion of the wzh gene that is amplified by the sequetyping assay is shown by the blue rectangle. The clear
blocks below the blue box shows regions were the genes that are not similar. BLASTN matches are shown as red bands between sequences, indicating the
degree of similarity between the sequences.
Fig 4. Comparative genome analysis of pneumococcal serotypes 16F and 9V genetic background. When the sequence identities of all four genomes
were compared using RAST(Rapid Annotation using Subsystem Technology), the genome backbone of all three 16F (103347 and 103385 from this study
and a 16F control strain) were mostly identical but divergent from 9V. The colour codes represent how close or divergent the genomes are. Therefore, similar
genome backgrounds will have similar colours.
a shared a 16F-like-MLST-type, except the third strain (103385) which had a unique glutamate
dehydrogenase gene (gdh) allele.
To identify a rapid high throughput molecular serotyping assay, rmPCR was compared to
sequetyping, in the first place using a panel of 40 control isolates. rmPCR is designed to detect
and identify 21 serotypes including all serotypes/-groups in PCV13, all of which were included
in our analysis. Concordance with Quellung was 100% (25/25) for those control isolates
included in the rmPCR panel. Sequetyping is designed to identify up to 46 different
serotypes/groups, concordance with Quellung for the 40 control strains was 88% (29/33), with failure of
sequetyping PCR for 7 strains.
Amongst the pneumococcal carriage isolates tested, the correct serotype/-group could be
assigned to 52% (69/132) and 85% (112/132) by rmPCR and sequetyping respectively. Of these,
63/132 isolates were not included in the rmPCR panel. However, the sequetyping assay was
theoretically expected to amplify and differentiate all except three isolates (Serotypes 25A and 38)
into the correct serotype/-group with some strains giving ambiguous results (serotype 13/20,
17F/33C, and 11A/D/1818F). For those serotypes included in the rmPCR assay, there was good
agreement between the results across all three assays. The high number of negative results from
rmPCR amongst nasopharyngeal isolates was not surprising since this assay is likely to be less
useful in areas where pneumococcal conjugate vaccines have been implemented resulting in
serotype replacement which may arise as a result of either serotype unmasking or capsular switching.
Data from the United States on invasive pneumococcal isolates showed a decline in serotypes
included in the rmPCR assay from 92% (3812/4106) prior to PCV7 implementation to 79%
(2939/3708) after PCV7 roll out and a further decrease to 74% (2581/3480) post PCV13
implementation (Unpublished US Active Bacterial Core surveillance data). Amongst our small cohort,
non-vaccine serotypes 11A, 13, 15B, 15C, 16F and 10A were the most prevalent serotypes
identified. Similarly, data from a number of other post PCV13 surveillance studies have reported
serotypes 11A, 15A/B/C, 16, but also 22F, 21 and 34 as prevalent non-vaccine serotypes [45–51]. The
original rmPCR protocol  did not make reference to any internal control in the assay set up.
However, as part of our assay set-up and validation, we screened all the samples with a 16S rRNA
PCR to check for inhibition and subjected all rmPCR negative samples to cpsA PCR to check the
integrity of the capsulation locus, although not applicable for serotypes 14, 25, 35A and 38 .
The broad range of serotypes that are theoretically detectable by sequetyping is a major
advantage of this technique. It is not clear why, in this study, amplification failed for 7/40
(18%) control strains. Interrogation of published gene sequences for these serotypes indicated
that these serotypes should generate PCR products with the protocol used here . PCR
inhibition was excluded based on successful lytA PCR in all 7 strains. Interestingly, four of the
nasopharyngeal cariage isolates that failed to amplify during sequetyping were serotypes for
which a similar problem was encountered when sequetyping the control isolates (serotypes
11A and 19A). The remaining three sequetype-negative isolates were of serotype 25A/38 which
were expected to be non-amplifiable because of absence of the reverse primer binding site in
the wzd gene [6,27,53]. Sequetyping misidentified three control strains (for which Quellung
and rmPCR were concordant). The sequence obtained from wzh amplicon of serotype 2 did
not match any of the sequences in GenBank with >98% sequence identity. The original study
describing sequetyping did not test this serotype although their insilico analysis had predicted
that the primer sets should be able to amplify serotype 2 . Misidentification of the serotype
46 isolate as 12A is explained by high relatedness between these serotypes as their cps gene
clusters are almost identical . Based on our observation of mistyping the 18C PCV13 serotype as
18B by sequetyping, it may be warranted to confirm all 18B results by Quellung.
Misidentification of 17F isolates as 33C was predicted by the original sequetyping paper as these serotypes
cannot be distinguished based on their wzh sequence .
We found one control strain and three nasopharyngeal strains that were serotyped as 16F
by Quellung, but sequetyped as 9V. Even though the wzh sequences of the queried 16F was
entirely 9V-like, the serotype specific wzy/wzx genes are entirely 16F-like. Based on analysis of
the core genome, the 16F control strain was identified as sequence type (ST) 5326, one of the
nasopharyngeal isolates was identified as ST4088, while the other nasopharyngeal isolates was
a new ST, which was a single-locus variant of ST5326. These sequence types are all commonly
associated with serotype 16F . Therefore, our strains seem to be 16F strains in almost every
sense, they only have a 9V-wzh gene. This structural difference is not expected to have occurred
as a result of vaccination, because none of the currently used vaccine formulations include
serotype 16F and the exchanged 9V gene does not result in a modified phenotype. In practice,
in our setting, each isolate with a 9V sequetype result should be investigated further.
Both molecular assays are able to type many pneumococcal strains only to the serogroup
level. Discrimination of individual serotypes within a serogroup may be important for more
detailed assessment of carriage and vaccine effectiveness. When selecting a serotyping method,
test characteristics other than accuracy may also be relevant. The sequetyping assay, which
involves a single amplification step, is inexpensive compared with the rmPCR assay, which is
labour intensive, includes many costly PCR probes and is constrained by the limited
multiplexing options of real-time PCR. Interpretation of the sequetyping results is based on the
publically available GenBank database. An advantage of this is its free accessibility, but the
uncontrolled and changing nature of this database could be a risk for the assignment of
serotypes. Our automated ‘sequetyper’ application makes analysis of the relevant sequence data for
sequetyping rapid and simple. A significant disadvantage of sequetyping is that the targeted
wzh gene is not serotype-specific and does itself not determine serotype—the results are
inferred based on association. It is entirely feasible therefore (as we found here) that for specific
serotypes and in particular populations of pneumococci that this association may not correctly
predict serotype. The technique is therefore likely only useful for typing pneumococci from
populations of pneumococci where such association has been confirmed using another typing
technique. In our case this would mean confirming serotype for a smaller subset such as
serotype 9V, 13, 20 and serogroup 33. The CDC Streptococcal laboratory has recently provided an
update of a conventional multiplex PCR assay (not available at the time of this study) that
utilises 41 serotype-specific primer sets to detect upto 70 different pneumococcal serotypes
basis/methodology for this assay is similar to the rmPCR employed here although less costly. The benefits of
deducing more than 70 serotypes by this assay needs to be weighed against sensitivity and risk
of amplicon contamination.
In conclusion, sequetyping is a useful technique for large scale molecular serotyping of
pneumococcal strains, particularly post-PCV introduction, because of the broad range of
nonvaccine serotypes that can be detected, low cost and ease of use. Our results suggest the need
for an extended and carefully curated database of serotype-specific sequence data, which will
increase the accuracy and expand the serotype coverage of the sequetyping method. However,
given the potential for gene exchange that could result in false assignment of serotype by
sequetyping, it is necessary to confirm serotype assignment using a different method. This may still
be cost-saving as it would involve, for example, testing only the specific serotype assigned by
serotyping, using the Quellung method, in most instances. The rmPCR assay, ideally extended
to include more serotypes is reliable but cost, time required to perform testing, and currently
restricted serotype coverage may limit its widespread application for large epidemiological
studies. In the future it is likely that WGS will be increasingly used as a tool for serotype
inference. WGS has many advantages, in that additional information (such as multi-locus sequence
type and antimicrobial resistance) can be inferred from the same dataset without additional
testing, and that serotype can be definitively assigned. As sequence costs decline further,
bioinformatic pipelines are increasingly automated and the technology is more widely available in
low-resource settings it is likely that WGS will replace conventional typing tools for
We thank Anne von Gottberg, Linda de Gouveia, Mushal Ali and the staff of the Centre for
Respiratory Diseases and Meningitis (CRDM), National Institute for Communicable Diseases
(NICD) of the National Health Laboratory Service for training, sharing of standard operating
procedures, supplying control isolates and performing phenotypic serotyping. We thank the
clinical research staff involved in the Drakenstein Child Lung Health Study for collection of
samples and the children and parents for participating in the study. We further wish to thank
Mamadou Kaba, Charmaine Barthus, Widaad Zemanay, Layla Hendricks, Nchimunya
Hapeela, Whitney Barnett and the rest of the Drakenstein Child Lung Health Study team for their
help and technical assistance. We thank the Western Health Department and the staff at Paarl
hospital, Mbekweni and TC Newman clinics for their support of the study.
Conceived and designed the experiments: LR SvM NW JM HZ MN FSD. Performed the
experiments: FSD SvM JM. Analyzed the data: FSD SvM PN NW JM HZ MN. Contributed reagents/
materials/analysis tools: NW JM HZ MN PN FSD. Wrote the paper: FSD SvM PN NW JM LR
molecular serotyping by microarray. J Clin Microbiol 49: 1784–1789. doi: 10.1128/JCM.00157-11
1. Wang H , Liddell CA , Coates MM , Mooney MD , Levitz CE , Schumacher AE , et al. ( 2014 ) Global, regional, and national levels of neonatal, infant, and under-5 mortality during 1990-2013: a systematic analysis for the Global Burden of Disease Study 2013 . Lancet.
2. Bogaert D , De Groot R , Hermans PWM ( 2004 ) Streptococcus pneumoniae colonisation: the key to pneumococcal disease . Lancet Infect Dis 4 : 144 - 154 . PMID: 14998500
3. Rudan I , O'Brien KL , Nair H , Liu L , Theodoratou E , Qazi S , et al. ( 2013 ) Epidemiology and etiology of childhood pneumonia in 2010: estimates of incidence, severe morbidity, mortality, underlying risk factors and causative pathogens for 192 countries . J Glob Health 3 : 010401 . PMID: 23826505
4. Coles CL , Sherchand JB , Khatry SK , Katz J , Leclerq SC , Mullany LC , et al. ( 2009 ) Nasopharyngeal carriage of S. pneumoniae among young children in rural Nepal . Trop Med Int Health 14 : 1025 - 1033 . doi: 10.1111/j.1365- 3156 . 2009 .02331.x PMID: 19563428
5. Antonio M , Dada-Adegbola H , Biney E , Awine T , O'Callaghan J , Pfluger V , et al. ( 2008 ) Molecular epidemiology of pneumococci obtained from Gambian children aged 2-29 months with invasive pneumococcal disease during a trial of a 9-valent pneumococcal conjugate vaccine . BMC Infect Dis 8: 81. doi: 10.1186/1471-2334-8-81 PMID: 18547404
6. Bentley SD , Aanensen DM , Mavroidi A , Saunders D , Rabbinowitsch E , Collins M , et al. ( 2006 ) Genetic analysis of the capsular biosynthetic locus from all 90 pneumococcal serotypes . PLoS Genet 2 : 0262 - 0269 .
7. Calix JJ , Porambo RJ , Brady AM , Larson TR , Yother J , Abeygunwardana C , et al. ( 2012 ) Biochemical, Genetic, and Serological Characterization of Two Capsule Subtypes among Streptococcus pneumoniae Serotype 20 Strains: DISCOVERY OF A NEW PNEUMOCOCCAL SEROTYPE. J Biol Chem 287 : 27885 - 27894 . doi: 10.1074/jbc. M112.380451 PMID: 22736767
8. Brugger SD , Frey P , Aebi S , Hinds J , Mühlemann K ( 2010 ) Multiple colonization with S. pneumoniae before and after introduction of the seven-valent conjugated pneumococcal polysaccharide vaccine . PLoS One 5: e11638. doi: 10.1371/journal.pone.0011638 PMID: 20661289
9. Rivera-Olivero IA , Blommaart M , Bogaert D , Hermans PWM , de Waard JH ( 2009 ) Multiplex PCR reveals a high rate of nasopharyngeal pneumococcal 7-valent conjugate vaccine serotypes co-colonizing indigenous Warao children in Venezuela . J Med Microbiol 58 : 584 - 587 . doi: 10.1099/jmm.0. 006726 -0 PMID: 19369519
10. Turner P , Hinds J , Turner C , Jankhot A , Gould K , Bentley SD , et al. ( 2011 ) Improved detection of nasopharyngeal cocolonization by multiple pneumococcal serotypes by use of latex agglutination or
11. Austrian R ( 1960 ) The quellung reaction, a neglected microbiologic technique . Mt Sinai J Med 43 : 699 - 709 .
12. Roca A , Hill PC , Townend J , Egere U , Antonio M , Bojang A , et al. ( 2011 ) Effects of community-wide vaccination with PCV-7 on pneumococcal nasopharyngeal carriage in the Gambia: a cluster-randomized trial . PLoS Med 8: e1001107. doi: 10.1371/journal.pmed.1001107 PMID: 22028630
13. Bronsdon MA , O'Brien KL , Facklam RR , Whitney CG , Schwartz B , Schwartz B , et al. ( 2004 ) Immunoblot method to detect Streptococcus pneumoniae and identify multiple serotypes from nasopharyngeal secretions . J Clin Microbiol 42 : 1596 - 1600 . PMID: 15071010
14. Richter SS , Heilmann KP , Dohrn CL , Riahi F , Diekema DJ , Doern GV ( 2013 ) Evaluation of pneumococcal serotyping by multiplex PCR and quellung reactions . J Clin Microbiol 51 : 4193 - 4195 . doi: 10.1128/ JCM.01876-13 PMID: 24025905
15. Brito DA , Ramirez M , de Lencastre H ( 2003 ) Serotyping Streptococcus pneumoniae by multiplex PCR . J Clin Microbiol 41 : 2378 - 2384 . PMID: 12791852
16. Pai R , Gertz RE , Beall B ( 2006 ) Sequential multiplex PCR approach for determining capsular serotypes of Streptococcus pneumoniae isolates . J Clin Microbiol 44 : 124 - 131 . PMID: 16390959
17. Batt SL , Charalambous BM , McHugh TD , Martin S , Gillespie SH ( 2005 ) Novel PCR-restriction fragment length polymorphism method for determining serotypes or serogroups of Streptococcus pneumoniae isolates . J Clin Microbiol 43 : 2656 - 2661 . PMID: 15956380
18. Selva L , del Amo E , Brotons P , Muñoz-Almagro C ( 2012 ) Rapid and easy identification of capsular serotypes of Streptococcus pneumoniae by use of fragment analysis by automated fluorescencebased capillary electrophoresis . J Clin Microbiol 50 : 3451 - 3457 . doi: 10.1128/ JCM.01368-12 PMID: 22875895
19. Massire C , Gertz RE , Svoboda P , Levert K , Reed MS , Pohl J , et al. ( 2012 ) Concurrent serotyping and genotyping of pneumococci by use of PCR and electrospray ionization mass spectrometry . J Clin Microbiol 50 : 2018 - 2025 . doi: 10.1128/ JCM.06735-11 PMID: 22442313
20. O'Sullivan MVN , Zhou F , Sintchenko V , Kong F , Gilbert GL ( 2011 ) Multiplex PCR and reverse line blot hybridization assay (mPCR/RLB) . J Vis Exp.
21. Pimenta FC , Roundtree A , Soysal A , Bakir M , du Plessis M , Wolter N , et al. ( 2013 ) Sequential triplex real-time PCR assay for detecting 21 pneumococcal capsular serotypes that account for a high global disease burden . J Clin Microbiol 51 : 647 - 652 . doi: 10.1128/ JCM.02927-12 PMID: 23224094
22. Dhoubhadel BG , Yasunami M , Yoshida L-M , Thi HAN , Thi THV , Thi TAN , et al. ( 2014 ) A novel highthroughput method for molecular serotyping and serotype-specific quantification of Streptococcus pneumoniae using a nanofluidic real-time PCR system . J Med Microbiol 63 : 528 - 539 . doi: 10.1099/ jmm.0. 071464 -0 PMID: 24464695
23. Azzari C , Moriondo M , Indolfi G , Cortimiglia M , Canessa C , Becciolini L , et al. ( 2010 ) Realtime PCR is more sensitive than multiplex PCR for diagnosis and serotyping in children with culture negative pneumococcal invasive disease . PLoS One 5: e9282. doi: 10.1371/journal.pone.0009282 PMID: 20174571
24. Blaschke AJ ( 2011 ) Interpreting assays for the detection of Streptococcus pneumoniae . Clin Infect Dis 52 Suppl 4 : S331 - S337 . doi: 10.1093/cid/cir048 PMID: 21460292
25. Moore CE , Sengduangphachanh A , Thaojaikong T , Sirisouk J , Foster D , Phetsouvanh R , et al. ( 2010 ) Enhanced determination of Streptococcus pneumoniae serotypes associated with invasive disease in Laos by using a real-time polymerase chain reaction serotyping assay with cerebrospinal fluid . Am J Trop Med Hyg 83 : 451 - 457 . doi: 10.4269/ajtmh.2010. 10-0225 PMID: 20810803
26. Resti M , Moriondo M , Cortimiglia M , Indolfi G , Canessa C , Becciolini L , et al. ( 2010 ) Communityacquired bacteremic pneumococcal pneumonia in children: diagnosis and serotyping by real-time polymerase chain reaction using blood samples . Clin Infect Dis 51 : 1042 - 1049 . doi: 10.1086/656579 PMID: 20883110
27. Leung MH , Bryson K , Freystatter K , Pichon B , Edwards G , Charalambous BM , et al. ( 2012 ) Sequetyping: serotyping Streptococcus pneumoniae by a single PCR sequencing strategy . J Clin Microbiol 50 : 2419 - 2427 . doi: 10.1128/ JCM.06384-11 PMID: 22553238
28. Jauneikaite E , Tocheva AS , Jefferies JMC , Gladstone RA , Faust SN , Christodoulides M , et al. ( 2015 ) Current methods for capsular typing of Streptococcus pneumoniae . J Microbiol Methods 113 : 41 - 49 . doi: 10.1016/j.mimet. 2015 . 03.006 PMID: 25819558
29. Von Gottberg A , Cohen C , de Gouveia L , Meiring S , Quan V , Whitelaw A , et al. ( 2013 ) Epidemiology of invasive pneumococcal disease in the pre-conjugate vaccine era: South Africa , 2003 - 2008 .
30. Wasas AD , Huebner RE , De Blanche M , Klugman KP ( 1998 ) Long-term survival of Streptococcus pneumoniae at room temperature on Dorset egg medium . J Clin Microbiol 36 : 1139 - 1140 . PMID: 9542956
31. Le Roux DM , Myer L , Nicol MP , Zar HJ ( 2015 ) Incidence and severity of childhood pneumonia in the first year of life in a South African birth cohort: the Drakenstein Child Health Study . Lancet Glob Heal 3 : e95 - e103 .
32. Satzke C , Turner P , Virolainen-Julkunen A , Adrian PV , Antonio M , Hare KM , et al. ( 2013 ) Standard method for detecting upper respiratory carriage of Streptococcus pneumoniae: updated recommendations from the World Health Organization Pneumococcal Carriage Working Group . Vaccine 32 : 165 - 179 . doi: 10.1016/j.vaccine. 2013 . 08.062 PMID: 24331112
33. Dube FS , Kaba M , Whittaker E , Zar HJ , Nicol MP ( 2013 ) Detection of Streptococcus pneumoniae from Different Types of Nasopharyngeal Swabs in Children. PLoS One 8: e68097 . PMID: 23840817
34. Winn WC , Allen SD , Janda WM , Koneman EW , Procop GW , Schreckenberger PC , et al. ( 2006 ) GramPositive Cocci Part II: Streptococci, Enterococci, and the “Streptococcus-like ” Bacteria , p. 672 - 764 . In: Konemans's Color Atlas and Textbook of Diagnostic Microbiology . 6th ed . Wilkins LW and, editor 672 - 764 p.
35. Leung MHY , Oriyo NM , Gillespie SH , Charalambous BM ( 2011 ) The adaptive potential during nasopharyngeal colonisation of Streptococcus pneumoniae . Infect Genet Evol 11 : 1989 - 1995 . doi: 10.1016/j. meegid. 2011 . 09.002 PMID: 21925618
36. Park IH , Pritchard DG , Cartee R , Brandao A , Brandileone MCC , Nahm MH , et al. ( 2007 ) Discovery of a new capsular serotype (6C) within serogroup 6 of Streptococcus pneumoniae . J Clin Microbiol 45 : 1225 - 1233 . PMID: 17267625
37. Nunes MC , Jones SA , Groome MJ , Kuwanda L , Van Niekerk N , von Gottberg A , et al. ( 2014 ) Acquisition of Streptococcus pneumoniae in South African children vaccinated with 7-valent pneumococcal conjugate vaccine at 6, 14 and 40 weeks of age . Vaccine.
38. Andrews S ( 2011 ) FastQC a quality-control tool for high-throughput sequence data .
39. Bolger AM , Lohse M , Usadel B ( 2014 ) Trimmomatic: a flexible trimmer for Illumina sequence data . Bioinformatics 30 : 2114 - 2120 . doi: 10.1093/bioinformatics/btu170 PMID: 24695404
40. Patel RK , Jain M ( 2012 ) NGS QC Toolkit: a toolkit for quality control of next generation sequencing data . PLoS One 7: e30619. doi: 10.1371/journal.pone.0030619 PMID: 22312429
41. Nurk S , Bankevich A , Antipov D , Gurevich AA , Korobeynikov A , Lapidus A , et al. ( 2013 ) Assembling single-cell genomes and mini-metagenomes from chimeric MDA products . J Comput Biol 20 : 714 - 737 . doi: 10.1089/cmb.2013.0084 PMID: 24093227
42. Aziz RK , Bartels D , Best AA , DeJongh M , Disz T , Edwards RA , et al. ( 2008 ) The RAST Server: rapid annotations using subsystems technology . BMC Genomics 9: 75. doi: 10.1186/1471-2164-9-75 PMID: 18261238
43. Li H , Handsaker B , Wysoker A , Fennell T , Ruan J , Homer N , et al. ( 2009 ) The Sequence Alignment/ Map format and SAMtools . Bioinformatics 25 : 2078 - 2079 . doi: 10.1093/bioinformatics/btp352 PMID: 19505943
44. Jolley KA , Maiden MCJ ( 2010 ) BIGSdb: Scalable analysis of bacterial genome variation at the population level . BMC Bioinformatics 11: 595. doi: 10.1186/1471-2105-11-595 PMID: 21143983
45. Guevara M , Ezpeleta C , Gil-Setas A , Torroba L , Beristain X , Aguinaga A , et al. ( 2014 ) Reduced incidence of invasive pneumococcal disease after introduction of the 13-valent conjugate vaccine in Navarre , Spain, 2001 - 2013 . Vaccine 32 : 2553 - 2562 . doi: 10.1016/j.vaccine. 2014 . 03.054 PMID: 24674661
46. Olarte L , Hulten KG , Lamberth L , Mason EO , Kaplan SL ( 2014 ) Impact of the 13-Valent Pneumococcal Conjugate Vaccine on Chronic Sinusitis Associated with Streptococcus pneumoniae in Children . Pediatr Infect Dis J.
47. Kaplan SL , Barson WJ , Lin PL , Romero JR , Bradley JS , Tan TQ , et al. ( 2013 ) Early trends for invasive pneumococcal infections in children after the introduction of the 13-valent pneumococcal conjugate vaccine . Pediatr Infect Dis J 32 : 203 - 207 . doi: 10.1097/INF.0b013e318275614b PMID: 23558320
48. Steens A , Bergsaker MAR , Aaberge IS , Rønning K , Vestrheim DF ( 2013 ) Prompt effect of replacing the 7-valent pneumococcal conjugate vaccine with the 13-valent vaccine on the epidemiology of invasive pneumococcal disease in Norway . Vaccine 31 : 6232 - 6238 . doi: 10.1016/j.vaccine. 2013 . 10.032 PMID: 24176490
49. Van Hoek AJ , Sheppard CL , Andrews NJ , Waight PA , Slack MPE , Harrison TG , et al. ( 2014 ) Pneumococcal carriage in children and adults two years after introduction of the thirteen valent pneumococcal conjugate vaccine in England . Vaccine.
50. Aguiar S , Brito M , Horacio A , Lopes J , Ramirez M , Melo-Cristino J , et al. ( 2014 ) Decreasing incidence and changes in serotype distribution of invasive pneumococcal disease in persons aged under 18 years since introduction of 10-valent and 13-valent conjugate vaccines in Portugal, July 2008 to June 2012 . Euro Surveill 19.
51. Cohen R , Levy C , Bingen E , Bechet S , Derkx V , Werner A , et al. ( 2012 ) Nasopharyngeal carriage of children 6 to 60 months during the implementation of the 13-valent pneumococcal conjugate vaccine . Arch Pediatr 19 : 1132 - 1139 . doi: 10.1016/j.arcped. 2012 . 07.013 PMID: 22925540
52. Da Gloria Carvalho M , Pimenta FC , Jackson D , Roundtree A , Ahmad Y , et al. ( 2010 ) Revisiting pneumococcal carriage by use of broth enrichment and PCR techniques for enhanced detection of carriage and serotypes . J Clin Microbiol 48 : 1611 - 1618 . doi: 10.1128/ JCM.02243-09 PMID: 20220175
53. Mavroidi A , Aanensen DM , Godoy D , Skovsted IC , Kaltoft MS , et al. ( 2007 ) Genetic relatedness of the Streptococcus pneumoniae capsular biosynthetic loci . J Bacteriol 189 : 7841 - 7855 . PMID: 17766424