Whole-genome evolutionary dynamics of human parainfluenza virus type 3 in Shanghai, China, 2016–2024

Highlights

• •Fifty whole-genome sequencing revealed co-circulating HPIV-3 C3 sub-lineages C3f and C3a in Shanghai, China.
• •Whole-genome phylogeny dated the HPIV-3 tMRCA to ∼1925.6 and revealed two post-1990 demographic expansions.
• •Recombination signals detected in the HN gene and other regions may lead to discordance in partial-gene phylogenies.
• •The L gene showed the highest variability and harbored the largest number of putative positively selected sites.

Dear Editor,

Human parainfluenza virus (HPIV) is a major viral pathogen causing acute respiratory infections (ARI) worldwide, particularly in young children. Based on genetic and antigenic variability, HPIV is classified into four serotypes (HPIV-1 to HPIV-4), among which HPIV-3 is the most frequently detected and is strongly associated with severe lower respiratory tract infections, especially pneumonia and bronchiolitis in children (Han et al., 2022). The HPIV-3 genome is approximately 15,462 nt in length, encoding six structural proteins in the order of nucleoprotein (N), phosphoprotein (P), matrix (M), fusion (F), hemagglutinin-neuraminidase (HN) and large RNA-dependent RNA polymerase (L) from the 3′ to 5′ end (Lefkowitz et al., 2018). Despite its substantial disease burden, no licensed vaccines or antiviral therapies are currently available for clinical use (Russell et al., 2018). Genomic surveillance is therefore crucial for understanding viral evolution and informing future prevention and control strategies. Nevertheless, molecular epidemiological studies of HPIV-3 have largely relied on partial sequencing of the hemagglutinin-neuraminidase (HN) gene (Aso et al., 2020). Whether such single-gene analyses fully capture viral evolutionary dynamics remains uncertain. Whole-genome sequencing (WGS) is an indispensable tool for accurate lineage classification, reconstruction of evolutionary history, and long-term surveillance. However, genome-wide studies of HPIV-3 remain limited globally, particularly in China.

In this letter, we performed whole-genome sequencing and evolutionary analyses on HPIV-3 strains obtained from routine acute respiratory infection (ARI) surveillance in Shanghai between 2016 and 2024. A total of 50 HPIV-3-positive clinical samples were successfully sequenced using a one-step tiled-amplicon strategy combined with Illumina sequencing (primers listed in Supplementary Table S1). The genome lengths ranged from 15,146 to 15,254 bp (mean: 15,189 bp), and the GC content ranged from 35.11% to 35.36% (mean: 35.25%). The clinical and demographic characteristics are summarized in Table 1. Among the 50 HPIV-3 positive clinical samples, the median age of the patients was 4 years (range: 1–84 years), and 52% of cases occurred in children under 5 years of age. The samples were collected from 2016 to 2024, with a peak in 2024, followed by 2017. High-quality genomes covering more than 98% of the reference genome were obtained and analyzed together with representative global HPIV-3 genomes retrieved from GenBank (Supplementary Methods and Table S2). The nucleotide identity among the sequences generated in this study ranged from 97.2% to 99.8%, and the corresponding amino acid identity ranged from 94.6% to 99.7%. The nucleotide and amino acid identities ranged from 94.7% to 99.8% and 89.2%–99.7% compared with the reference strain (Human respirovirus 3 C243/ATCC VR-93 RNA/LC817393.1), respectively.

Table 1.

Basic clinical characteristics of HPIV-3-positive samples in this study.

Genbank ID	Sample name	Ct value	Post-amplification concentration (ng/μL)	Coverage (%)	Average coverage depth ( × )	Genome	Genotype	Gender	Age	Collection date	Clinical symptom	Clinical diagnose
PX245585	HPIV-3/Shanghai/China/2016/01	26.98	>60	99	2366.53	WGS	C3f	Female	1	2016/3/29	Fever, cough, shortness of breath	Pneumonia
PX245583	HPIV-3/Shanghai/China/2016/02	25.78	58	99	2310.16	WGS	C3f	Female	1	2016/2/14	Fever, cough, headache, fatigue	Pneumonia
PX245479	HPIV-3/Shanghai/China/2016/03	27.60	53	98	346.32	WGS	C3a	Female	1	2016/5/5	Cough, runny nose	Pneumonia
PX132646	HPIV-3/Shanghai/China/2017/01	18.42	>60	98	4952.91	WGS	C3a	Male	76	2017/5/22	Fever, cough	Pneumonia
PX132640	HPIV-3/Shanghai/China/2017/02	23.57	>60	99	3948.43	WGS	C3a	Male	1	2017/5/10	Cough	Pneumonia
PX132641	HPIV-3/Shanghai/China/2017/03	20.05	>60	99	1572.75	WGS	C3a	Male	1	2017/4/23	Fever, cough, runny nose, expectoration	Bronchitis
PX122091	HPIV-3/Shanghai/China/2017/04	22.76	>60	98	5752.52	WGS	C3a	Female	1	2017/4/7	Fever, cough, expectoration	Pneumonia
PX236538	HPIV-3/Shanghai/China/2017/05	19.64	>60	100	8054.14	WGS	C3b	Male	1	2017/3/8	Cough	Pneumonia
PX245581	HPIV-3/Shanghai/China/2017/06	17.55	>60	99	9101.81	WGS	C3f	Male	71	2017/4/7	Cough, expectoration, shortness of breath	AECOPD
PX245578	HPIV-3/Shanghai/China/2017/07	26.10	>60	99	8383.59	WGS	C3f	Female	5	2017/4/8	Fever, cough, runny nose	Pneumonia
PX236530	HPIV-3/Shanghai/China/2017/08	26.73	>60	99	1636.93	WGS	C3a	Male	2	2017/4/13	Fever, cough, runny nose	Pneumonia
PX245589	HPIV-3/Shanghai/China/2017/09	24.53	56	99	180.67	WGS	C3f	Male	5	2017/10/26	Cough	Pneumonia
PX242663	HPIV-3/Shanghai/China/2018/01	19.35	>60	98	720.21	WGS	C3b	Male	1	2018/5/17	Cough, shortness of breath	Pneumonia
PX236532	HPIV-3/Shanghai/China/2018/02	24.70	50	98	1173.80	WGS	C3a	Male	1	2018/6/5	Cough, runny nose	Pneumonia
PX236539	HPIV-3/Shanghai/China/2018/03	25.22	>60	99	2485.75	WGS	C3b	Female	7	2018/8/9	Fever, runny nose, sore throat	URTI
PX245582	HPIV-3/Shanghai/China/2018/04	21.27	>60	98	2508.48	WGS	C3f	Female	46	2017/8/18	Fever, cough, sore throat, expectoration	URTI
PX245577	HPIV-3/Shanghai/China/2018/05	19.96	>60	98	2951.02	WGS	C3f	Male	1	2018/5/18	Fever, cough, expectoration	Pneumonia
PX132638	HPIV-3/Shanghai/China/2018/06	20.28	>60	99	3334.84	WGS	C3a	Male	3	2018/5/10	Fever, cough, sore throat	URTI
PX132639	HPIV-3/Shanghai/China/2018/07	28.34	>60	98	2694.88	WGS	C3a	Female	1	2018/4/12	Cough, sore throat, expectoration	Pneumonia
PX236534	HPIV-3/Shanghai/China/2018/08	23.18	>60	99	1700.90	WGS	C3a	Male	1	2018/5/9	Cough, runny nose	URTI
PX132643	HPIV-3/Shanghai/China/2018/09	21.60	>60	99	1772.15	WGS	C3a	Male	3	2018/5/10	Fever, cough, sore throat	URTI
PX245576	HPIV-3/Shanghai/China/2019/01	22.32	>60	99	2811.97	WGS	C3f	Female	3	2019/8/29	Fever, cough, expectoration	Pneumonia
PX245587	HPIV-3/Shanghai/China/2019/02	27.91	>60	98	4245.31	WGS	C3f	Female	3	2019/6/20	Fever, cough, runny nose, sore throat	URTI
PX132644	HPIV-3/Shanghai/China/2019/03	26.72	>60	98	4965.06	WGS	C3a	Female	1	2019/6/20	Fever	URTI
PX245580	HPIV-3/Shanghai/China/2019/04	14.14	>60	99	2518.95	WGS	C3f	Female	81	2019/6/17	Fever, fatigue	URTI
PX236533	HPIV-3/Shanghai/China/2019/05	22.39	>60	98	7477.58	WGS	C3a	Male	1	2019/7/4	Fever, cough, sore throat, expectoration	URTI
PX132642	HPIV-3/Shanghai/China/2019/06	26.82	>60	98	4625.65	WGS	C3a	Male	30	2019/7/18	Fever, cough	URTI
PX242662	HPIV-3/Shanghai/China/2019/07	23.89	>60	99	5114.03	WGS	C3b	Male	3	2019/6/5	Fever, cough	URTI
PX236531	HPIV-3/Shanghai/China/2019/08	26.25	>60	98	1304.25	WGS	C3a	Female	30	2019/9/18	Fever, sore throat	URTI
PX242664	HPIV-3/Shanghai/China/2020/01	20.69	>60	98	4563.99	WGS	C3b	Male	68	2020/12/8	Fever, cough, sore throat	URTI
PX245579	HPIV-3/Shanghai/China/2021/01	26.70	>60	98	1120.40	WGS	C3f	Female	9	2021/10/27	Fever, cough	URTI
PX236536	HPIV-3/Shanghai/China/2023/01	23.12	>60	99	4667.19	WGS	C3a	Female	62	2023/10/25	/	URTI
PX242665	HPIV-3/Shanghai/China/2023/02	21.27	>60	99	3242.22	WGS	C3f	Male	20	2023/11/8	Fever, cough	URTI
PX245477	HPIV-3/Shanghai/China/2023/03	24.68	>60	98	1365.90	WGS	C3f	Male	66	2023/11/1	Cough, expectoration	URTI
PX245575	HPIV-3/Shanghai/China/2024/01	23.95	>60	98	5119.27	WGS	C3f	Male	3	2024/7/3	Fever, cough	Pneumonia
PX132645	HPIV-3/Shanghai/China/2024/02	21.13	>60	99	8016.55	WGS	C3a	Male	1	2024/7/10	Fever, cough	Pneumonia
PX132647	HPIV-3/Shanghai/China/2024/03	26.09	>60	98	5277.31	WGS	C3a	Female	61	2024/7/11	Fever, cough, runny nose, sore throat, expectoration, dyspnea, headache	Pneumonia
PX245586	HPIV-3/Shanghai/China/2024/04	17.81	>60	99	8957.91	WGS	C3f	Female	64	2024/7/18	Fever, cough, runny nose, sore throat, headache, fatigue	Infectious fever
PX245593	HPIV-3/Shanghai/China/2024/05	17.8	>60	98	3294.49	WGS	C3f	Male	68	2024/5/8	Cough, expectoration	Pneumonia
PX245584	HPIV-3/Shanghai/China/2024/06	23.81	>60	98	1727.05	WGS	C3f	Male	84	2024/8/14	Cough, runny nose, expectoration, sore throat	URTI
PX236535	HPIV-3/Shanghai/China/2024/07	24.93	>60	99	1086.25	WGS	C3a	Female	57	2024/4/25	Cough, runny nose, expectoration	Infectious fever
PX245591	HPIV-3/Shanghai/China/2024/08	28.27	>60	99	1021.11	WGS	C3f	Female	9	2024/4/29	Fever, cough, runny nose, expectoration	Bronchitis
PX245478	HPIV-3/Shanghai/China/2024/09	35.72	>60	98	2161.00	WGS	C3f	Female	46	2024/4/22	Cough	Pneumonia
PX245588	HPIV-3/Shanghai/China/2024/10	34.73	>60	98	84.41	WGS	C3f	Female	67	2024/4/23	/	Bronchitis
PX245592	HPIV-3/Shanghai/China/2024/11	29.58	48	98	191.50	WGS	C3f	Male	4	2024/6/5	Fever, cough, shortness of breath	Pneumonia
PX245480	HPIV-3/Shanghai/China/2024/12	27.71	44	98	1767.50	WGS	C3f	Female	1	2024/6/12	Fever, cough, runny nose, expectoration	Pneumonia
PX242666	HPIV-3/Shanghai/China/2024/13	27.25	48	99	1940.00	WGS	C3f	Male	1	2024/6/27	Fever, cough	Infectious fever
PX245590	HPIV-3/Shanghai/China/2024/14	31.52	46	98	1462.00	WGS	C3a	Female	30	2024/7/3	Fever, cough, runny nose	Pneumonia
PX236537	HPIV-3/Shanghai/China/2024/15	28.74	66	98	1775.50	WGS	C3a	Female	4	2024/8/1	Fever, cough	Infectious fever
PX245476	HPIV-3/Shanghai/China/2024/16	21.23	52	99	2511.12	WGS	C3f	Female	55	2024/8/1	Cough, expectoration	Bronchitis

WGS: Whole-genome sequencing; URTI: upper respiratory tract infection; AECOPD: acute exacerbation of chronic obstructive pulmonary disease.

Although the HN gene has been widely used for phylogenetic classification of HPIV-3 because of its antigenic relevance and sequence variability (Aso et al., 2020; Elusah et al., 2020), different genomic regions, including the F gene (Godoy et al., 2016) and L gene (Guo et al., 2024), have also been used for lineage classification, reflecting the lack of a universally standardized genotyping framework. In our dataset, the comparison of phylogenetic classifications based on whole-genome sequences and the HN gene revealed discrepancies in lineage assignment for certain clinical isolates, which showed topological incongruence between the two phylogenies (Supplementary Fig. S1). This finding further supports the use of whole-genome analysis for improved phylogenetic resolution compared with single-gene approaches. Maximum-likelihood and Bayesian phylogenetic analyses based on whole-genome sequences consistently classified all Shanghai strains within genotype C3. Three sub-lineages were identified, including C3f (n = 24), C3a (n = 21), and C3b (n = 5), indicating co-circulation of multiple lineages within a single metropolitan region (Fig. 1A). This pattern is consistent with findings from a recent study in China (Guo et al., 2025). The simultaneous circulation of multiple sub-lineages within one city highlights the dynamic nature of local HPIV-3 populations. Phylogenetically, C3a strains clustered more closely with strains from the United States, whereas C3f strains were more closely related to Vietnamese strains, suggesting shared ancestry or cross-regional transmission.

Fig. 1.

Whole-genome evolutionary dynamics and diversity of HPIV-3. A Time-calibrated Bayesian Markov chain Monte Carlo (MCMC) phylogenetic tree and Bayesian skyline plot (BSP) shown in the inset (black box) are based on whole-genome sequences of HPIV-3. The root node represents the estimated time to the most recent common ancestor (tMRCA) of all analyzed HPIV-3 sequences, whereas the labeled internal nodes indicate lineage-specific divergence times between corresponding branches. B Entropy plot of nucleotide positions across the HPIV-3 whole-genome sequences. C Entropy plot of amino acid positions. For each nucleotide or amino acid position, higher entropy values indicate greater sequence variability. Abbreviated protein names are shown at the top of the plot in the genomic order of coding sequences.

To further investigate the potential causes of these phylogenetic discrepancies, we performed recombination analysis across the whole-genome sequences. Recombination analysis identified potential recombination events in 5 of the 50 sequenced strains (10.0%). Notably, HPIV-3/China/Shanghai/2019/03 exhibited discordant clustering between HN-based and whole-genome phylogenies (Supplementary Fig. S1) and showed seven potential recombination events involving sub-lineages A, C3b, and C3f. Two recombination events were located within the HN gene region (Supplementary Fig. S2), whereas the remaining events were identified in the N, M, and L genes. These findings are consistent with the observed phylogenetic incongruence and suggest that recombination may contribute to discrepancies between HN-based and whole-genome phylogenetic analyses. Another strain also exhibited potential recombination signals primarily within the L gene region, further supporting a role for recombination in shaping HPIV-3 genomic diversity. These findings are consistent with previous evidence that recombination occurs naturally in HPIV-3 (Bose et al., 2019; Shao et al., 2021). In the context of earlier studies that predominantly relied on HN-based analyses, such mosaic patterns may partly explain inconsistencies between HN-based phylogenetic inferences and those derived from whole-genome data, further underscoring the necessity of whole-genome sequencing for robust evolutionary reconstruction.

Understanding the divergence time of HPIV-3 lineages is crucial for reconstructing its evolutionary history. Time-scaled phylogenetic reconstruction estimated the time to the most recent common ancestor (tMRCA) of HPIV-3 to be approximately 1925.6 (95% HPD: 1894.8–1950.9), which is earlier than estimates derived from partial HN gene datasets (Aso et al., 2020). This discrepancy may reflect the limited evolutionary information captured by partial gene datasets and suggests that single-gene analyses could underestimate the evolutionary depth of HPIV-3. The mean nucleotide substitution rate was 5.63 × 10⁻⁴ substitutions/site/year, which was also lower than those reported in several countries during 2003–2011 (Bose et al., 2019), potentially reflecting regional transmission dynamics or differences in selective pressures. Bayesian skyline analysis indicated a relatively stable effective population size prior to 1990, followed by two phases of sustained population expansion thereafter, coinciding with the diversification of C3a, C3f, and C3e sub-lineages (Fig. 1A). At the lineage level, whole-genome analysis further suggested that Cluster C emerged around 1950 following divergence from Cluster B, contrasting with earlier studies reporting largely constant population sizes between 1950 and 1998 (Takahashi et al., 2018). These findings underscore the improved temporal resolution afforded by whole-genome data in reconstructing HPIV-3 population dynamics and suggest possible influences of changing host immunity or ecological factors in recent decades.

Beyond population-level evolutionary dynamics, characterization of genome-wide variability may provide additional insights into the molecular basis of HPIV-3 diversification and adaptation (Greenbaum and Ghedin, 2015). Genome-wide entropy analysis demonstrated uneven distribution of sequence variability across the genome. At the nucleotide level, 199 hypervariable sites with entropy values greater than 0.6 were identified. Among these, the L gene coding region (nt 8638–15339) harbored the highest number of variable sites (n = 68), accounting for 34.17% of all identified positions. This pattern likely reflects, at least in part, the large size of the L gene and its multifunctional role within the viral RNA-dependent RNA polymerase complex. Notably, only a single nucleotide site exhibited high entropy (> 1.0), which was located within the F gene coding region (Fig. 1B). Given the central role of the F protein in mediating viral membrane fusion, this finding may reflect localized mutational hotspots within structurally or functionally redundant regions that tolerate sequence variation without compromising fusion activity (Stearns et al., 2024). Amino acid sequences exhibited higher homology compared to nucleotide sequences. Most amino acid variable sites (entropy > 0.6) were again found within the L gene (65 sites, 51.18%), mirroring the nucleotide-level pattern. The P and HN genes remained relatively conserved at the amino acid level, consistent with previous reports (Bose et al., 2019). This observation suggested that strong purifying selection acted to preserve protein function and further indicated that the P gene may represent a relatively conserved molecular target for future assay development. Only one highly variable site (entropy > 1.0) was identified in a non-coding genomic region between the HN and L genes (Fig. 1C). Non-coding regions are generally subject to weaker selective constraints than coding regions and may therefore accumulate substitutions more readily. Such variability may contribute to fine-tuning viral replication or transcription efficiency while preserving the structural and functional integrity of essential viral proteins, although this hypothesis warrants further investigation.

To further assess adaptive evolution, we performed codon-level selection analyses across the entire HPIV-3 genome. Among the four complementary algorithms applied (Supplementary Fig. S3), MEME identified the largest number of positively selected sites, whereas FUBAR detected the fewest, highlighting the value of integrating multiple approaches to capture both pervasive and episodic selection. Collectively, the L gene harbored the greatest number of putative positively selected sites, followed by the P and N genes. As the L protein functions as the viral RNA-dependent RNA polymerase and plays a central role in genome replication and transcription (Matsumoto et al., 2015), the enrichment of positively selected sites within L gene may reflect heterogeneous adaptive pressures acting on the viral replication machinery during HPIV-3 evolution. In the HN gene, MEME, FEL, and SLAC identified several positively selected codons in the alignment, whereas no positively selected sites were detected by FUBAR. Given the role of the HN protein in receptor binding and viral entry, these sites may be associated with host-driven selective pressures. Overall, these patterns suggest that distinct viral proteins may experience heterogeneous evolutionary pressures during HPIV-3 evolution. The biological significance of these putative positively selected sites remains unclear and requires further functional investigation to determine their potential effects on viral replication, fitness, or host interactions.

In summary, we highlighted the limitations of single-gene analyses and demonstrated the added resolution provided by whole-genome sequencing for reconstructing HPIV-3 evolutionary history. However, it should be noted that the implementation of whole-genome sequencing in routine surveillance remains challenging due to factors such as higher costs, sample requirements for sufficient viral load and quality, and the need for advanced bioinformatics capacity. Based on our findings, a tiered surveillance strategy may be feasible in practice, in which routine detection in primary clinical settings relies on targeted assays or conserved gene regions (e.g., HN gene or P gene), while whole-genome sequencing is reserved for representative samples, outbreak investigations, or cases with unusual clinical or epidemiological features. Such a combined approach may balance feasibility and resolution in large-scale surveillance. Overall, long-term whole-genome surveillance, when feasible, will be critical for monitoring viral diversity, recombination, and evolutionary dynamics, and for informing future strategies for ARI prevention and control.

Footnotes

This work was supported by the Shanghai Municipal Three-year Action Plan for Strengthening the Construction of the Public Health System (2023–2025) and the Key Subject (grant no. GWVI-11.1-02-Infectious Diseases); Young Medical Talents Training Program of Shanghai Pudong New Area Health Commission (PWRq2025-03). This study was approved by the ethical review board of Shanghai Pudong New Area Center for Disease Control and Prevention (Record number: PDCDCLL-20241024-001). The authors declared no conflict of interest. The nucleotide sequences were deposited in the GenBank database under accession numbers PX122091, PX132638–PX132647, PX236530–PX236539, PX242662–PX242666, PX245476–PX245480, PX245575–PX245593, and in Science Data Bank with doi: 10.57760/sciencedb.36547. Supplementary data to this article can be found online at https://doi.org/10.1016/j.virs.2026.07.005.

The following are the Supplementary data to this article:

Supplementary Figure S1.

Phylogenetic trees constructed based on (A) whole-genome sequences and (B) HN gene sequences of HPIV-3 strains. Red squares indicate strains with inconsistent clustering patterns between the two trees.

Supplementary Figure S2.

Similarity and bootscan analyses of the putative recombinant HPIV-3 strain HPIV-3/China/Shanghai/2019/03. Analyses were performed using the whole-genome sequence of strain HPIV-3/China/Shanghai/2019/03, which exhibited discordant lineage assignments between HN-based and whole-genome phylogenies, as the query sequence. A sliding window of 200 bp with a step size of 20 bp was applied. Black arrows indicate putative recombinant regions.

Supplementary Figure S3.

Amino acid sites under selection pressure in HPIV-3 identified in this study. LRT: Likelihood Ratio Test. Positive selection was defined as dN − dS > 0 with P < 0.05 for SLAC and FEL, P < 0.001 for MEME, and posterior probability >0.95 for FUBAR.