Nasopharyngeal aspirates were obtained from children enrolled in a vaccine efficacy study that was conducted in Soweto, South Africa (22). The samples were obtained from children hospitalized for a lower respiratory tract infection in a 3-year period (2000–2002). Samples were stored at –70°C until processed for this study. Details regarding the cohort of children, procedure for collecting nasopharyngeal aspirate samples, and other viruses isolated from these samples have been published in part (22,23). The samples used in this study were from the entire year and not confined to samples obtained from the winter-spring months. Because of resource constraints, we sequenced a minimum of 30% of the hMPV-positive samples from each month; 92 (45%) of 206 hMPV-positive samples were sequenced for the F gene, and 61 (30%) of 206 were sequenced for the G gene. All samples that were sequenced for the G gene were sequenced for the F gene. Viral RNA was isolated from the stored frozen nasopharyngeal aspirate samples by using the QIAamp viral RNA kit (Qiagen, Inc., Valencia, CA, USA) according to the manufacturer's instructions. The study was approved by the Committee for Research on Human Subjects at the University of the Witwatersrand, South Africa.

A nested reverse transcription–polymerase chain reaction (RT-PCR) assay to amplify a fragment of the hMPV F gene was used to detect hMPV. RT-PCR was performed with the SUPERSCRIPT One-Step RT-PCR kit (Invitrogen, Carlsbad, CA, USA) with primers 5´-ATGTCTTGGAAAGTGGTG-3´ (corresponding to nucleotide position 3052–3069 in the NL/1/00 genome accession no. AF371337) and 5´- CCATGTAAATTACGGAGCT-3´ (nucleotide position 3844–3862 in NL/1/00 in genome) under the following conditions: 50°C for 30 min; 94°C for 2 min; 94°C for 30 s, 45°C for 45 s, and 68°C for 1 min for 35 cycles; 68°C for 7 min.

The nested PCR was performed with primers 5´-TCATGTAGCACTATAACT-3´ (nucleotide position 3130–3149) and 5´-TCTTCTTACCATTGCAC-3´ (nucleotide position 3794–3810) under the following conditions: 94°C for 2 min; 94°C, 48°C, and 72°C for 1 min for 30 cycles; and 72° for 7 min. The PCR product was analyzed by electrophoresis on a 2% ethidium bromide–stained agarose gel.

The G gene open reading frame (ORF) was amplified with the following primers: HMPVGunivF: 5´-GAGAACATTCGRRCRATAGAYATG-3´ (nucleotide position 6262–6285 of NL/1/00, GenBank accession no. AF371337) and HMPVGunivR: 5´-AGATAGACATTRACAGTGGATTCA-3´ (nucleotide position 7181–7204) under the following conditions: 50°C for 30 min; 95°C for 3 min; 94°C for 1 min, 59°C for 1 min, and 72°C for 2 min for 38 cycles; and 72°C for 7 min. The PCR product was analyzed on a 2% ethidium bromide–stained agarose gel. When necessary to increase the yield for sequencing, a nested PCR was performed with the same primer set.

The PCR product generated for both F and G genes was purified with the QIAquick gel extraction kit (Qiagen, Inc.) and sequenced in both directions by using the nested primers for the F gene and the primers used for detecting the G gene. The PCR product was sequenced by using the BigDye Terminator Cycle sequencing kit (Applied Biosystems, Foster City, CA, USA) on the ABI 310 Genetic Analyzer (Applied Biosystems).

Nucleotide sequence alignments were generated with the ClustalX 1.81 software (24). Phylogenetic analysis was performed by using MEGA version (2.1) (25). Strains from the Netherlands, NL/1/00, NL/17/00, NL/1/99, and NL/1/94 (GenBank accession nos. AF371337, AY304360/AY296021, AY304361/AY296034, and AY304362/AY296060, respectively) and Canada, CAN97-83, hMPV13-00, CAN98-75, and hMPV33-01 (GenBank accession nos. AY485253/AY145296, AY485232, AY485245/AY145289, and AY485242, respectively) were used as prototypes of the 2 groups and subgroups. The hMPV sequences for the F and G protein genes presented in this article have been deposited in GenBank under the accession numbers AY694693–AY694784 and AY848859–AY848919, respectively.

We examined the circulation pattern of hMPV during the 3 years by sequencing part of the hMPV F gene and performing phylogenetic analysis for 92 hMPV-positive samples. These samples, numbering 40, 34, and 18 from years 2000, 2001, and 2002 respectively, were distributed over all the months when hMPV was identified and accounted for ≈30%–100% of the samples from each month. Phylogenetic analysis indicated 2 major groups (A and B), each divided into 2 subgroups (1 and 2), causing a complex circulation pattern over the course of the study. Both groups A and B viruses cocirculated throughout the study period. Of the 92 hMPV-positive samples that were categorized, 56 (60.9%) of the viruses belonged to group A and 36 (39.1%) to group B.

During the 2000 epidemic, subgroups B2 and A2 cocirculated, with 72.5% of the circulating viruses belonging to subgroup B2. In 2001 subgroups A1, A2, and B2 cocirculated. Subgroup B2 virus significantly declined (4 [11.8%] of 34) in 2001 compared to 2000 (29 [72.5%] of 40), p<0.0001. Subgroup A1 emerged as the dominant strain, causing 67.7% of infections, compared to 20.5% observed for subgroup A2 and 11.8% for subgroup B2. Subgroup A1 also dominated in 2002, causing 83.3% of infections, and cocirculated with the emergent subgroup B1.

Sixty-one (66.3%) of the 92 hMPV-positive samples sequenced for the F gene were further sequenced for the G gene (Figure). Phylogenetic analysis of both the F and G gene nucleotide sequences also showed 2 major genetic groups (A and B) that could be further divided into 2 subgroups (1 and 2). The existence of these 2 major genetic groups was strongly supported by bootstrap values (100% of bootstrap replicas into 2 major groups and 99%–100% of bootstrap replicas into 2 minor subgroups).

From the topology of the trees, subgroup B2 was the most divergent. Although the South African hMPV clustered with both Canadian and Netherlands prototypes, the South African subgroup A1 virus clustered more closely with the Canadian prototype.

From the topology of the tree and supported by strong bootstrap values (70%–100%), the subgroups can be further divided into genotypes. We attempted to group the G sequences of the 4 subgroups into genotypes by using the criteria previously described for RSV (26), in which sequences were arbitrarily considered a genotype if they clustered together with bootstrap values of 70% to 100% (internal nodes at the internal branches). When these criteria were used (only on South African isolates), subgroup A1 could be divided into 5 genotypes, subgroup A2 into 2 genotypes, B1 into 2 genotypes, and B2 into 6 possible genotypes.

The estimated nucleotide and amino acid identities showed a high percentage of identity for the F gene and more variability for the G gene (Table). The estimated identities for the F gene between the 2 major groups, A and B, were 83%–85% at the nucleotide level and 93.2%–95.8% at the amino acid level. In contrast, the G gene estimated identities were 45.1%–53.1% at the nucleotide level and 22.4%–27.6% at the amino acid level. There was also a higher percentage of identity between members of the same group (e.g., A1–A2) for the F gene than for the G gene (Table).

Amino acid alignments of hMPV F gene were compared to those of prototype isolates from the Netherlands and Canada (data not shown). Cysteine residues were conserved in all South African strains at positions 60 and 182. Group-specific amino acid residue at positions 122, 135, 139, 167, 175, and 233 differentiated between groups A and B. Further amino acid substitutions at various positions were exclusive to subgroups A1 (amino acids [aa] 61, 82, 143), A2 (aa 61, 143, 185), and subgroups B1 (aa 46, 143, 179) and B2 (aa 143).

The predicted G ORF amino acid alignments of selected South African strains with prototypes from the Netherlands and Canada are shown in Figure A1. Sequence variation due to nucleotide substitutions and insertions led to variable lengths in polypeptides, which ranged from 228 aa residues (subgroup A2) to 240 aa residues (subgroup B2). The hMPV G ORFs of subgroups A2 and B1 terminated at the TAA codon, whereas the subgroup B2 isolates terminated at the TAG codon. For both genetic groups A and B, a cysteine residue was present in the intracellular domain. In addition, group B isolates had a cysteine residue in the extracellular domain except in 2 isolates (RSA/71/00 and RSA/90/00).

The region of the predicted G ORF sequenced in this study had a high serine and threonine content (30.7%–34.9% for group A, 30.6%–36.6% for group B isolates). Proline content varied among the subgroups: 7.6%–9.0% for subgroup A2, 9.0%–9.9% for subgroup A1, 7.8%–8.7%for subgroup B1, and 3.7%–5.2%, the lowest content, for subgroup B2. Only 1 potential N-linked glycosylation site was conserved at the junction of the intracellular and transmembrane domains.