Adenovirus Polymerase Chain Reaction Assay for RapidDiagnosis of Conjunctivitis
RobertJ. Cooper,1 Adrian C. Yeo,x Andrew S. Bailey,2 and Andrew B. Tullo5
PURPOSE. TO evaluate newly designed primers in a polymerase chain reaction (PCR) for thedetection of adenovirus DNA in conjunctiva! swabs.
METHODS. Oligonucleotides were derived from the adenovirus liexon gene and modified such that:a maximum of only two mismatches occurred with adenovirus types 2 through 5. 7, and 16.Specificity was determined against adenovirus types 2 through 4, 7, 8 through II, 14, 19, 37, 40,and 4 1 and from non-adenoviral DNA and the sensitivity by PCR amplification of purified adeno-virus type 2 DNA. The assay was compared retrospectively with cell culture and a PCR withdifferent primers on 59 stored conjunctiva! swab samples. The new PCR also was used prospec-tive!}' in comparison with cell culture on 2743 conjunctiva! swabs.
RESULTS. The l4()-bp product was amplified from all the adenovirus serotypes tested except types40 and 41, which have not been isolated from the eye. There were no amplified products from thenon-adenoviral DNA tested. With adenovirus type 2 DNA. despite two deliberate mismatches, 40copies of the target were detectable after PCR and ethidium bromide-staining. In the retrospectivestudy, 51 of 55 (92.7%) were positive by this new PCR compared with 42 of 55 (76.4%) by the olderPCR and 40 of 55 (72.7%) by cell culture. In the prospective study, the new PCR detected 386 of415 (93%) adenovirus-positive specimens compared with 248 of 415 (59.8%) by cell culture. Of 167specimens positive for herpes simplex virus by cell culture, none were positive by the adenovirusPCR.
CONCLUSIONS. PCR with the newly designed primers shows a much increased sensitivity over cellculture and previous PCRs for the detection of adenoviruses in conjunctiva! swabs. (InvestOphthalmol Vis Sci. 1999:40:90-95)
A broad range of adenoviruses (principally subgenus 1$,types 3 and 7; subgenus C, types 1, 2 and 5; subgenusD, type 10; and subgenus E, type 4) cause conjunctivitis
and arc also linked with community outbreaks of follicularconjunctivitis or pharyngoconjunctiva! fever.1 In addition, ep-idemics of subgenus D (particularly types 8, 19, and 37) ade-novirus keratoconjunctivitis can occur in eye clinics.2* Theneed for an accurate and rapid laboratory identification ofadenoviral conjunctivitis is important because the clinical di-agnosis of adenovirus eye disease can be difficult in the ab-sence of characteristic features.' Also, the control of nosoco-mial outbreaks of epidemic keratoconjunctivitis can befacilitated by rapid laboratory investigation3 to promote rein-forcement of hand washing and tonometer disinfection proce-d u r e s / 6 Studies on the antiviral chemotherapy of adenovirusconjunctivitis with (S)-1 -(3-hydroxy-2-phosphon\iniethoxy-
From the 'Division of Virology, Department of Pathologic Sci-ences, University of Manchester; the ^Clinical Virology Laboratory,Manchester Royal Infirmary: and the *Royal live Hospital, Manchester.England.
Submitted for publication April 15. 1998: revised July 21, 1998;accepted August 14. 1998.
Proprietary interest category: N.Present address: Adrian C. Yeo, Chemical Process and Biotechnol-
Reprint requests: Robert J. Cooper, University Virology, Depart-ment of Pathologic Sciences. University of Manchester, 3rd Floor,Clinical Sciences Building, Manchester Royal Infirmary. Oxford Road.Manchester M 13 9WL England.
propyl)-cytosine (Cidofovir)' may result in effective treatmentthat may require laboratory confirmation to be justified.
Until recently, detection of adenovirus in conjunctiva!swabs relied on virus isolation in cell culture or antigen detec-tion,K'9 methods that are either slow or insensitive, respec-tively. Polymerase chain reaction (PCR) assays provide diagnos-tic tests for adenoviral eye infections that are more rapid thanculture and more sensitive than antigen detection.10"12 Ade-novirus PCR avoids the low specificity of viral antigen detec-tion1* but can lack sensitivity for some serotypes found in eyeswabs. 1O'I~ We describe a new primer pair for an adenovirusPCR assay and its application to large numbers of eye swabssent to the diagnostic virology laboratory.
MATERIALS AND METHODS
Clinical Specimens
For the retrospective studies, extracts of 59 conjunctival swabsthat had previously been tested for adenovirus by culture andby PCR using II1 and 112 primers'' and stored at — 30°C for upto 18 months were analyzed by the new PCR. This was donewithout prior knowledge of the original diagnostic laboratoryresults. Subsequently, for the prospective study, conjunctivalswabs arriving in the laboratory in viral transport medium wereimmediately inoculated onto cell cultures for detection of ad-enoviruses or herpes simplex virus (1ISV) and for adenovirustyp ing . 8 " 1 2 These samples were examined within 1 week ofreceipt using the new adenovirus PCR. All the conjunctival
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933210/ on 02/04/2018
10VS, January 1999, Vol. 40, No. 1 PCR for Adenoviral Conjunctivitis 91
swabs were from patients presenting with conjunctivitis andshowing symptoms and signs consistent with a I-ISV or adeno-virus etiology.
Sample PreparationDNA was extracted from the viral transport medium with lysisbuffer." Briefly, equal volumes (75 JU.1) of sample and lysisbuffer (20 mM Tris-hydrochloric acid |pH 8.31, 2 mM EDTA, 1%Triton X-100, 0.002% sodium dodecyl sulfate, and 500 Wj/mlproteinase K) were incubated at 56°C for 2 hours and thenboiled for 10 minutes. For a small number of specimens, analternative procedure was used in which 100 JLL! of the viraltransport medium was mixed with 50 /u.1 30% polyethyleneglycol (PEG-9000; Sigma Chemical, St. Louis, MO) in 3 Msodium chloride and kept on ice for 30 minutes. After centrif-ugation (lO.OOOg, 15 minutes), the supernatant was removed,and the pellets were resuspended in 20 fx\ of 10 mM Tris-hydrochloric acid (pH 7.6), 05% Nonidet P-40 and incubated atroom temperature for 10 minutes before adding 5 jul directly tothe PCR mixture.
PCR ConditionsThe reaction mixture for the PCR was composed of 10 mMTris-hydrochloric acid (pll 8.3). 50 mM potassium chloride.200 /xM each deoxynucleoside triphosphate, 1.5 mM MgCl2,0.2 /AM primers, 1.25 U Taq DNA polvmerase (Advanced Bio-technologies Ltd, Epsom, UK), and 5 jxl appropriate DNAsample or sterile distilled water, as contamination control, toform a final volume of 50 /xl.'' One drop of mineral oil (Sigma)overlay was added to each reaction mixture to prevent evap-oration. The assay was performed on a Programmable Dri-Block Gene Ataq thermal cycler (Pharmacia LK.B, Uppsala,Sweden) using one cycle each of94°C for 7 minutes, 55°C for1 minute, and 72°C for 1.5 minute followed by 40 cycles eachof 94°C for I minute, 55°C for I minute, and 72°C for 1.5minute. The amplification products were analyzed by electro-phoresis in 6% polyacrylamide gels,12 and the anticontamina-tion measures in all the PCRs were as previously described.1 l l ' i
They included separate rooms for preparation of the reactionmixtures, for preparation and addition of DNA extracts, and forproduct analysis. Plugged pipette tips were used throughout.
Control DNA SamplesThe control DNA samples were composed of adenovirus type2 genomic DNA (GIBCO-BRL, Paisley, Scotland, UK) and DNAextracted from adenovirus strains derived from clinical iso-lates"; DNA extracted from HSV type 1 strain Synl7 ' infectedVero cells; purified plasmid designated pCtl.2 from Chlamycliatrachomalis strain L2/43415; and human fibroblast DNA.16
Statistical AnalysisThe results of the various assays of clinical specimens werecompared using McNemar's test:.1' P < 0.05 was consideredsignificant.
RESULTS
Primer DesignThe two adenovirus primer pairs (HI and H2,IH JCH1 andJCH2iy) evaluated on eye swabs in this laboratory1 l l 2 had
nucleotide sequences identical to those of adenovirus types 2and 52 0 but showed, respectively, 4, 5, 3. and 2 mismatcheswith adenovirus type 3 (Fig. 1). This included a mismatch atthe 3' terminus of primer H2, a site crucial in the binding ofTaq polvmerase. Thus, the apparently lower sensitivity of PCRsusing primers 111 and 112 rather than JCH1 and JCH2 fordetection of subgenus B adenoviruses,12 and the suboptimalperformance of the former primers with the same strains evenafter lowering of the annealing temperature to 40°C" sug-gested that further improvement in the diagnostic perfor-mance of adenovirus DNA amplification required the design ofa novel oligonucleotide pair showing minimal nonterminalsequence mismatching with the genomes of as many adenovi-rus types as possible. The recent release of partial sequences ofhexon DNA for the subgenus B and E adenoviruses types 3, 4,7, and 16 (PC/Gene CD-ROM, Release 14.0; lntelli Genetics,Oxford, UK) offered the possibility of deriving primers show-ing minimal mismatches with these sequences and those of thesubgenus C adenoviruses types 2 and 5.2(> The sequences of thenew adenovirus primer set evaluated in this study (ADRJC1 andADRJC2; Fig. 2) were derived from the highly conserved DNAregion coding for the carboxyl end of the monomeric proteinII that forms the trimeric pseudohexagonal base of the adeno-virus hexon. This is the same region from which JCH 1 andJCH2 primers were derived,19 and, indeed, ADRJC2 has con-siderable overlap with JCH 1. The new primers yielded a prod-uct size of 140 bp. Mismatches were deliberately introducedinto the primer sequence such that each primer had a maxi-mum of two nonterminal mismatches when compared withknown DNA sequences of the hexon proteins of adenovirustypes 2 through 5, 7, and 16 (Fig. 2).
Sensitivity and Specificity of the PCRThe detection limit with primer set ADRJC1 and ADRJC2 was40 molecules adenovirus type 2 genomic DNA per reactionmixture. The adenovirus PCR amplified control subgenus B(types 3, 7, 11, and 14), C (type 2), D (types 8 through 10, 19,and 37), and E (type 4) DNA but not subgenus F (types 40 and41), control HSV, C. trachornatis plasmid or human DNA.
Retrospective Detection of AdenovirusOf the 59 lysis buffer-extracted conjunctival swabs tested ret-rospectively using ADRJC1 and ADRJC2, 54 were from swabspositive for adenovirus by culture, or PCR, or both with the H 1and H2 primer pair at the time of their collection (Table 1). Theremaining 5 extracts were negative by both of these tests. Thenew PCR detected 15 positive samples that were culture-negative and a total of 9 positive samples that were negativewith the HI and H2 primer pair. There were also 4 culture-positive specimens that were negative by both PCR primersets. If all adenovirus culture-positive and all PCR-positive,culture-negative eye swabs were considered to indicate ocularadenovirus infection, then the new primer pair had the highestsensitivity. Thus, for ADRJC1 and ADRJC2, the sensitivity was51 of 55 (92.7%), higher than HI and 112 at 42 of 55 (76.4%, P0.004), and culture at 40 of 55 (72.7%, ^0.019). The sensitiv-ities of culture and the PCR of primer pair H 1 and H2 were notsignificantly different (P 0.845).
Prospective Analysis of Clinical SamplesAll eye swabs sent to the diagnostic virology laboratory forvirus isolation over a 2-year period were subjected also to the
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933210/ on 02/04/2018
92 Cooper et al. /OVA", January 1999, Vol. 40, No. 1
5'-GCC GCA GIG GTC IIA CAT CCA CAT C-3' H1 Primer
A (12 and Ad5Ad3
5-GCC GCA GTG GTC TTA CAT GCA CAT C-3'5 -GCC CCA ATG GGC ATA CAT GCA CAT C-3'
H2 Primer(complementary sequence)
3'-TG AAA CTG TAG ULU IAA, C"AL l.ACO5-AC TIT GAC ATC CGC GGC GIG CTG-3'
A (12 and A (15Ad3
5'-AC TTI GAC ATC CGC GGC GTG CTG-3'5'-TC IT I GAC ATT CGG GGG GTG C I 1-3'
FIGUKI: 1. Location of primers II1 and 1-12.IH JCH1 and JCH219 within the hexon gene of aclenovirus types2 and 520 and 3 (PC/Cicnc CD-KOM, Release 14.0). Underscore indicates position of primer mismatch.
aclenovirus PCR. A total of 2743 specimens were tested overthis period. These data are shown in Table 2, and an exampleof the PCR results is shown in Figure 3. The aclenovirus PCRdetected 167 positive samples that were negative by culturecompared with 29 culture-positive samples that were missedby the PCR. These 29 specimens were examined further. Fourof the isolates were chosen at random (aclenovirus type 3 \n =3J and aclenovirus type 4 \n = 11) and subjected to PCR. Allfour isolates were positive, indicating that failure to detectthem in the original virus transport medium was not due to asequence heterogeneity with the new primer pair. Of the 29original specimens, 23 were available for reiesting. After repeatextraction with lysis buffer, these specimens remained nega-tive by PCR. However, when 15 of them were treated using thePEG precipitation method, 14 became positive by PCR. sug-gesting very low levels of virus, the presence of PCR inhibitorsthat were not being removed by the lysis buffer extraction, orboth. The specimen that remained negative by PCR after bothextraction procedures was an aclenovirus type 1. If all thespecimens that were aclenovirus culture-positive or lysis buff-er-extracted PCR-positive were taken to indicate ocular infec-tion, then PCR was much more sensitive, detecting 386 of 415(930%) compared with culture that detected 248 of 415(598%, P < 0.001). None of the specimens positive by culturefor HSV (167) were positive by PCR for aclenovirus. a specific-ity for the aclenovirus PCR of 100%.
Overall, aclenovirus was isolated from 248 of 2743 (90%)of conjunctival swabs compared with the PCR. which detectedaclenovirus DNA in 386 of 2743 (14.1%) specimens. HSV wasisolated from 167 of 2743 (6.1%) of the specimens.
DISCUSSION
The design of primers ADRJC1 and ADRJC2 achieved only twononterminal mismatches when compared with the hexon DNAsequences of six aclenovirus types (2 through 5, 7, and 16)representative of three of the four subgenera infecting the eye(H. C, and E). Despite these deliberate mismatches, the PCRwas still capable of detecting 40 copies of aclenovirus type 2DNA. In the retrospective studies, the ADRJC1 and ADRJC2primer pair identified 9 positive samples (8 of which were alsoculture-positive) that were not detected by the HI and H2primer pair, presumably reflecting the difference in primer-target mismatching of the two pairs of primers. An alternativeapproach would be the use of degenerate primers. However,these tend to be more useful in circumstances in which targetcopy numbers are relatively high because they can lack sensi-tivity. In addition, they require a more complex thermal cy-cling program and, because of the multiplicity of primers, canresult in nonspecific product bands on the gel / 1 Clinical sam-ples often have low target copy numbers, and it is an advantage
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933210/ on 02/04/2018
IOVS, January 1999, Vol. 40, No. 1 PCR for Adcnoviral Conjunctivitis 93
5'-GAC ATG ACT TTC GAG GTC GAT CCC ATG GA-3' Primer ADRJCl
Ad2 5 -GAC ATG ACT I IT GAG GTG GAT CCC ATG GAC GAG CCC ACC C IT CTT-3'
Fir.UKi; 2. Location of primers ADRJCl and ADRJC2 within the hexon gene of adenovirus types 2 and 52<l
and 3, A, 7, and 16 (PC/Gene CD-ROM, Release 14.0). Underscore indicates position of primer mismatch.
for a busy routine diagnostic laboratory when tests are simpleand straightforward. TABLE 1. Retrospective Comparison of Virus Culture
The new PCR is significantly more sensitive than culture for and Polymerase Chain Reaction with Either H1/H2 orthe detection of adenovirus in conjunctival swabs. This may ADRJC1/ADRJC2 Primer Pairs for Detection ofreflect a loss of viable virus during transport to the laboratory, a Adenoviruses in Conjunctival Swabsvery slow growth rate of some strains of adenovirus, toxicity of virus No ofthe samples, and, occasionally, bacterial contamination of the cell Culture H1/H2 ADRJC1/ADRJC2 Specimenscultures. The possibility of contamination of the PCR in culture-negative, PCR-positive specimens is unlikely because of the neg- + + + 28ative results of the contamination controls included in each PCR + ~ + 8run, the technical precautions taken to minimize DNA carryover, + ~~ ~ ^the random distribution of these specimens, and the fact that of '167 swabs culture-positive for MSV, none were positive by the _ _ _ /
adenovirus PCR. Also, the culture-positive, PCR-negative sped-mens remained PCR-negative when reextracted with lysis buffer +. positive; - , negative (for adenoviruses).
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933210/ on 02/04/2018
94 Cooper et al. IOVS, January 1999, Vol. 40, No. 1
TABLE 2. Prospective Comparison of Virus Cultureand Polymerase Chain Reaction with ADRJC1 andADRJC2 Primer Pair for Detection of Adenoviruses inEye Swabs
and only became PCR-positive when concentrated by PEG pre-cipitation.
In view of the latter observation we considered replacingthe lysis buffer extraction method by PEG precipitation. In asmall pilot study (results not shown), fewer positive specimenswere detected with PEG than with lysis buffer. There is noperfect extraction system for PCR, and different methods re-move a different range of nonspecific inhibitors. The pilotstudy suggests that those found in eye swabs are more likely tobe removed by lysis buffer than PEG, but neither method is100% successful. In addition, the PEG procedure requires moremanipulations than the lysis buffer method, which is simplerand the one we continue to use.
The single-target PCR using primers ADRJC1 and ADRJC2could replace H1/H2 PCR and cell culture for detection ofadenoviruses in conjunctival swabs. No false-positive PCR re-
L C + 1 2 3 4 5 6
140bp=
FIGURE 3. Polyacrylamide gel of the adenovirus polymerase chainreaction products. L, 1 kb Plus DNA ladder (Gibco-BRL, Paisley, Scot-land, UK); C, contamination control (sterile distilled water); +, positivecontrol (adenovirus type 2 DNA); lanes 1 through 6, conjunctivalswabs of which 2 and 6 are positive for adenovirus DNA.
suits were obtained, confirming the success of the simpleanticontamination measures used1114 and the applicability ofDNA amplification as a technique for the routine diagnosis ofocular infections. In addition, the PCR result can be availablefor the physician within 8 hours of the laboratory receiving thespecimen compared with 1 to 4 weeks for conventional cellculture. However, we do continue to use cell culture forconjunctival swabs to detect HSV. After 1 week, by which timeHSV will have become apparent, cultures are discarded exceptfor those inoculated from swabs positive by the adenovirusPCR, where incubation is continued for a further 3 weeks. Thisenables us to isolate many of the adenoviruses detected by PCRthat can then be typed for epidemiologic studies.
AcknowledgmentsAdrian C. Yeo was the recipient of a postgraduate studentship (Over-seas Staff Development Program) funded by the Singapore Polytechnic,Republic of Singapore.
References1. Sharp IR, Wadell G. Adenoviruses. In: AJ Zuckerman, JE Banatvala,
JR Pattison, eds. Principles and Practice of Clinical Virology.Chichester, UK: John Wiley; 1994:287-308.
2. Richmond SJ, Burman R, Crosdale E, et al. A large outbreak ofkeratoconjunctivitis due to adenovirus type 8./Hyg (Cambridge).1984;93:285-291.
3. Warren D, Nelson KE, Farrar JA, et al. A large outbreak of epidemickeratoconjunctivitis: problems in controlling nosocomial spread.
prevention of ocular viral and chlamydial infections. Crit RevMicrobiol. 1989; 16:369-418.
5. Ankers HE, Klapper PE, Cleator GM, Bailey AS, Tullo AB. The roleof a rapid diagnostic test (adenovirus immune dot-blot) in thecontrol of an outbreak of adenovirus type 8 kerato-conjunctivitis.Eye. 1993;7(suppl):15-17.
6. Morris DJ, Klapper PE, Cleator GM. Viral cross infection in oph-thalmology. Postgrad Doctor Caribbean. 1995; 11:152-160.
7. Gordon YJ, Naesens L, DeClerq E, Maudgal PC, Veckeneer M.Treatment of adenoviral conjunctivitis with topical Cidofovir. Cor-nea. 1996; 15:546.
8. Killough R, Klapper PE, Bailey AS, Sharp IR, Tullo A, Richmond SJ.An immune dot-blot technique for the diagnosis of ocular adeno-virus infection./ Virol Methods. 199O;3O:197-2O4.
9. Wiley L, Springer D, Kowalski R, et al. Rapid diagnostic test forocular adenovirus. Ophthalmology. 1988;95:431-433.
10. Kinchington PR, Turse SE, Kowalski RP, Gordon YJ. Use of poly-merase chain amplification reaction for the detection of adenovi-ruses in ocular swab specimens. Invest Ophthalmol Vis Sci. 1994;35:4126-4134.
11. Morris DJ, Bailey AS, Cooper RJ, et al. Polymerase chain reactionfor rapid detection of ocular adenovirus infection. / Med Virol.1995;46:126-132.
12. Hussain MAS, Costello P, Morris DJ, et al. Comparison of primersets for detection of faecal and ocular adenovirus infectionusing the polymerase chain reaction. J Med Virol. 1996;49:187-194.
13. Morris DJ, Klapper PE? Kiltough R, Bailey AS, Nelson J, Tullo AB.Prospective study of adenovirus antigen detection in eye swabs byradioimmune dot-blot. Eye. 1995;9:629-632.
14. Turner PC, Bailey AS, Cooper RJ, Morris DJ. The polymerasechain reaction for detecting adenovirus DNA in formalin-fixed,paraffin-embedded tissue obtained post mortem./Infect. 1993;27:43-46.
15. Lusher M, Storey CC, Richmond SJ. Plasmid diversity within thegenus Chlamydia. J Gen Microbiol. 1989;135:ll45-1151.
16. McElhinney LM, Cooper RJ, Morris DJ. Multiplex polymerase chainreaction for human herpesvirus-6, human cytomegalovirus andhuman (3-globin DNA / Virol Methods. 1995;53:223-233.
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933210/ on 02/04/2018
IOVS, January 1999, Vol. 40, No. 1 PCR for Adenoviral Conjunctivitis 95
17. McNemar Q. Note on the sampling error of" the difference betweencorrelated proportions or percentages. Psychometrika. 1947; 12:153-157.
18. Allard A, Girones R. Juto P, Wadell G. Polymcrase chain reactionfor the detection of adenoviruses in stool samples. / Cliu Micro-biol. 1990:28:2659-2667.
19. Hierholzcr JC, llalonen PI;, Dahlcn PC), Hingham PG, McDon-ough MM. Detection of adenovirus in clinical specimens bypolymcrase chain reaction and liquid-phase hybridization quan-
titated by time-resolved fluoromciry../ Clin Microbiol. 1993:31:1886-1891.
20. Kinloch R, Mackay N, Mautner V. Adenovirus hexon: sequencecomparison for subgroup C serotypes 2 and 5. J Biol Cbem.1984:9:6431-6436.
21. Compton T. Degenerate primers for DNA amplification. In: MAInnes, DM Gelfand, JJ Sninsky, T| White, eds. PCR Protocols: AGuide to Methods and Applications. London: Academic Press;1990:39-45.
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933210/ on 02/04/2018
