A Novel Real-Time PCR Method for Determination and Quantification of Each Cytomegalovirus Glycoprotein H Subtype in Clinical Samples

Abstract

To investigate reinfection in patients with congenital cytomegalovirus (CMV) infection, we established a CMV subtype-specific real-time quantitative PCR method targeting the CMV gH epitope region that can be used for evaluating pathogenic CMV strains in cases of mixed CMV infection.

TEXT

Cytomegalovirus (CMV) is a causative pathogen of congenital infection that leads to multiple clinical defects, not only at birth but also during the early stages of development (6). Primary infection during pregnancy is a significant risk factor for intrauterine CMV infection. Approximately 30% of pregnant women are CMV seronegative, and 1.1% of them experienced primary infection during pregnancy (10, 13, 14, 18). Compared with primary infection, the chance of CMV reinfection or reactivation is said to be much lower; however, the possibility still exists in neonates born to women who were seropositive prior to pregnancy (1–4, 11, 12, 15, 16).

Frequency of reinfection differs greatly according to race, age, and infectious disease (7, 15, 17), and reinfection in CMV-seropositive individuals is, therefore, not a rare event. Although there are many CMV strains, immunity against CMV subtype-specific neutralizing epitopes of glycoprotein H (gH) is acquired when CMV-seropositive individuals are infected with a previously unexposed strain (2, 19). Serology by gH offers a clue to understanding immunity against CMV (2, 8). However, the detection of the pathogen itself, not just its antibody, is required for the identification of the pathological strain in cases of CMV infection. We therefore established a novel quantitative PCR (qPCR) method that enables the detailed clarification of CMV reinfection by distinguishing between and quantifying CMV subtypes by gH epitope.

Seventy-three of 23,757 neonates were diagnosed with congenital CMV infection by the Japanese Congenital Cytomegalovirus Study Group (9). CMV genomic DNA was isolated from urine samples in seven of the 73 cases by using a QIAamp DNA minikit (Qiagen, Valencia, CA) according to the manufacturer's protocol.

CMV-specific IgG in blood from neonates was tested by a commercial kit (Enzygnost anti-CMV IgG; Dade Behring Marburg GmbH, Germany). CMV subtype-specific CMV IgG was tested by enzyme-linked immunosorbent assay (ELISA) as described previously (8).

Quantitative PCR for CMV DNA was carried out with a TaqMan probe as previously reported (5). For CMV subtype-specific quantification and determination, the gH region (amino acid positions 27 to 52) containing an epitope (positions 34 to 43) was targeted for qPCR testing (2, 19). Primers to amplify the region were designed based on a consensus sequence between the AD169 and Towne strains, with gH78-97 as the forward primer and gH157-137 as the reverse primer (Fig. 1). An amplified PCR product of the gH region from each strain was cloned into the pGEM-T Easy vector (Promega, Madison, WI) and transformed in the Escherichia coli DH5α strain (Wako Pure Chemical Industries, Ltd., Osaka, Japan). The plasmid DNA was then harvested by the HiSpeed plasmid midikit (Qiagen) as the standard DNA for the CMV subtype-specific qPCR. Probes to distinguish each strain were designed based on the uniquely sequenced portion of the gH epitope region (positions 34 to 39). The qPCR was performed using a StepOnePlus real-time PCR system (Applied Biosystems, Foster City, CA). We carried out nested PCR for samples with lower CMV copy numbers. The initial PCR was performed prior to qPCR using primers gH21-38 (5′-CTA CCT CAY CGT CYT CR-3′) and gH600-580 (5′-ATC AAA GAG GAT ACA GGT CTG-3′) with initialization at 95°C for 5 min, followed by 30 cycles of 95°C for 30 s, 50°C for 1 min, and 72°C for 1 min and a final extension at 72°C for 5 min.

Fig 1.

Sequence of the HCMV gH region. (A) Sequences of gH in CMV were compared between 19 isolates. Twelve sequences were from GenBank, accession numbers NC_001347, AB275151.1, AB275155.1, AB275152.1, AB275157.1, AB275158.1, AB275153.1, U53565.1, FJ616285.1, AB275156.1, AB275159.1, and AB275154.1. (B) Primers and probes for CMV subtype-specific real-time PCR. R is A or G; M is A or C; Y is C or T; S is G or C; D is G, A, or T.

PCR products from case numbers 22383 and 20117 (see Table 2) were cloned into a pGEM-T Easy vector (Promega), and CMV subtypes were investigated in each clone by qPCR analysis.

Table 2.

Determination of CMV strains in neonates with congenital CMV infection^a

ID	CMV DNA in urine (no. of copies/μl)			CMV gH sequencing	Presence of CMV IgG antibody
	CMV	AD	To		CMV	AD	To
20026	2.4 × 105	6.6 × 104	ND	AD	+	+	–
20040	6.6 × 103	7.2 × 102	ND	AD	+	+	–
20270	2.3 × 103	1.6 × 102	ND	AD	+	+	–
22383	1.3 × 105	3.1 × 104	ND	AD	+	+	+
20117	2.2 × 101	ND	P	To	+	–	+
18189	2.4 × 104	ND	1.4 × 104	To	+	–	+
19382	1.5 × 105	ND	1.1 × 105	To	+	–	+

^aP, positive by nested PCR; ND, not determined.

The nucleotide sequences of the CMV gH gene from 7 clinical isolates were determined by fluorescent dye-terminator sequencing. Seminested PCR was performed using primers gH21-38 and gH600-580, followed by primers gH69-85 (5′-ACG ATA TGG CGC AGA MG-3′) and gH600-580. PCR products were labeled using a DYEnamic ET terminator cycle sequencing kit (GE Healthcare) and read on the ABI PRISM 3100 Genetic Analyzer (Applied Biosystems). CMV strain was determined by sequence alignment with data from AD169 and Towne strains (GenBank accession numbers NC_001347 and FJ616285.1, respectively).

Sensitivity and specificity of the CMV subtype-specific qPCR were tested using standard plasmids containing the gH region from the AD169 or Towne strain. The AD169 or Towne plasmid DNA was detected from 10⁸ to 10² copies/reaction with 6-carboxyfluorescein (FAM)-labeled AD or Vic-labeled To probes, respectively. There were no nonspecific detections; i.e., no AD169 or Towne plasmid DNA was detected (at 10⁸ to 10¹ copies/reaction) with the Vic-labeled To or FAM-labeled AD probes, respectively. The sensitivity was increased sufficiently to detect 10¹ copies/reaction by nested PCR.

The sensitivity of the CMV subtype-specific qPCR was also tested by mixing both types of standard plasmid DNA at different ratios to simulate clinical isolates from cases with mixed CMV infection (Table 1). The target plasmid DNA was serially 10-fold diluted from 10⁷ to 10² copies/reaction, and the other plasmid DNA was mixed in the reaction at 10⁷, 10⁵, or 10³ copies/reaction. The signals from the qPCR were then detected with both AD (FAM) and To (Vic) probes as a multiplex qPCR. The results were no longer close to the copy numbers we added nor detectable when the counterpart plasmid DNA exceeded the target 100-fold; for example, it was impossible to detect 10⁴ copies/reaction of AD169 plasmid DNA under contamination with 10⁷ copies/reaction of Towne plasmid DNA (Table 1). A level of 10³ copies/reaction of AD169 plasmid DNA became detectable and closer to the expected copy numbers when the contaminated Towne plasmid DNA was decreased to 10⁵ copies/reaction. The results were similar when detecting Towne plasmid DNA by the To (Vic) probe under contamination with AD169 plasmid DNA (Table 1). The relationship between the expected copy numbers (added copy numbers/reaction) and measurements obtained from the qPCR was positively correlated when the copy numbers of the mixed constituents were up to 100-fold that of the plasmid DNA on which the detection was focused.

Table 1.

Sensitivity of HCMV strain-specific real-time quantitative PCR

Expected no. of copies	No. of mixed copiesa
	107	105	103
AD169
107	1.08 (±0.07) × 107	1.07 (±0.01) × 107	1.23 (±0.09) × 107
106	0.95 (±0.01) × 106	1.12 (±0.02) × 106	1.19 (±0.15) × 106
105	0.42 (±0.01) × 105	1.34 (±0.01) × 105	0.90 (±0.04) × 105
104	ND	1.13 (±0.09) × 104	1.24 (±0.07) × 104
103	ND	0.65 (±0.13) × 103	1.08 (±0.12) × 103
102	ND	ND	1.33 (±0.21) × 102
Towne
107	0.83 (±0.01) × 107	1.04 (±0.04) × 107	1.00 (±0.01) × 107
106	3.73 (±0.01) × 106	1.02 (±0.18) × 106	1.10 (±0.21) × 106
105	4.78 (±0.61) × 103	1.08 (±0.24) × 105	0.80 (±0.12) × 105
104	ND	0.58 (±0.04) × 104	0.77 (±0.05) × 104
103	ND	ND	0.71 (±0.08) × 103
102	ND	ND	1.27 (±0.31) × 102

^aND, not detected.

DNA in urine from seven neonates with congenital CMV infection was provided to evaluate the performance of the method. All seven neonates had IgG antibodies against CMV; three against AD169, three against Towne, and one against both strains (Table 2). Their mothers showed the same CMV IgG antibody patterns as the neonates, indicating that the predominant IgG antibodies in the neonates were the maternal antibodies (data not shown). Direct sequencing revealed that four neonates shed AD169-type and three shed Towne-type CMV in their urine, and these results matched the serological results. Results from the CMV subtype-specific qPCR were consistent with the direct sequencing; only the AD169-type or Towne-type strain was detected from neonates who had AD169-type or Towne-type strains by direct sequencing from the urine. Case number 22383, who had antibodies against both CMV strains, shed only AD169-type CMV strain in his urine. We could not find any cases in which both subtypes where shed in the urine, even though one of the neonates had IgG against both AD169-type and Towne-type strains. We tried to find multiple CMV infections in urine by cloning PCR products. All the clones (308 clones from urine, blood, or breast milk from case number 22383 and 43 clones from urine from case number 20117) were screened by subtype-specific qPCR, and only a single subtype was detected from each individual (data not shown). In agreement with our results, CMV gH subtype (as assessed by antibody assay) was previously reported to match that in the individual's urine (2). In that report, only a single subtype was shed in the urine despite having CMV IgG against both AD169 and Towne subtypes. A previous study also reported that the analysis of the genotype of gB, gH, gN, UL144, and UL149 CMV genes revealed no evidence of infection with multiple strains (20).

In conclusion, our established subtype-specific qPCR is a highly effective method for evaluating multiple CMV infections that escaped neutralization.