Journal of Clinical Virology 39 (2007) 113–118

Hybridization assay performed at ambient temperature

for typing high-risk human papillomaviruses
Ivan Brukner a,∗ , Razan El-Ramahi a , Jacob Sawicki b ,
Izabella Gorska-Flipot b,d , Maja Krajinovic a,c , Damian Labuda a,c,∗∗
a Centre de Recherche, Hôpital Sainte-Justine, Université de Montréal, 3175 Côte Sainte-Catherine,
Montréal, Que. H3T 1C5, Canada
b Centre de Recherche, Hôpital Hôtel-Dieu, Université de Montréal, Montréal, Que., Canada
c Département de Pédiatrie, Université de Montréal, Montréal, Que., Canada
d Départment de Pathologie, Université de Montréal, Montréal, Que., Canada

Received 21 September 2006; received in revised form 21 March 2007; accepted 27 March 2007

Abstract

Background: Human papillomavirus (HPV) infection with oncogenic types is a prerequisite for cervical cancer development. HPV typing is
required in the management of pre-cancerous lesions, epidemiological studies, and vaccination trials. None of the available HPV assays are
satisfactory for routine diagnosis.
Objectives: In order to develop an assay for clinically relevant HPV types, we generated HPV probes using in vitro selection scheme of
iterative hybridization.
Study design: Starting from a mixture of random oligonucleotides, through several rounds of hybridization with 39 type-specific GP5+/6+
L1 sequences, we aimed to obtain specific probes to discriminate between these HPV types.
Results: In vitro selection led to pools of specific probes, from which individual probes were cloned and tested for their diagnostic performance
at ambient temperature. Typically, 10-fold stronger hybridization signals were obtained between each of the selected probes and their specific
targets compared to signals with the remaining 38 HPV types. High sensitivity and specificity of selected probes was demonstrated a series
of clinical samples in the hybridization assay.
Conclusions: A new panel of probes for detecting HPV types is described. Probes can be adapted for use in a simple clinical setting, or
incorporated into different detection systems.
© 2007 Elsevier B.V. All rights reserved.

Keywords: Ambient temperature hybridization; Oligonucleotide probes; HPV typing; Point-of-care device

1. Introduction
Infection with the human papillomavirus (HPV) is the
cause of cervical cancer (Munoz et al., 2003), the second most
common malignancy affecting women worldwide (Parkin et
∗ Corresponding author. Tel.: +1 514 345 4931x5647;
fax: +1 514 345 4731.
∗∗ Corresponding author at: Centre de Recherche, Hôpital Sainte-Justine,
3175 Côte Sainte-Catherine, Montréal, Que. H3T 1C5, Canada.
Tel.: +1 514 345 4931x5647; fax: +1 514 345 4731.
E-mail addresses: ibrukner@hotmail.com (I. Brukner),
damian.labuda@umontreal.ca (D. Labuda).
al., 2001). HPV consists of more than 100 types, many of
which are detected in the anogenital mucosa. They are divided
into high-risk types that are associated with anogenital cancer,
and low-risk types that are not carcinogenic and are primarily
found in genital warts (Munoz et al., 2003). HPV diagnosis,
including both detection and typing, is a recommended mea-
sure in the management of pre-cancerous lesions (ACOG,
2003). Typing is necessary to investigate the epidemiology
of particular HPV types, and for monitoring the efficacy
of the HPV vaccines. Several diagnostic kits are commer-
cially available and numerous diagnostic systems have been
described (see Cuschieri and Cubie, 2005; Klaassen et al.,

1386-6532/$ – see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.jcv.2007.03.013
114 I. Brukner et al. / Journal of Clinical Virology 39 (2007) 113–118
2004; Kleter et al., 1999; Lin et al., 2005; Molijn et al., 2005;
van den Brule et al., 2002; van Ham et al., 2005).
A recent study by the World Health Organization (Quint
et al., 2006) was undertaken to examine detection of 24
samples of the seven most frequent HPV types, using com-
mercially available individual typing kits, such as PGMY
line blot (Gravitt et al., 2000), SPF10-LiPa (Kleter et al.,
1999; van Hamont et al., 2006), Deg GP5+/6+ reverse line
blot (van den Brule et al., 2002) and DNA chip (Biomed
Lab Seoul, Korea). The measurements were performed in
29 independent laboratories and 12 different countries. The
overall detection rate of HPV16 was 62% and that of HPV18
was 73.9%; approximately half of the laboratories failed to
identify HPV31 and many did not detect types 35, 52 and
6. New, multiplex high throughput assays were developed,
using system of primers MY09/MY11 (Jiang et al., 2006),
PGMY09/11 (Wallace et al., 2005), or GP5+/GP6+ (Schmitt
et al., 2006), aiming at detection of 22 up to 45 distinct HPV
types. However, they either miss some clinically relevant
types (Jiang et al., 2006; Schmitt et al., 2006), or the vali-
dation of the full set of probes was not performed (Wallace
et al., 2005).
We describe an application of a novel empirical probe
design, the iterative hybridization method (Brukner et al.,
Fig. 1. Alignments of 39 HPV target sequences. The alignment of HPV sequences, between positions 6647 and 6740 as in HPV16 (GI: 333031), as obtained
by Clustal (http://www.ebi.ac.uk/cgi-bin/clustalw/).
2002, 2007), which yielded clinically useful probes for typ-
ing 39 mucosal HPV types. The procedure was optimized
to generate probes that differentiate between distinct short
HPV amplicons with as much as 87% sequence identity, in a
simple typing assay at ambient temperature.
2. Materials and methods
2.1. Oligonucleotides
Oligonucleotides were synthesized by Integrated DNA
Technologies (Coralville, IA). Targets were 91–100
nucleotides-long type-specific segments (GP5+ strands) of
the GP5+/6+ amplicons (Fig. 1), located between posi-
tions 6647 and 6740 of HPV16 sequence (accession number
K02718) (Seedorf et al., 1985). They were synthesized
non-modified or with biotin at their 5 ends. Biotiny-
lated targets were immobilized in streptavidin-coated tubes
(Roche Diagnostics GmbH, Mannheim, Germany) or 96-
well plates (Pierce Reacti-Bind Streptavidin Coated High
Binding Capacity Black plates, Rockford, Il) for preparative
and analytical purposes, respectively, following manufac-
turers’ instructions. Initial random oligonucleotides mix-
 I. Brukner et al. / Journal of Clinical Virology 39 (2007) 113–118
ture, ROM22, was GCCTGTTGTGAGCCTCCTGTCGAA-
(N)22 -TTGAGCGTTTATTCTTGTCTCCCA.
The 5 block TTCGACAGGAGGCTCACAACAGGC
and 3 block TGGGAGACAAGAATAAACGCTCAA were
used to “block” the anchoring segments of ROM22. The
forward primer GCCTGTTGTGAGCCTCCTGTCGAA and
the 5 -phosphorylated 3 block (used as reverse primer)
served to amplify ROM22 and the derived probes. The for-
ward GP5+ primer was as described by de Roda Husman et
al. (1995) and we used a mixture of four oligonucleotides as a
reverse GP6+ primer: GAAAAATAAACTGTAAATCATAT-
TC, GAAAAATAAACTGTAAATCATACTC, GAAAAA-
TAAACTGTAAATCAAATTC and GAAAAATAAACTG-
TAAATCAAACTC.
2.2. Iterative hybridizations
ROM22 (1 nmol) was used initially to obtain the first
affinity-selected mixture of oligonucleotide probes, referred
to as pooled probes, PP. All subsequent hybridization cycles
used PP from the previous cycle. Prior to use, PP were PCR-
amplified and converted to single stranded (GP5+ strand)
form as described (Brukner et al., 2007). ROM22 or PP was
mixed with block oligonucleotides (0.05–0.25 and 0.5 ␮M,
respectively) in 200 ␮l of TMN buffer. After heating to 90 ◦ C
the solution was transferred to tubes containing pre-bound
targets (1 pmol, except in the first step, when 100 pmol was
used), cooled down to 22–24 ◦ C and left for 4 h. After rins-
ing, oligonucleotides bound to the target were dissociated by
heating at 90 ◦ C for 2 min. Following positive hybridization
cycles described above, we carried subtractive hybridization
that differed only by the presence, in the hybridization solu-
tion, of 0.5 ␮M (total) of the remaining 38 non-biotinylated
target oligonucleotides (i.e. all other than the immobilized
Fig. 2. Hybridization of the selected pooled probes, PPs (A) and of the individual cloned probes, CPs (B) with each of the HPV type. PPs were obtained
after five rounds of positive and two rounds of subtractive hybridization, CPOs as described in the text. Gray scale expresses relative extent of hybridization
intensities.
115
target). Ultimately, the affinity-selected PPs were cloned into
plasmids, using TOPO TA Cloning kit (Invitrogen, CA), to
obtain individual cloned probes, CPs. The presence of insert
was confirmed by sequencing, using the LiCor apparatus
(Lincoln, NE). PCR-amplified PPs and CPs were validated
by hybridization. Hybridization with FAM-6 label ampli-
cons was quantified by measuring fluorescence in Spectra
MAX Gemini XS (22 ◦ C, λex = 485 nm and λem = 538 nm)
and expressed in relative fluorescence unit, RFU.
2.3. HPV typing assay
We used synthetic HPV16 GP5+/6+ oligonucleotide and
pre-characterized clinical samples as a source of HPV tem-
plates for PCR amplification (van den Brule et al., 2002).
The resulting amplicons were converted into single stranded
DNA, mixed with blocking oligonucleotides, and hybridized
to immobilized CPs (100 pmol). To detect hybridization sig-
nal, we used the FAM-6 labeled GP6+ primer serving here as
a detection probe: an exact complement of the GP5+ primer
and FAM-6 labeled GP6+ primer were added at 1 ␮M each.
After washing off the unbound material, the hybridization
was expressed in RFU.
3. Results
3.1. Selection, cloning and validation of specific probes
The probes were selected in a series of iterative
hybridization experiments (Brukner et al., 2007). Briefly,
biotinylated target oligonucleotides representing “GP5+/6+”
L1 sequences of 39 distinct HPV types (Fig. 1) were
immobilized in the streptavidin-coated tubes and hybridized
116 I. Brukner et al. / Journal of Clinical Virology 39 (2007) 113–118

to a mixture of random oligonucleotides (ROM22). Fol-

lowing hybridization, the unbound oligonucleotides were
washed away, whereas the bound ones were dissociated from
their target, re-amplified by PCR, and hybridized again.
Every additional hybridization cycle “iteratively” enriched
the resulting mixture of probes, PPs, in sequences that bind
efficiently to their target. The non-specific binders, which
would also recognize other HPV types, were eliminated dur-
ing cycles of subtractive hybridization. As shown in Fig. 2, all
39 PPs (5+2−), each obtained after five cycles of the positive
and two cycles of the subtractive hybridization, bound their
specific targets efficiently. Yet, a few of them (PPs 22, 23, 25,
27 and 29—see Fig. 2A) also cross-reacted with few other
HPVs. To eliminate the cross-reacting species, we added
another cycle of subtractive hybridization using only these
five targets. Thereafter, all the resulting type-specific PPs
(5+3−) efficiently discriminated each of their corresponding
39 HPV targets. On average we observed a 10-fold stronger
hybridization signal with the specific target as compared to
the remaining non-specific ones.
The unique probe sequences, cloned probes, CPs were
obtained by cloning PPs into plasmids. Individual cloned Fig. 4. HPV typing of pre-characterized clinical samples containing HPV6
probes were tested for binding to the whole panel of 39 and HPV16 to the array of 39 immobilized type specific CPs. (A) the arrange-
targets. In 29 out of 39 such experiments, the first tested ments of CP probes; (B) hybridization with HPV6; (C) hybridization wit
HPV16. Arrows indicate the orientation of the probes array.
clone displayed signal/noise ratio of at least 7. Cloned probes
obtained from the remaining PPs, for HPV types 6, 34, 40,
43, 45, 52, 64, 70, 72 and MM7, cross-hybridized with (Fig. 4). Clinical samples containing HPV6 and HPV16 types
up to four HPV types. Therefore, we performed an addi- were amplified by PCR using GP5+/6+ primers. The ampli-
tional subtractive hybridization to obtain PPs (5+4−) that cons, converted to single stranded form, were hybridized to
were subsequently cloned. Fig. 2B shows a summary of the the panel of immobilized probes in the presence of the FAM6-
hybridization results of 39 type-specific CPs, tested against labeled detection probe and blocking oligonucleotides. As
the whole panel of immobilized HPV targets. In turn, Fig. 3 shown in Fig. 4B and C significant hybridization signal was
shows hybridization intensities of all selected type-specific only detected with CP6 and CP16.
CPs with the immobilized HPV16, the most common onco-
genic HPV variant. The signal obtained with CP16 was ∼20 3.3. Origin of specificity of the selected probes
times stronger than with the remaining non-specific CPs.
Sequences of the selected CPs are shown in Fig. 5 and
3.2. HPV typing assay compared with the fragments of the HPV targets identified
by BLAST analysis. Clearly, there is no full complemen-
The HPV typing assay was performed in a reverse format, tarity between probes and targets. Complementary segments
in which all 39 HPV type-specific CPs, biotinylated at the are short and scattered, separated by mismatches. The speci-
5 terminus were immobilized in streptavidin-coated plates ficity of the probes appears thus to depend on different
factors. These include the position of the probe with respect
to the target sequence, its length and sequence context. A
better understanding can be obtained by analyzing each
probe–target combination. The alignment of the sequence
segment of CP33 and its HPV 33 target with nine additional
HPV types sharing high sequence identity with type 33 is
very revealing (Fig. 6). We note A/C mismatch at the sixth
position of the otherwise fully complementary CP33:HPV33
probe–target duplex (in gray). Interestingly, the same A is
also present in other nine HPV targets, all of which also
show mismatches in the adjacent sequence positions (2–5
nucleotides apart). One may speculate that “logic of probe
Fig. 3. Performance of 39 CPs with HPV16 target. Probes are in the same selection” was to introduce mismatch to further destabilize
linear order as HPV targets in Fig. 1. mismatches when interacting with non-specific targets.
 I. Brukner et al. / Journal of Clinical Virology 39 (2007) 113–118 117

Fig. 5. Partial sequence alignment of HPV targets with sequences of their specific CPs as obtained by BLAST. Sequences of HPV targets are in plain text (upper
row) and the corresponding CPs, devoid of flanking priming segments are in italics below. Only 22-nucleotide long HPV target segment is shown, corresponding
to the length of the CP sequence, in the region identified by BLAST (http://www.ncbi.nlm.nih.gov/blast/bl2seq/wblast2.cgi). Full matches between probe and
its target are highlighted in grey with local insertions/deletions indicated by dashes. The number following each HPV type corresponds to the distance from
the GP5+ primer 3 end to the first nucleotide position of the probe–target matching segment shown in gray. For HPV 35, where two BLAST alignments were
similarly scored and for HPV 52 and MM7, where BLAST was equivocal, the shown alignments were obtained using Clustal (Fig. 1).

4. Discussion pairs. The advantage of the method described here is the

We applied in vitro selection scheme to obtain a new
generation of probes that discriminate among 39 clinically
relevant HPV types, based upon the previously character-
ized GP5+/6+ L1 segment of the HPV (de Roda Husman
et al., 1995; Evans et al., 2005; Jacobs et al., 1997). In
a series of hybridization steps, starting from a mixture
of random oligonucleotides, we iteratively enriched it in
oligonucleotides that selectively recognized a specific HPV
out of the 39 HPV targets (Figs. 2 and 3). The performance
of the reverse format of the assay (CPs on the solid support)
was examined using clinical samples containing HPV 16 and
HPV 6 (Fig. 4).
In a classic multi-probe typing experiment, hybridization
conditions represent a compromise to accommodate quasi-
optimal conditions for each of the participating probe–target
possibility of enforcing and optimizing uniform conditions
for all probes that are to be used in the resulting diagnostic
assay. These conditions are those that are imposed/used dur-
ing iterative selections of different probes and correspond to
non-denaturing buffer hybridization at ambient temperature
between 20 and 28 ◦ C. Cross-hybridization was, on average,
less then 10% of the specific signal between the selected
probe and its corresponding HPV target, and sensitivity was
in the range of low pmol level.
Sequence analysis (Fig. 5) permits some speculations on
the “strategies” used by the selection process to achieve bind-
ing specificity of the probes. This specificity is achieved by
selecting short probe segments that are complementary to the
target and avoid matching with unintended targets. Similarly,
the complementary segments tend also to be interrupted by
mismatches, as illustrated by the CP33-HPV33 complex that
118 I. Brukner et al. / Journal of Clinical Virology 39 (2007) 113–118
Fig. 6. Partial sequence alignment of CP33 with its specific and nine similar
HPV targets. The mismatch that breaks an elongated stretch of complemen-
tarity between CP33 and its target is highlighted in gray. Dots represent
nucleotide identity with the uppermost CP33 and different sequences below.
Note that targets are in usual 5 –3 orientation, while upper CP33 is repre-
sented by its antisense strand to facilitate the comparison.
is characterized by the presence of a characteristic mismatch
(Fig. 6). This mismatch destabilizes probe–target complex,
but still allows for the complex to form. However, the same
mismatch prevents binding to very similar sequence motifs
present in the other 9 HPV targets to be discriminated (Fig. 6),
presumably by interrupting stacking interactions and destabi-
lizing even further its nearest-neighbor mismatched positions
(Brukner et al., 2007).
In conclusion, we describe a new panel of probes for
detecting different types of HPV. These probes can be used
in a simple clinical setting incorporating different detec-
tion systems. For example, combining the sensitivity of the
bead-based platform (Jiang et al., 2006; Schmitt et al., 2006;
Wallace et al., 2005) with a wide spectrum of viral types
detected by our probes could produce a robust HPV typ-
ing assay. Its miniaturization will be an important asset. Our
approach can be easily applied to generate diagnostic tools
for typing other pathogens or even in the detection of point
mutations in genes coding for known hereditary disorders.
Acknowledgements
We would like to thank Drs Irina Nazarenko (Digene
Corporation) and Kerry Kunning (IDT) for troubleshoot-
ing fluorescence quenching problem. This research was
supported by the Canadian Institutes of Health Research
(NTA-71859) and Research Center of Sainte-Justine Hos-
pital. MK is a scholar of the Fonds de la Recherche en Santé
du Québec.
References
ACOG. American College of Obstetricians and Gynecologists, News
release, July 31, 2003. Cervical Screening: Testing Can Start and Later
Occur Less Often Under New ACOG Recommendations.
Brukner I, Tremblay GA, Paquin B. Generation of amplifiable
genome-specific oligonucleotide probes and libraries. Biotechniques
2002;33:874–6, 878, 880 passim.
Brukner, I., El-Ramahi, R., Gorska-Flipot, I., Krajinovic, M., Labuda, D.,
2007. Generation of oligonucleotide probes to discriminate between
closely related DNA sequences. Nucleic Acids Res, 2007; April 10;
[epub ahead of print].
Cuschieri KS, Cubie HA. The role of human papillomavirus testing in cer-
vical screening. J Clin Virol 2005;32(Suppl 1):S34–42.
de Roda Husman AM, Walboomers JM, van den Brule AJ, Meijer CJ, Sni-
jders PJ. The use of general primers GP5 and GP6 elongated at their 3
ends with adjacent highly conserved sequences improves human papil-
lomavirus detection by PCR. J Gen Virol 1995;76(Pt 4):1057–62.
Evans MF, Adamson CS, Simmons-Arnold L, Cooper K. Touchdown
General Primer (GP5+/GP6+) PCR and optimized sample DNA concen-
tration support the sensitive detection of human papillomavirus. BMC
Clin Pathol 2005;5:10.
Gravitt PE, Peyton CL, Alessi TQ, Wheeler CM, Coutlee F, Hildesheim A,
et al. Improved amplification of genital human papillomaviruses. J Clin
Microbiol 2000;38:357–61.
Jacobs MV, Snijders PJ, van den Brule AJ, Helmerhorst TJ, Meijer CJ, Wal-
boomers JM. A general primer GP5+/GP6(+)-mediated PCR-enzyme
immunoassay method for rapid detection of 14 high-risk and 6 low-risk
human papillomavirus genotypes in cervical scrapings. J Clin Microbiol
1997;35:791–5.
Jiang HL, Zhu HH, Zhou LF, Chen F, Chen Z. Genotyping of human papil-
lomavirus in cervical lesions by L1 consensus PCR and the Luminex
xMAP system. J Med Microbiol 2006;55:715–20.
Klaassen CH, Prinsen CF, de Valk HA, Horrevorts AM, Jeunink MA, Thun-
nissen FB. DNA microarray format for detection and subtyping of human
papillomavirus. J Clin Microbiol 2004;42:2152–60.
Kleter B, van Doorn LJ, Schrauwen L, Molijn A, Sastrowijoto S, ter Schegget
J, et al. Development and clinical evaluation of a highly sensitive PCR-
reverse hybridization line probe assay for detection and identification
of anogenital human papillomavirus. J Clin Microbiol 1999;37:2508–
17.
Lin H, Moh JS, Ou YC, Shen SY, Tsai YM, ChangChien CC, et al. A
simple method for the detection and genotyping of high-risk human
papillomavirus using seminested polymerase chain reaction and reverse
hybridization. Gynecol Oncol 2005;96:84–91.
Molijn A, Kleter B, Quint W, van Doorn LJ. Molecular diagnosis of human
papillomavirus (HPV) infections. J Clin Virol 2005;32(Suppl 1):S43–51.
Munoz N, Bosch FX, de Sanjose S, Herrero R, Castellsague X, Shah KV, et al.
Epidemiologic classification of human papillomavirus types associated
with cervical cancer. N Engl J Med 2003;348:518–27.
Parkin DM, Bray FI, Devesa SS. Cancer burden in the year 2000. The global
picture. Eur J Cancer 2001;37(Suppl 8):S4–66.
Quint WG, Pagliusi SR, Lelie N, de Villiers EM, Wheeler CM. Results of the
first world health organization international collaborative study of detec-
tion of human papillomavirus DNA. J Clin Microbiol 2006;44:571–9.
Schmitt M, Bravo IG, Snijders PJ, Gissmann L, Pawlita M, Waterboer T.
Bead-based multiplex genotyping of human papillomaviruses. J Clin
Microbiol 2006;44:504–12.
Seedorf K, Krammer G, Durst M, Suhai S, Rowekamp WG. Human papil-
lomavirus type 16 DNA sequence. Virology 1985;145:181–5.
van den Brule AJ, Pol R, Fransen-Daalmeijer N, Schouls LM, Meijer CJ,
Snijders PJ. GP5+/6+ PCR followed by reverse line blot analysis enables
rapid and high-throughput identification of human papillomavirus geno-
types. J Clin Microbiol 2002;40:779–87.
van Ham MA, Bakkers JM, Harbers GK, Quint WG, Massuger LF, Melch-
ers WJ. Comparison of two commercial assays for detection of human
papillomavirus (HPV) in cervical scrape specimens: validation of the
Roche AMPLICOR HPV test as a means to screen for HPV genotypes
associated with a higher risk of cervical disorders. J Clin Microbiol
2005;43:2662–7.
van Hamont D, van Ham MA, Bakkers JM, Massuger LF, Melchers WJ.
Evaluation of the SPF10-INNO LiPA human papillomavirus (HPV)
genotyping test and the roche linear array HPV genotyping test. J Clin
Microbiol 2006;44:3122–9.
Wallace J, Woda BA, Pihan G. Facile, comprehensive, high-throughput geno-
typing of human genital papillomaviruses using spectrally addressable
liquid bead microarrays. J Mol Diagn 2005;7:72–80.