Journal of Medical Virology 37:149-157 (1992)

Detection of Adenoviruses in Stools From Healthy Persons and Patients With Diarrhea by Two-step Polymerase Chain Reaction Annika Allard, Bo Albinsson, and Goran Wadell Department of Virology, University of Ume&,Urnec, Sweden The use of the polymerase chain reaction (PCR) for detection of human adenoviruses in diluted stool samples was investigated. Two sets of nested primers, including primers specific for the hexon-coding region and for the E l B region of enteric adenoviruses (EAd), were assessed by two-step amplification. The primers constitute two different PCR systems designed for the detection of adenoviruses belonging t o all six subgenera (A-F), and the t w o EAds Ad40 and Ad41, respectively. In a two-step PCR mediated amplification a single virus particle was detected when the two sets of general hexon primers or EAd specific primers were used. Earlier results from PCR detection of adenoviruses in stool from children suffering from diarrhea gave indications that adenovirus particles are commonly shed in stools without being identified as the cause of ilI ness [Allard et a I.: Journal of Clinical Microbiology 28:2659-2667, 1990). Therefore, the general and the EAd specific PCR assays were assessed on 150 stool specimens from three groups including 50 healthy children, 50 healthy adults, and 50 adults suffering from diarrhea. When the t w o sets of general hexon primers were used, 25 of the 50 specimens from the healthy children (mean age 21 months) were found positive by two-step PCR amplification. Nine of the 50 specimens from the healthy adults (mean age 32 years) were found positive whereas 12 of the 50 specimens from sick adults (mean age 31 years) gave amplification products, using the t w o sets of general hexon primers in a nested fashion. None of the 150 specimens were found to be positive by two-step PCR amplification using the two sets of EAd-specific primers. The sensitivity of the PCR, together with its simplicity and reduced time scale compared t o other detection methods, support the potential of this technique as an additional method for routine detection of human adenovirus infections.

INTRODUCTION Adenoviruses are ubiquitous lytic DNA viruses that infect most animal species. There are 47 known adenovirus serotypes [Hierholzer et al., 19881 which fall into six subgenera (A to F) based on DNA homology and biological and biochemical properties. A great intraserotypic genomic variability was demonstrated for all serotypes analyzed so far by use of restriction enzyme analysis, and several distinct viral entities designated genome types have been identified in each case [Wadell et al., 19801. Tropism varies according to subgenus and type and clinical presentations range from mild to severe infections of the upper respiratory tract caused by types in subgenera B, C , and E, to epidemic keratoconjunctivitis caused by members of subgenus D, infantile diarrhea by subgenus F, and subclinical infections by most other types [Straus, 1984; Wigand and Adrian, 19861. Nothing is known about the molecular basis for these differences in pathogenicity. The subgenus F contains the two so-called fastidious or enteric adenoviruses (EAds) serotypes 40 and 41. After rotaviruses, the EAds are now recognized as the second most commonly identified agent in stools of infants and young children with gastroenteritis. Although acute infection with adenoviruses is sometimes severe, it is rarely fatal in otherwise normal adults. Fatal adenovirus infections have been reported, however, in very young children and in immunocompromised individuals [Zahradnik et al., 1980; Siegal et al., 1981; Flomenberg e t al., 1987; Ljungman et a]., 19891. Adenoviruses are regarded as opportunistic pathogens of immunosuppressed patients [Burns and Saral, 19851. This is also corroborated by epidemiological studies [Fox et al., 19771. In addition to acute disease, adenoviruses also cause persistent infections characterized by prolonged and intermittent fecal excretion. Approximately half of infants and young children infected with adenovirus type l (Adl), Ad2, Ad3, or Ad5 excrete viruses for long periods, sometimes

C 1992 Wiley-Liss, Inc.

KEY WORDS: EAd, persistence, nested PCR, diagnostic method C 1992 WILEY-LISS, INC.

Accepted for publication December 9, 1991. Address reprint requests to Annika Allard, Department of Virology, University of Umea, ,5901 85 Ume5, Sweden.

150 years [Fox et al., 19771. The cell types responsible for virus replication in persistently infected individuals have not been identified, although several lines of evidence indicate that lymphoid tissues may be the primary source of persistent virus. Viral DNA, from subgenus C adenoviruses has been detected in lymphoid tissues and peripheral blood lymphocytes from apparently normal individuals [Green e t al., 1979; Horvath et al., 1986; Ginsberg et al., 19871. Like most other viruses associated with diarrhea, the EAds are fastidious with a limited capacity to grow in primary cells such a s human embryonic kidney cells (HEK) or human diploid fibroblasts [Albert, 1986; deJong et al., 19831. However, 293 cells, immortalized by transformation with the human Ad5 E l A and E1B region [Takiff et al., 19811 or HEp-2 cells [deJong et al., 19831 support limited replication of EAds. A549 cells are useful for the isolation of all other adenoviruses. Alternative methods for direct detection of adenoviruses have been developed such as latex agglutination [Grandien e t al., 1987 I, immunofluorescence, and dot blot hybridization [Hyypia 1985; Stiilhandske et al., 19851. Furthermore, enzyme-linked immunosorbent assay (ELISA) [Johansson et al., 1980, 19851, monoclonal antibodies [Herrmann et al., 1987; van der Avoort et al., 1989; and Wood et al., 19891, SDS-polyacrylamide gel electrophoresis [Wadell, 19791 and DNA restriction analysis [Uhnoo et al., 1983; Wadell et al., 1980, 19861 have been used to diagnose diarrheal disease caused by adenoviruses. The polymerase chain reaction (PCR) is a n in vitro method for primer-directed enzymatic amplification of specific target DNA sequences [Saiki e t al., 19851. A few copies of target DNA can be amplified to a level detectable by gel electrophoresis or Southern blot hybridization. Recently we have presented a PCR method for detection of adenoviruses in stool samples [Allard et al., 19901. The results indicated a high prevalence of persistent adenoviruses in children suffering from diarrhea. We have now identified general nucleotide sequences that can serve a s oligonucleotide primers in a two-step PCR-assay to investigate the persistence of adenoviruses in stools of healthy children and healthy adults as well a s in adults suffering from diarrhea.

MATERIALS AND METHODS Study Group and Specimens Stools of 50 adults suffering from diarrhea were selected from specimens submitted to the Regional Diagnostic Virus Laboratory a t the University Hospital of Umeh from July to December 1990. The 50 samples were shown to be negative for enterovirus by virus isolation in cell cultures, negative for rotavirus by Rotalex (Orion Diagnostica, Helsinki, Finland) and negative for adenovirus by Adenolex (Orion Diagnostica). The 100 samples from healthy individuals originated from volunteers, all living in Umeii. These samples were collected during July to December 1990. The 50 healthy children were all below 3 years of age (mean age 21 months) and all the adults included in the study, both sick and healthy, were between 20 and 50 years old

Allard et al. (mean age 31 and 32 years, respectively). The inclusion criterium was the absence of signs of either respiratory or diarrheal infection during two weeks prior to sampling. The fecal specimens were collected in transport medium (phosphate-buffered saline, 20mM N-2-hydroxyethylpiperazine-N’-2-ethanesulfonic acid [HEPES] [PH 7.31, 1%bovine serum albumin, 5% sucrose, 50 pg of gentamicin per ml), and clarified by centrifugation for 5 min at 8 , 8 0 0 ~ g.

Pretreatment of Samples A NaOH extraction method was used to disrupt the adenovirus capsid. All reactions were carried out with clarified stool specimens in transport medium. A 50 pl sample was treated with NaOH at a final concentration of 0.5 M to denature the capsids. This treatment was followed by incubation in a 37°C waterbath for 15 min and neutralization with HC1 a t a final concentration of 0.5 M. The samples had to be diluted 10-100-fold before the PCR amplification due to the fact that the presence of inhibiting factors in the original stool suspension may impair the function of the Taq-polymerase. Preparation of Viral DNA Lysates of cells infected with the adenovirus prototype or reference strains (Table 2) were inoculated onto subconfluent monolayers of A549 cells or 293 cells in roller tubes. Three to 5 days postinfection cells were harvested and intracellular viral DNA was extracted by the method of Shinagawa et al. [ 19831. Latex Agglutination Test For routine detection of adenovirus and rotavirus in stool-specimens, latex reagents, Adenolex and Rotalex, respectively, (Orion Diagnostica, Finland) were used according to the instructions of the manufacturer. The reagents consist of particles coated with antibodies specific for adenovirus or rotavirus. The method is based on agglutination of latex beads when the relevant antigens are present. Primers The primer pair used for detection of human adenoviruses in general by a one-step amplification, hexAA1885lhexAA1913, together with the specific pair of primer 41AA142141AA358, used for detection of enteric adenoviruses type 40 and 41, have been described earlier [Allard et al., 19901. The DNA sequences of the open reading frame of the hexon genes of Ad2 [Akusjarvi et al., 19841, Ad5 [Kinloch et al., 19841, Ad40 [Toogood et al., 19891, and Ad41 [Toogood and Hay, 19881were examined with the GCG computer program (version 6) to locate two conserved sequences. The identified oligonucleotides were arranged as primers for detection of adenoviruses by two-step amplification in a nested fashion. For the specific detection of EAds, Ad40 and Ad41, primers used in a nested two-step amplification were chosen from the gene encoding early region E1B (Allard e t al., unpublished data) (Fig. 1;Table I).

Two-step PCR for Detection of Adenoviruses

151

Elo Elb

Hexon

/ \ HexAA1085

+

. )

neHexAA1893 \

41AA142

+

+

ne41AA206 . )

3’ Size of product

HexAA1913

300 bp

neHexAA1905

142 bp

41AA350

2107 bp

4- ne41AA356

1523 bp

Fig. 1. Locations of the four sets ofoligonucleotide primers in the 36-kb adenovirus genome

TABLE I. Oligonucleotide Primers for PCR Amplification of Adenoviruses Virus type (region)

Position

Ad2 (hexon) Ad40 (hexon) Ad41 (hexon) Ad2 (hexon) Ad2 (hexon) Ad41 (ElB) Ad41 (ElB) Ad41 (ElB) Ad41 (EIB)

Primera

Sequence

Amplimer length (bp)

5’-GCCGCAGTGGTCTTACATGCACATC-”‘ 300 hexAA1885 5’-CAGCACGCCGCGGATGTCAAAGT-3’“ hexAA1913 nehexAA1893 5’-GCCACCGAGACGTACTTCAGCCTG-3’ 142 nehexAA1905 5’-TTGTACGAGTACGCGGTATCCTCGCGGTC-3’ 5’-TCTGATGGAGT’M’TGGAGTGAGCTA3’ 2,187 41AA142 41AA358 5‘-AGAAGCATTAGCGGGAGGGTTAAG-3’ 5’-GTCTGGTGGGCTGATTTGGAAGATG3‘ 1,523 ne41AA2O6 5’-CAGGGCCACTTTGGCAAACAAATC-3’ ne41AA356 aPrimers were namedin such a way a s to describe the adenovirus type or adenovirus gene, the initials of the individual who discovered the 18,858-18,883b 19,136-19,158b 18,937-18,960b 19,051-19,079b 1,421-1,446d 3,585-3,608d 2,061-2,085d 3,561-3,584d

rimer, and the sequence position with the last figure deleted. ‘The sequence positions of the hexon primers are referred to the Ad2 hexon region [Akusjarvi et al., 19841. ‘These primer sequences are shared amongAd2 [Akusjarvi et al., 19841,Ad40 [Toogood et al., 19891and Ad41 [Toogood and Hay, 19881. dThe sequence position is from Allard et al. (unpublished data).

All primers were synthesized on a Beckman DNA SM automated DNA synthesizer.

Polymerase Chain Reaction Amplification of adenovirus DNA was carried out according to two different protocols with two different sets of primers. la. The general primer pair hexAA1885lhexAA1913 was used in a one-step amplification of the DNA prepared from infected cells or directly on pretreated stools (30 cycles). lb. The general primer pair nehexAA1893l nehexAA1905 is positioned between the two hexAA1885lhexAA1913 primers. This primer pair was used in combination with the hexAA1885lhexAA1913 primer pair in a nested primer two-step amplification of the adenovirus DNA (30 cycles hexAA1885/ hexAA1913,25 cycles nehexAA1893lnehexAAl905~. 2a. The specific enteric adenovirus primer pair 41AA142141AA358 was used in a one-step amplification of the DNA prepared from infected cells or on pretreated stools (30-35 cycles). 2b. The second specific primer pair for enteric adenoviruses ne41AA206lne41AA356 is positioned between the two 41AA142141AA358 primers. This primer pair was used in combination with the 41AA142141AA358

primer pair in a nested primer two-step amplification of the enteric adenovirus DNAs (30-35 cycles 41AA1421 41AA358.25 cycles ne41AA206Ine41AA356). For a typical one-step amplification reaction 10 p1 of a pretreated stool suspension diluted 10-fold in distilled water or 5-20 ng of extracted viral DNA were used. Amplification was carried out in a 50 p1 reaction mixture containing 16.6 mM (NH,) SO,, 67 mM Tris-HC1 ( P H 8.8 at 25”C), 6.7 mM MgCl,, 10 mM P-mercaptoethanol, 200 pM each deoxynucleoside triphosphate (i.e., dATP, dCTP, dGTP, and dTTP), 100 pg BSA per ml, 0.08 pM of each primer needed for the specific reaction, and 2 U of thermostable Taq DNA polymerase (Perkin-Elmer, Cetus). The samples were overlaid with 75 pl of mineral oil to prevent evaporation. Thermal cycling of the amplification mixture was performed in a programmable heat block (Hybaid Ltd, Teddington, Middlesex, UK, or Techne PHC-2 Cambridge, UK). In all PCR assays t h e first cycle of denaturation was carried out for 4 minutes at 94°C. The conditions for amplification using the two sets of general hexon primers were denaturation at 92°C for 30 sec, annealing at 55°C for 30 sec, and extension at 72°C for 30 sec, respectively. With the use of the two different sets of specific enteric adenovirus primers, which both yield amplified products of more than 1,500 bp, a cycle comprises dena-

Allard et al.

152 turation for 30 sec at 92"C, annealing for 30 sec a t 60°C, and primer extension for 1min at 72°C. For a typical two-step amplification 1 pl (1150) was taken from a 30 cycled one-step amplification and added to a new batch of 50 p1 PCR reaction mixture containing PCR buffer, the four nucleotides a t a final concentration of 200 pM, 100 pg of BSA per ml, 2 U of Taq polymerase, and 0.16 p M of each primer of nehexAA1893lnehexAA1905 or ne41AA2061 ne41AA356, depending on which system was used. Twenty-five additional cycles were then performed. As positive controls in all experiments, lOpg of Ad2 or Ad41 DNA were used. PCR mixtures without DNA were used a s negative controls placed in between every fifth sample.

Detection and Hybridization After the last cycle, 10 pl of the reaction mixture were analyzed for amplification products by gel electrophoresis. Large fragments (1,500-2,200 bp) of amplified DNA were separated on a 1% SeaKem ME agarose gel, whereas small fragments of amplified DNA (142300 bp) were separated on a 2% NuSieve GTG + 1% SeaKem ME agarose gel (FMC Bioproducts, Rockland, MA, USA) by electrophoresis and stained with ethidium bromide [Sambrook et al., 19891. Gels were blotted onto nylon filters which were used for hybridization. The entire genomes of Ad31 (subgenus A), Ad3 (subgenus B), Ad2 (subgenus C), Adl9a (subgenus D), Ad4 (subgenus E), and Ad40 and Ad41 (subgenus F) were labeled with [ c Y - ~ ' P ] ~ C(Amersham, TP England) by the multipriming DNA labeling technique (Amersham) to a n activity of lo8 to lo9 dpm/lJ-g and used as probes under stringent hybridization and washing conditions (Tm -12°C). Filters were exposed to a Cronex 4 film (Dupont)for 16 to 40 h r a t -70°C.

TABLE 11. Adenovirus Prototvoe and Reference Strains Subgenus Type Strain or description A

12

B

31 3 7

C

11 1 2 20

D

E F

35 4 40 41

F-3072-86, prototypelike pattern (Huie)a 1315, prototype GB, prototype Gomen, prototype Slobitski, prototype Ad71, prototype Ad6, prototype 931, prototype Holden, prototype RI-67, prototype HoviX, prototype D389, pattern like that of representative strain N597"

aDNA restriction enzyme pattern.

in Ad2, was tested against 12 different adenovirus types representing all six subgenera (Table 11).Positive results were obtained with all types when analyzed on a n agarose gel (Fig. 2b). The hexon region of only four human adenovirus types has been sequenced so far, e.g., Ad2 [Akusjarvi et al., 19841, Ad5 [Kinloch et al., 19841, Ad40 [Toogood et al., 19891, and Ad41 [Toogood and Hay, 19881, respectively. The general hexon primers hexAA1885 and hexAA1913 show a 100%homology with the hexons of these four types. However, the upstream internal hexon primer used in the two-step amplification, nehexAA1893, has two mismatches in Ad40 and one mismatch in Ad41. The downstream internal primer, nehexAAl905, has three mismatches in Ad40 and four mismatches in Ad41 compared to the sequences of Ad2 and Ad5, which are 100% homologous to the primers. These mismatches do not decrease the effiRESULTS ciency of annealing, at least not to detectable levels. Specificity of Primers Therefore the annealing temperature was not altered. In the two-step amplification for the detection of enThe hexon primers, hexAA1885 and hexAA1913, used in the one-step amplification were demonstrated teric adenoviruses, a pair of primers internal of the two to be general in the sense that they were able to detect specific EAd primers, 41AA142141AA358, was used 18 different adenovirus types representing all six sub- (Fig. 1).The specificity of these new internal primers genera [Allard e t al., 19901. This pair of primers flanks ne41AA206 and ne41AA356, which flank the region the region between nucleotides 18,858 and 19,158 in between nucleotide 2061-3584 in Ad41 and the correAd2 and the corresponding regions of the other human sponding region in Ad40, was tested on the same set of adenoviruses. The specific primers for the two EAds, 12 adenovirus types. Amplification was only detected 41AA142 and 41AA358, which flank the region be- when the two EAds Ad40 and Ad41 were used as temtween nucleotides 1,421-3,608 in Ad41 and the corre- plates (Fig. 2a). In hybridization experiments the responding region in Ad40 have earlier been shown to gion E1B of Ad40 and Ad41 showed very little homolonly amplify DNA sequences specific for Ad40 and ogy to the corresponding region of other human adenovirus types (Allard et al., unpublished data). Ad41 IAllard et al., 19901. Two conserved nucleotide sequences localized within Therefore this area was chosen for selection of specific the first amplimer created by the primers hexAA18851 primers. The only ElB-sequences available for comhexAA1913, were identified. These two new primers, puter comparison were those of Ad2, Ad4, Ad5, Ad7, nehexAA1893 and nehexAA1905, could be arranged in and Ad12. The aim was to find primers creating short a nested fashion to direct a two-step amplification that amplimers, but areas with low homology to other adeallowed detection of adenoviruses belonging to all sub- novirus types seemed to be located a t the terminal ends genera (Fig. 1).The specificity of the primers, which of the Ad40 and Ad41 E1B genes, respectively. In a test flank a region between nucleotides 18,937 and 19,079 with 12 different randomly chosen pairs of primers

Two-step PCR for Detection of Adenoviruses

Fig. 2. A: PCR amplification of DNA from 12 different adenovirus types representing all six subgenera with the EAd-specific primers ne41AA206 and ne41AA356 used in the two-step amplification. Amplified DNA resulted in a sequence of 1,523 bp. B: Adenovirus DNA amplified by PCR with the general hexon primers nehexAA1893 and nehexAA1905 used in the two-step amplification. Agarose gel electrophoresis of amplified products shows the characteristic 142 bp band. In all panels lane m = molecular weight standards (lambda HindIII digest and/or 6x174 HaeIII digest).

(21-30 mers), contained within the region amplified by the primers ne41AA206 and ne41AA356, we failed to find specific sequences for detection of EAds. To ensure the completion of the reaction for the two sets of enteric specific primers 41AA142141AA358 and ne41AA2061 ne41AA356, we recommend a n extension time of 1min.

Sensitivity The sensitivity of the different sets of primers differed with respect to the size of the amplimer produced. If pure Ad41 virions were treated with NaOH and amplified by use of EAd specific primers, using both sets separately in one-step amplifications, it was possible to detect lo3 virus particles by PCR. If the two different pairs of general hexon primers were used separately under the same conditions, the sensitivity was increased 10-fold to lo2 virus particles (data not shown). The determination of sensitivity by using DNA from pure virions was not comparable to the sensitivity obtained when virus DNA from stool samples was amplified. Consequently it was necessary to simulate the characteristic conditions of the clinical sample. Therefore, plasmids containing the viral regions to be amplified were diluted in supernatants of negative stool sam-

153

Fig. 3. Sensitivity of PCR for the detection of adenovirus DNA using plasmids containing the viral region to be amplified, diluted in supernatants of negative stool samples. The numbers above the figures represent the numbers of viral particles per genome copy contained in 10 p1 (one-fifth) of the reaction mixture prior to amplification. A: Ad41 DNA was amplified with the general hexon primers hexAA1885ihexAA1913. A l i l 0 0 aliquot was transferred to a second amplification, primed by nehexAA1893inehexAAl905 in a nested fashion, creating an amplimer of 142 bp. B: Ad41 DNA was amplified with the EAd-specific primers 41AA142141AA358 and ne41AA2061 ne41AA356 in a two-step amplification, as described in A, creating a n amplimer of 1,523bp. Lane n = negative control. Lane m= molecular weight standards (lambda HindIII digest and/or 6x174 HaeIII digest).

ples. In this PCR experiment the sensitivity of all sets of primers was still the same as in the experiment with the virion DNA as template. In order to determine the sensitivity of the two-step amplification, we performed a second amplification of the products obtained from PCR of the plasmids diluted in negative stool supernatants. A two-step amplification using the internal set of EAd specific primers lowered the detection limit to a single virus particle. By two-step amplification under the same conditions using the internal hexon primers i t was possible to detect 1 copy of the virus genome (Fig. 3). Stool supernatants may inhibit the PCR. Eight different negative samples were mixed with a positive sample a t a ratio of 1:1, respectively, in eight independent reactions. One sample became negative in PCR amplification primed with the general hexon primers hexAA1885hexAA1913 (data not shown). These results confirm assumptions of quantitative differences of inhibitor factors in different stool supernatants which may affect the activity of the Taq polymerase. All the

154

Allard et al.

Fig. 4. Calibration of the two-step PCR amplification. A: Five different aliquots of amplified DNA from a NaOH-treated stool sample (typed to be Ad41 by DNA restriction enzyme analysis) primed by the hexon primers hexAA1885ihexAA1913, were transferred to a second PCR amplification performed with different number of cycles (15-30) using the nested hexon primers nehexAA1893inehexAAl905. This was made to optimize the conditions for the two-step amplification

system. Lane a = 0.1 pl,lane b = 0.5 pl, lane c = 1.0 pl, lane d = 2.5 pl, lane e = 5.0 pl, lane m = Molecular weight standards (lambda Hind111 digest and +X174 HaeIII digest). B: Calibration of the twostep amplification using the two sets of EAd-specific primers 41AA142141AA358and ne41AA206ine41AA356, creating a final amplified DNA product of 1523bp. The calibration was performed as described in A.

tests for sensitivity were done after NaOH treatment of the samples. The PCR-buffer recommended by New England Biolabs was used in all reactions. A specific calibration for each pair of primers concerning different concentrations of the compounds t h a t make up the buffer was not done is this study. We have compared the PCR buffers recommended by New England Biolabs, Promega (50 mM Tris-HC1 [PH 9.0 a t 25"C], 50 mM NaCl, 10 mM MgCl,, 200 pM of each dATP, dCTP, dGTP, and dTTP) and Perkin Elmer Cetus (50 mM KC1,lO mM Tris-HC1 [PH8.4 a t room temp.l,l.5 mM MgCl,, 100 pgiml gelatin, 200 FM of each dATP, dCTP, dGTP, and dTTP). In our hands the buffer recommended by New England Biolabs was preferred concerning sensitivity. The primer concentration of 0.08 pM,which was used, corresponds to approximately 2 x 10l2 molecules of each primer. An increased primer concentration was not required since none of the primers is made by a compromise of alternative bases or inosine to create degenerated oligonucleotides. The hexon primers hexAA1883 and hexAA1913 were shown to be general, and the mismatches found in some of the other primers are not located too close to the 3' ends where the elongation starts.

be considered. Therefore, the use of aerosol resistant tips is recommended, together with a n additional number of negative controls placed in between the samples to be analyzed. The two-step amplification can be performed in many different ways. We have chosen to transfer a small aliquot from the first amplification step to a new batch of PCR-buffer containing the internal set of primers. We have noted that the amount to be transferred from the first to the second step is more important than the number of cycles used in the second amplification. If a l / l O O aliquot of a given stool sample from the first amplification step of 30 cycles was transferred to the second amplification another 15 cycles were sufficient for detection of hexon-primed DNA whereas at least 20 cycles were required for detection of EAd-primed DNA (Fig. 4).If a s much as 1/10 or 1/20 of a n 30-cycled amplified stool sample was transferred to the second amplification step a degradation of the final amplified product was regularly seen. This phenomenon was noticed in the two-step amplification primed by both the set of general hexon primers and the set of EAd-specific primers. All 150 stool samples in this study were pretreated by the NaOH method, diluted 10-fold, and used in two-step PCR amplifications with the two sets of general hexon primers and the two sets of EAd specific primers. Among the 50 healthy children only one was positive for adenovirus by a one-step amplification using the general hexon primers hexAA1885/hexAA1913. By a two-step amplification with a n additional 25 cycles using the nested hexon

Adenovirus Detection in Clinical Samples When the two-step amplification is used a n increase of the contamination risk, caused by additional steps, compared to the one-step amplification reaction should

Two-step PCR for Detection of Adenoviruses

155

novirus positive individuals was equally spread among the groups. None of the 50 adults suffering from diarrhea was positive for adenovirus by a one-step amplification using the general hexon primers. However, 12 samples (24%) became positive in a two-step amplification primed by the two sets of hexon primers (data not shown). None of the 50 samples from healthy adults displayed the characteristic 300 bp amplified product on a n agarose gel when assayed by one-step amplification using the primer pair hexAA1885lhexAA1913. After the twostep amplification, nine samples exhibited the 142 bp product amplified by use of the primer pair nehexAA1893lnehexAAl905, when analyzed on a n agarose gel (data not shown). The selected individuals in both groups were between 20 and 50 years of age, but there was no predilection for a certain age group concerning adenovirus positivity. The samples from healthy children and the samples from the two groups of adults were not positive for enteric adenoviruses using the two sets of EAd-specific primers in a two-step PCR amplification. Southern blots of the agarose gels did not reveal additional amplified products.

Comments on the Method For detection of adenovirus particles in stool samples by PCR, a one-step amplification of 30-35 cycles primed by the general hexon primers hexAA1885ihexAA1913 Fig. 5. Two-step PCR amplification of DNA extracted by the NaOHis recommended for routine diagnosis. The positive method from 50 stool samples of healthy children under 3 years of age. samples can be confirmed to be enteric by use of the All samples were subjected to 30 cycles of amplification with the hexAA1885ihexAA1913 primer pair, followed by 25 cycles of amplifi- EAd specific pair of primers, 41AA142t41AA358. The cation with the nehexAA1893inehexAA1905 primer pair. Positive and negative controls transferred from the first to the second step of ampli- EAd specific primers can be useful for fast detection of fication are marked with exponent 1, and controls from the second step EAds directly in stool specimens, since enteric adenoviof amplification only, are marked with exponent 2. Lane m = Molecu- ruses are excreted in large amounts, up to 10'l particles lar weight standards (lamda Hind111 digest and (bX174 HaeIII digest). per gram of faeces at the acute stage of the disease. In a one-step PCR amplification using the EAd specific primers a large amplimer is synthesized. The length of this product reduces the efficiency of the PCR 10-100primers nehexAA1893lnehexAAl905, a s many a s 50% fold compared to the hexon-primed PCR. Therefore, the of the samples from healthy children exhibited the ex- PCR detection of EAds could be supplemented with a pected 142-bp amplimer when analyzed on a n agarose second step of PCR, DNA restriction enzyme analysis or gel (Fig. 5). The results are summarized in Table I11 monoclonal antibodies. However, the use of EAd type together with some additional data about the children. specific primers in a two-step amplified PCR represents By principal component analysis, (PCA) [Wold et al., the most sensitive method developed so far for detection 19841 we could not find any correlation of these data of enteric adenoviruses. To keep the virions as intact a s and the findings of adenovirus particles in the stools. possible, the stools should be kept a t -7O"C, preferably PCA is a projection method that combines the included for less than 1year to avoid false-negative results. Durvariables to a few underlaying descriptive dimensions, ing prolonged storage the viral capsid may be disrupted summarizing the systematic variation in a data matrix. and viral DNA become degraded mainly due to lysis of PCA is primarily used to get a n overview of dominant bacteria in the stools resulting in release of proteolytic patterns or major trends in the data. By PCA there was enzymes and DNase. Correct storage of stool samples is no difference between the 12 children that were nursed critical for the successful application of the EAd specific a t home, and those 38 that were nursed a t day care primers, since they direct the synthesis of long amcenters. The fact that 27 of the children analyzed had plimers of more than 2,000 bp. If the viral DNA is siblings did not seem to increase the possibility of a partially degraded the target sequence for the primer positive result. Moreover, if the children were subdi- pair 41AA142141AA358 may be damaged. By use of vided into three groups-less than 1year, between 1to this primer pair we have occasionally noted a n extra 2 years, and between 2 to 3 years-the presence of ade- amplified product of 530 bp. This product can be more

156

Allard et al.

TABLE 111. Results From Detection of Adenoviruses in Stools of 50 Healthy Children No. of individuals with No. of individuals Two-step PCR Siblings General hexon Nursed a t Nursed a t More than nursed at home Siblings one sibling day care center primers day care center 5 11 6 11 Positive for 25 19 adenoviruses 8 14 25 19 6 16 Negative for adenoviruses Total number 50 38 12 27 13 25

or less dominant depending on the amount of the DNA amplified and length of the storage time of the sample. Hybridization revealed that this fragment was located within the E1B region, but we cannot explain its origin since no further sequence homologies for the primers can be found in this region. Application of the two sets of general hexon primers requires only a 142 bp or a 300 bp template; these are less likely to be damaged than the longer templates. To eliminate the presence of inhibitor factors a proteinase K/phenol treatment of the stool samples can be a n alternative method [Allard et al., 19901. A further dilution of the stool samples (1OOX) can also be recommended. Sometimes i t is necessary to use hybridization as a complementary method to increase the sensitivity and to confirm the specificity as well. In our hands, with the use of all the different sets of adenovirus primers, a hybridization with specific probes did not increase the sensitivity.

DISCUSSION PCR provides a unique possibility to detect previously unrecognized adenoviruses or enteric adenoviruses that are fastidious, due to a lack of a n optimal isolation system. The two-step amplification system on NaOH-treated stool samples offers the possibility to detect one copy of adenovirus DNA (Fig. 3 ) . In a previous study inoculation of cell layers with patient specimens followed by DNA extraction was used a s a first biological amplification step before a one-step amplification by PCR. In our hands the combination of these two methods has given similar results as the two-step PCR amplification system. Therefore, considering the intrinsic risk for contamination during virus replication in cell culture prior to PCR amplification and the increasing need for a rapid diagnostic procedure, the two-step PCR should be preferred. We wished to determine whether adenoviruses were shed into stools of healthy adults and children. In the previous study of sick children suffering from diarrhea [Allard et al., 19901, we found that a high number of them shed adenovirus particles in their stools. By using the combined method of virus isolation in cell culture and subsequent one-step PCR amplification, as many as 85% of the sick children were positive for adenoviruses in general, and 27% were shedding enteric adeno-

Siblings nursed at home 0 2 2

virus type 41. Consequently, it was pertinent to assay also adults with diarrhea. In this study amplified products were obtained in 50% of the healthy children by a two-step PCR amplification using the two sets of general hexon primers. The investigation should be extended to allow conclusions concerning possible differences in the degree of interaction with other children regarding the risk to acquire adenovirus infections. It will be of interest to examine the stools of the members of the family with a healthy shedding child to have a more complete picture. Among the healthy adults a n amplified product was detected in 18% of the samples by a two-step amplification using the two sets of general hexon primers, whereas 24% of the stools of adults suffering from diarrhea contained adenovirus particles. The two-step PCR amplification is very sensitive and can easily pick up persistent adenovirus particles. Therefore, adenoviruses should not always be identified as the cause of illness when they are isolated from or detected in the stools of a n individual with disease. Infection with members of subgenus C and Ad3 a r e characterized by a prolonged intermittent excretion (up to 906 days) [Kidd et al., 1982; Fox et al., 1977; Brandt et al., 19851. No amplified products were found by two-step amplification using the two sets of EAd-specific primers. This confirms the notion that infections caused by enteric adenoviruses are rarely found in adults. Information on possible persistence of enteric adenoviruses is scanty. We have followed a 3-year-old boy infected with Ad41 for 4 months. In the acute stage of diarrheal disease and fever, 30 cycles of a one-step PCR amplification were sufficient for detection with both the general hexon primers hexAA1885lhexAA1913 or the EAd-specific primers 41AA142141AA358. After both 1 and 2 months, a two-step amplification primed by the general sets of hexon primers was required to detect virus particles in the stools. However, negative results were obtained already in the sample taken 1 month after the outbreak when the two sets of EAd-specific primers were used in a two-step amplification. After 3 months no viral DNA was detected by PCR regardless of what primer system was used. Thus it seems that Ad41 particles are frequently shed in stools during the acute stage but t h a t the excretion declines quite rapidly after recovery. The different sensitivity of the primer sets

Two-step PCR for Detection of Adenoviruses used in the two-step amplification may be due to the fact that the general hexon primers are more sensitive than the EAd specific primers or less likely that the general hexon primers detected persistent adenovirus particles from a n earlier infection caused by a nonenteric type. Consequently, adenovirus particles are quite frequently shed in stools of both healthy and sick individuals, more frequently in children than adults. This is also expected since both respiratory and enteric adenovirus infections are common during the first years of life. Finally, the two-step amplification procedure with general hexon primers could also be useful for detecting adenoviruses in other sources than stools, as in liquor or biopsies, or to assess the degree of fecal contamination in wells, sewage, or seawater.

ACKNOWLEDGMENTS We thank Dr Per Jut0 for kindly providing clinical stool specimens. We also thank Rolf Sjoberg for help with photography, and Katrine Isaksson for excellent typing.

REFERENCES Akusjarvi G, Alestrom P, Pettersson M, Lager M, Jornvall H, Pettersson U (1984): The gene for the adenovirus 2 hexon polypeptide. Journal of Biological Chemistry 259:13976-13979. Albert M (1986): Enteric adenoviruses. A brief review. Archives of Virology 88:l-17. Allard A, Girones R, Jut0 P, Wadell G (1990): Polymerase chain reaction for detection of adenoviruses in stool samples. Journal of Clinical Microbiology 28:2659-2667. Brandt CD, Kim HW, Rodriguez WJ, Arrobio J O , Jeffries BC, Stallings EP, Lewis C, Miles AJ, Gardner MK, Parrott RH (1985): Adenovirus and pediatric gastroenteritis. Journal of infectious Diseases 151:437443. Burns WH, Sara1 R (1985): Opportunistic viral infections. British Medical Bulletin 41:46-49. deJong JC, Wigand R, Kidd AH, Wadell G, Kapsenberg JG, Muzerie CJ, Wermenbol AG, Firtzlaff RG (1983): Candidate adenovirus 40 and 41: fastidious adenoviruses from human infant stool. Journal of Medical Virology 11:215-231. Flomenberg PR, Chen M, Munk G, Horwitz MS (1987): Molecular epidemiology of adenovirus type 35 infections in immunocompromised hosts. Journal of Infectious Diseases 155:1127-1134. Fox J P , Hall CE, Cooney MK (1977): The Seattle virus watch. VII. Observations of adenovirus infections. American Journal of Epidemiology 105:362-386. Ginsberg HS, Lundholm-Beauchamp U, Prince G (1987): Adenovirus as a model of disease. In Russell WC, Almond J W (eds):“Molecular Basis of Virus Disease.” Cambridge University Press, pp 245258. Grandien M, Pettersson CA, Svensson L, Uhnoo I (1987): Latex agglutination test for adenovirus diagnosis in diarrheal disease. Journal of Medical Virology 23:311316. Green M, Wold WSM, Mackey JK, Rigden P (1979): Analysis of human tonsil and cancer DNAs and RNAs for DNA sequences of Group C (serotypes 1, 2, 5, and 6) human adenoviruses. Proceedings of the National Academy of Sciences of the United States of America 76:66066610. Herrmann JE, Perron-Henry DM, Blacklow NR (1987): Antigen detection with monoclonal antibodies for the diagnosis of adenovirus gastroenteritis. Journal of Infectious Diseases 155:1167-117 1. Hierholzer JC, Wigand R, Anderson LJ,Adrian T, Gold JWM 11988): Adenoviruses from patients with AIDS: a plethora of serotypes and a description of five new serotypes of subgenus D (type 43-47). Journal of Infectious Diseases 158:804-813. Horvath J, Palkonyay L, Weber J (1986): Group C adenovirus DNA sequences in human lymphoid cells. Journal of Virology 59:189192. Hyypia T (1985): Detection of adenoviruses in nasopharyngeal speci-

157 mens by radioactive and non-radioactive DNA probes. Journal of Clinical Microbiology 21:730-733. Johansson ME, Uhnoo I, Kidd AH, Madeley CR, Wadell G (1980): Direct identification of enteric adenovirus, a candidate new serotype, associated with infantile gastroenteritis. Journal of Clinical Microbiology 12:95100. Johansson ME, Uhnoo I, Svensson L, Pettersson CA, Wadell G (1985): Enzyme-linked immunosorbent assay for detection of enteric adenovirus 41. Journal of Medical Virology 17:19-27. Kidd AH, Cosgrove BP, Brown RA, Madeley CR (1982): Faecal adenoviruses from Glasgow babies. Studies on culture and identity. Journal of Hygiene 88:463474. Kinloch R, Mackay N, Mautner V (1984): Adenovirus Hexon: sequence comparison of subgroup C serotypes 2 and 5 . Journal of Biological Chemistry 259:6431-6436. Ljungman P, Gleaves CA, Meyers J D (1989): Respiratory virus infection in immunocompromised patients. Bone marrow Transplantation 4:35-40. Saiki RK, Scharf S, Faloona F, Mullis KB, Horn GT, Erlich HA, Arnheim N (1985): Enzymatic amplification of p-globin genomic sequences and restriction site analysis for diagnosis of sickle cell anemia. Science 230:1350-1354. Sambrook J , Fritsch EF, Maniatis T (1989): “Molecular Cloning: A Laboratory Manual.” Cold Spring Harbor, NY: Cold Spring Harbor Laboratory. Shinagawa M, Matsuda A, Ishiyama T, Goto H, Sat0 G (1983): A rapid and simple method for preparation of adenovirus DNA from infected cells. Microbiology and Immunology 272317-822. Siegal FP, Dikman SH, Arayata RB, Bottone E J (1981):Fatal disseminated adenovirus 11 pneumonia in a n agammaglobulinemic patient. American Journal of Medicine 71:1062-1067. Stglhandske P, Hyypia T, Allard A, Halonen P, Pettersson U (1985): Detection of adenoviruses in stool specimens by nucleic acid spot hybridization. Journal of Medical Virology 16:213-218. Straus SE (1984) Adenovirus infections in humans. In Ginsberg HS (ed): “The Adenoviruses.” New York: Plenum Press, p 451. Takiff HE, Straus SE, Garon CF (1981): Propagation and in vitro studies of previously non-cultivable enteral adenoviruses in 293cells. Lancet ii:832-834. Toogood CIA, Hay RT (1988): DNA sequence of the adenovirus type 4 1 hexon gene and predicted structure of the protein. Journal of General Virology 69:2291-2301. Toogood CIA, Murali R, Burnett RM, Hay RT (1989): The adenovirus type 40 hexon: sequence, predicted structure and relationship to other adenovirus hexons. Journal of General Virology 70:32033214. Uhnoo I, Wadell G, Svensson L, Johansson M (1983): Two new serotypes of enteric adenovirus causing infantile diarrhoea. Development of Biological Standards 53:311318. van der Avoort HGAM, Wermenbol AG, Zomerdijk TPL, Kleijne JAFW, van Asten JAAM, Jensma P, Osterhaus ADME, Kidd AH, deJong JC (1989): Characterization of fastidious adenovirus types 40 and 41 by DNA restriction enzyme analysis and by neutralizing monoclonal antibodies. Virus Research 12:139-158. Wadell G (1979): Classification of human adenovirus by SDS-polyacrylamide gel electrophoresis of structural polypeptides. Intervirology 11:47-57. Wadell G, Hammarskjold ML, Winberg G, Varsanyi TM, Sundell G (1980): Genetic variability of adenoviruses. Annals of the New York Academy of Sciences 354:16-42. Wadell G, Allard A, Evander M, Li Q (1986): Genetic variability and evolution of adenoviruses. Chemica Scripta 26B:325-335. Wigand R, Adrian T (1986):Classification and epidemiology of adenoviruses. In Doerfler W (ed): “Adenovirus DNA, the Viral Genome and Its Expression.” Boston: Martinus Nijhoff, p 409. Wold S, Albano C, Dunn WJ, Edlund U, Esbensen K, Geladi P, Hellberg s,Johansson E, Lindberg W, Sjostrom M (1984): Multivariate data analysis in chemistry. In Kowalski BR (ed): “Chemometrics: Mathematics and Statistics in Chemistry, NATO AS1 Series C 138.”Dordrecht Reidel, pp 6 9 6 . Wood DJ, Bijlsma K, deJong JC, Tonkin C (1989): Evaluation of a commercial monoclonal antibody-based enzyme immunoassay for detection of adenovirus types 40 and 41 in stool specimens. Journal of Clinical Microbiology 27:1155-1158. Zahradnik JM, Spencer MJ, Porter DD (1980):Adenovirus infection in the immunocompromised patient. American Journal of Medicine 68:725-732.