Enterovirus 71 Activates Plasmacytoid Dendritic Cell–Dependent PSGL-1 Binding Independent of Productive Infection

Enterovirus 71 (EV71) is the main cause of hand, foot, and mouth disease (1, 2). Although most children infected with EV71 experience mild symptoms, a small percentage of patients can develop severe complications in the CNS, sometimes leading to fatalities (1, 3). EV71 uses scavenger receptor class B, member 2 (SCARB2) as its cellular receptor for attaching to the cell surface and uncoating in the endosomal compartment (4, 5). Besides SCARB2, other molecules such as P-selectin glycoprotein ligand-1 (PSGL-1) (6) and heparan sulfate (HS) (7) may also assist in EV71 attachment to target cells (8). In addition to binding with EV71, PSGL-1 is involved in initiating caveola-dependent endocytosis of the virus (9).

The binding specificity of EV71 to PSGL-1 and HS is influenced by the positively charged lysine residues surrounding its 5-fold axes, particularly VP1-242K and VP1-244K (7, 10, 11). VP1-145, located near the 5-fold axis, plays a role in determining EV71’s binding specificity to PSGL-1. The negatively charged glutamic acid at VP1-145 (VP1-145E) interacts with VP1-244K, reducing the positive charge at the 5-fold axes and resulting in decreased affinity to HS and PSGL-1. Conversely, viruses with glycine (G), alanine (A), or glutamine (Q) at VP1-145 are rich in positively charged amino acids, facilitating binding with HS and PSGL-1. VP1-145 acts as a switch that regulates the virus’s binding to HS and PSGL-1 (10, 11). VP1-145E isolates were the most common (81%), whereas VP1-145 A/G/Q strains were less frequently found in clinical samples (10, 12, 13).

The molecular epidemiology analysis revealed that the VP1-145 polymorphism is associated with severe neurologic symptoms. VP1-145E strains are more frequently isolated from patients with mild disease, whereas VP1-145 A/G/Q strains are more frequently isolated from patients with severe symptoms (14–17). Recent in vitro studies have confirmed the link between the amino acid at VP1-145 and virulence, suggesting that the strain with E at VP1-145 is less virulent compared with A/G/Q strains (18). However, there are conflicting results regarding the role of VP1-145 variants in EV71 virulence and in vivo pathogenicity. EV71 VP1-145E strains have been observed to replicate in monkeys, whereas VP1-145G strains show limited viral replication and a tendency to revert to glutamic acid in vivo (12). This reduced replication of VP1-145G has also been noted in mice (19, 20). The in vivo disadvantage of VP1-145G virus replication has been linked to either heightened susceptibility to neutralizing Abs (21) or binding to HS on the surface of cells with low or no expression of SCARB2, which are resistant to EV71 infection (19).

Plasmacytoid dendritic cells (pDCs), a specialized subset of cells that produce type I IFN (IFN-I), play a critical role in the antiviral immune response (22, 23). Unlike other cell types, pDCs can generate IFN-I in a manner independent of productive infection (23). Various mechanisms have been identified through which EV71 inhibits IFN-I production in infected cells (24–29). Therefore, the productive infection-independent production of IFN-I by pDCs is essential for initiating antiviral immune responses. Nevertheless, the specific role of pDCs in EV71 infection remains unclear.

Our research has shown that SCARB2 is predominantly expressed in pDCs (30). Additionally, we have found high levels of PSGL-1 expression on pDCs. Furthermore, we have demonstrated that EV71 can replicate within pDCs and induce the production of IFN-α in a PSGL-1–dependent manner. Specifically, only EV71 strains capable of binding to PSGL-1 can infect and activate pDCs to produce IFN-α. Importantly, pDC activation for IFN-α production does not require productive infection, as even UV-inactivated virions can stimulate IFN-α production. pDCs with high PSGL-1 expression secrete IFN-α in response to virulent EV71 strains early in the virus entry process, effectively inhibiting viral replication. Based on these findings, we suggest that the level of pDC activation following EV71 infection could serve as an indicator of disease progression.

Wild-type EV71 (0804232Y) (31) and wild-type EV71 (FUYANG) (32) were provided by Dr. Honglin Xu from the National Vaccine and Serum Institute and Dr. Xiangxi Wang from the Institute of Biophysics, Chinese Academy of Science, respectively. Infectious cDNA clones were constructed using the 0804232Y strain as a template. Specifically, four fragments encoding EV71 nt 1–900 with a T7 promoter sequence, nt 848–3743, nt 3696–5808, and nt 5758–7405 linked with a T7 terminator were amplified from EV71 cDNA and fused into the pWSK29 vector using a previously described method (33). The primers used for constructing cDNA clones can be found in Supplemental Table I. Point mutant viruses were also generated based on infectious cDNA clones, following established procedures (10), with the overlapping PCR primers for mutant cDNA clones listed in Supplemental Table I. Additionally, an EV71-GFP virus was produced from the 0804232Y strain using the infectious clone technique, in accordance with prior publications (4, 33).

In terms of cell subset purification, PBMCs were isolated from the whole blood of healthy volunteers via density gradient centrifugation using Ficoll-Paque (TBD, Tianjin, China). pDCs and conventional DCs were isolated through lineage (CD3, CD14, CD16, CD19, CD56)⁻ magnetic sorting, followed by subsequent flow cytometric sorting based on live, lineage (CD3, CD14, CD16, CD20, CD56)⁻, HLA-DR⁺CD123⁺, cells or HLA-DR⁺CD11c⁺ cells to a purity of 95%. T cells, B cells, monocytes, and NK cells were isolated from PBMCs by flow cytometric sorting of live, CD3⁺, CD19⁺, CD14⁺, or CD56⁺ cells, respectively.

Surface expression was detected after blocking nonspecific binding with FcR-blocking solution (BioLegend) followed by staining the cells with Abs for 30 min at 4°C. For intracellular molecules, cells were fixed with 2% (v/v) paraformaldehyde in PBS, and then permeabilized with intracellular fixation and permeabilization buffer (eBioscience) according to the manufacturer’s protocol. Flow cytometry analysis was conducted on an LSRFortessa cell analyzer (BD Biosciences), and data were analyzed using Summit 5.2 software (Dako, Glostrup, Denmark) or FlowJo 10.8.1 software (BD Biosciences).

To locate SCARB2 in pDCs, purified pDCs were fixed with 2% (v/v) paraformaldehyde in PBS, permeabilized using intracellular fixation and permeabilization buffer (eBioscience), and then stained with FITC anti-human SCARB2 Ab for 30 min at 4°C. Subsequently, the cells were affixed to glass slides using a cytospin centrifuge. The slides were sealed using Vectashield mounting medium (Vector Laboratories) with 1.5 mg/ml DAPI (Sigma-Aldrich) and visualized under a confocal microscope (Olympus FV1000). Data were collected using a FV10-ASW 1.7 viewer (Olympus).

Freshly purified pDCs were seeded in 96-well round plates at a density of 1 to 10 × 10⁴ cells per well. They were then incubated with wild-type EV71, EV71-GFP, or point mutant viruses that were diluted in pDC culture medium. After 1 h of incubation, the cells were washed three times with PBS containing 2% FBS and cultured in pDC culture medium. For wild-type EV71 or point mutant virus infection, supernatants were collected at designated time points and stored at −80°C for further analysis. In the case of EV71-GFP infection, pDCs were fixed with 2% (v/v) paraformaldehyde in PBS at 24 h postinfection. The GFP-positive cells were then examined using a confocal microscope (Olympus FV1000) or a BD FACSCalibur flow cytometer (BD Biosciences).

Freshly purified pDCs were exposed to EV71 at 37°C. Following a 1-h incubation period, the cells underwent three washes with PBS containing 2% FBS before being cultured in growth medium. Culture supernatants were collected at 1 and 24 h postinfection. These supernatants were then used to infect 293-hSCARB2 cells in six-well plates for 1 h. Subsequently, the supernatants were removed, and agarose solution was added to each well to facilitate plaque formation. The plates were then incubated at 37°C until visible plaques appeared. The cells were washed with PBS, fixed with 2% paraformaldehyde for 15 min at room temperature, and the plaques were visualized by staining with 0.1% crystal violet (w/v).

pDCs were infected by EV71 as described above. RNA from supernatants was extracted using the Biospin virus RNA extraction kit (BioFlux) according to the manufacturer’s instructions. cDNAs were generated from viral RNA by using Moloney murine leukemia virus reverse transcriptase (M1705, Promega) according to the manufacturer’s instructions. Real-time quantitative PCR was performed using a Rotor-Gene Corbett 65H0 system (Corbett Life Science). The primers used for quantitative PCR were as follows: probe, 5′-CGGAACCGACTACTTTGGGTGTCCGT-3′; forward primer, 5′-TCCTCCGGCCCCTGA-3′; reverse primer, 5′-AATTGTCACCATAAGCAGCCA-3′ for viral RNA quantification (34).

To assess the impact of soluble SCARB2 (R&D Systems) or PSGL-1 fusion protein (R&D Systems), EV71 was first incubated with SCARB2-Fc or PSGL-1-Fc (1, 10 μg/ml) for 1 h at 37°C. The mixture was then added to purified pDCs for 1 h at 37°C, followed by washing and culturing in pDC culture medium. Additionally, the effect of Abs against SCARB2 or PSGL-1 was evaluated by preincubating purified pDCs with the respective Abs for 1 h at 37°C, then adding EV71 to the cells, followed by washing and culturing. Furthermore, the impact of the Ab to IFNAR was determined by preincubating purified pDCs with the Ab for 1 h at 37°C, then adding EV71 to the cells, followed by washing and culturing. Supernatants were collected at specified time points and stored at −80°C for further analysis.

pDCs were infected by EV71 as described above. ELISA was performed to detect IFN-α in the collected supernatants using a human IFN-α pan ELISA kit (3425-1H-20, Mabtech) according to the manufacturer’s instructions. Absorbance was determined at 450 nm.

PBMC subsets were washed twice with cold PBS and then lysed in radioimmunoprecipitation assay buffer (0.01 M Tris [pH 8.0], 0.14 M NaCl, 2 mM EDTA, 1% NaDOC, 0.1% SDS, 1% Triton X-100, 1 mM PMSF) supplemented with 10% protease inhibitor mixture (Roche). Subsequently, the lysates underwent centrifugation at 13,300 × g for 20 min at 4°C. The proteins were separated using 10% SDS-PAGE gels and transferred onto polyvinylidene difluoride membranes. To prevent nonspecific binding, the membranes were blocked with 3% BSA in PBS containing 0.1% Tween 20. Following this, the membranes were incubated overnight at 4°C with a goat Ab against SCARB2 (R&D Systems). Ab labeling was detected using HRP-conjugated secondary Abs (ZsBio) and visualized with Immobilon Western HRP substrate (Millipore).

All graphs in this study were produced, and statistical analyses were performed, using GraphPad Prism 8.0 software. Statistically significant differences were determined by an unpaired, two-tailed Student t test. A p value <0.05 was considered significant.

In our prior study, we reported that SCARB2 expression is chiefly observed in pDCs (30). To further validate this discovery, we conducted an experiment isolating various immune cells from healthy volunteers. Our results were consistent with previous data, showing higher levels of SCARB2 expression in pDCs compared with other immune cells (Fig. 1A). Intracellular localization of SCARB2 was observed in pDCs (Fig. 1B–D).

Previous studies have reported that PSGL-1, another functional receptor of EV71, is primarily expressed on leukocytes (35, 36). In our study, we found that PSGL-1 expression was highest in pDCs in PBMCs (Fig. 2A). Furthermore, we analyzed the expression of PSGL-1 on the surface of other leukocytes and found that it was much higher on pDCs compared with other leukocytes (Fig. 2B, 2C). Both SCARB2 and PSGL-1 are preferentially expressed in pDCs, with PSGL-1 mainly located on the surface of pDCs.

The high expression of both SCARB2 and PSGL-1 in pDCs raises the question of whether pDCs are susceptible to EV71 infection in human peripheral blood. To address this question, we isolated subsets from healthy volunteers and exposed them to a PSGL-1 binding strain (EV71 0804232Y). Our findings indicate that only pDCs were susceptible to EV71 infection (Fig. 3A). Furthermore, we assessed the productivity of EV71 infection in pDCs by infecting purified pDCs with EV71 0804232Y and collecting the culture supernatants at 1 and 24 h postinfection. Subsequently, we infected EV71-susceptible 293-SCARB2 cells, which were established by stable overexpression of human SCARB2 in 293A cells as described previously (33), with the collected culture supernatants for plaque assay detection. We observed cytopathic effects in 293-SCARB2 cells infected with culture supernatants for 24 h but not in 1-h samples (Supplemental Fig. 1). This result suggests that EV71 infection of pDCs is productive.

To examine EV71 viral growth kinetics, we infected purified pDCs with EV71 0804232Y. EV71 efficiently proliferated in pDCs (Fig. 3B). The efficiency of infection was also dose-dependent (Fig. 3C).

To further confirm EV71 replication in pDCs, we then detected purified pDCs by infecting them with EV71-GFP, which expressed GFP upon infection (33). As shown in Fig. 3D, we detected GFP-positive pDCs upon EV71-GFP infection. Consistent with previous results, the efficiency of EV71-GFP infection was also dose-dependent (Fig. 3E).

These results indicate that EV71 preferentially infects and replicates in pDCs in human peripheral blood leukocytes.

To determine whether the interaction of SCARB2 or PSGL-1 with EV71 facilitates EV71 infection of pDCs, two infection inhibition assays were conducted. Initially, EV71 0804232Y was preincubated with PSGL-1–Fc fusion protein or SCARB2-Fc fusion protein separately, with untreated virus serving as a control. Subsequently, purified pDCs were infected with the pretreated or untreated virus. It was observed that the replication of EV71 in pDCs decreased after preincubation with both the PSGL-1–Fc fusion protein and the SCARB2-Fc fusion protein. The inhibition efficiency of the PSGL-1–Fc fusion protein was greater than that of the SCARB2-Fc fusion protein (Fig. 4A). Subsequently, a mAb recognizing SCARB2 was generated, which was able to block the replication of EV71 in 293-hSCARB2 cells (Supplemental Fig. 2A, 2B). Purified pDCs were preincubated with the Abs against SCARB2 (anti-SCARB2) or PSGL-1 (anti–PSGL-1, KPL1 [6]), followed by infection with EV71 0804232Y. In line with the inhibition results obtained with the fusion protein, EV71 infection was impeded by pretreatment of pDCs with both Abs. Anti–PSGL-1 was also more effective than anti-SCARB2 (Fig. 4B).

However, despite the low expression level of SCARB2 on pDCs (Fig. 1C), the mechanism by which the Ab anti-SCARB2 blocks virus infection is unclear. To address this question, pDCs were isolated from healthy volunteers and incubated with FITC-labeled anti-SCARB2 for an extended period at 37°C. This resulted in a notable increase in intracellular Abs (Supplemental Fig. 2C, 2D), indicating that low-level SCARB2 expression on the pDC membrane facilitates anti-SCARB2 endocytosis and accumulation in the endosomes. Therefore, it is hypothesized that anti-SCARB2 reaches the endosomes and obstructs viral entry.

Taken together, these results indicate that EV71 infection of human pDCs depends on both SCARB2 and PSGL-1.

It has been documented that only certain EV71 isolates are able to bind to PSGL-1. The presence of alanine (A), glycine (G), or glutamate (Q) at residue 145 of VP1 enables EV71 isolates to bind to PSGL-1, designating them as PB strains. Conversely, EV71 isolates with glutamic acid (E) at position 145 do not bind to PSGL-1, and are categorized as non-PB strains (10). Our data suggest that both SCARB2 and PSGL-1 are essential for EV71 infection of pDCs, with replication in pDCs being limited to PB strains.

This was confirmed by infecting purified pDCs with mutated viruses at VP1-98 and VP1-145 based on a PB virus (0804232Y), showing that certain mutations at VP1-145 can impact pDC infection (Fig. 5A, 5B). Notably, a wild-type non-PB virus (FUYANG) also failed to replicate in pDCs (Fig. 5C, 5D), underscoring the significance of specific amino acid residues in EV71’s ability to infect pDCs. In summary, our data suggest that only PB strains of EV71 have the capacity to infect pDCs.

To further investigate the correlation between infection rate in pDCs and individual cell expression of PSGL-1, we infected human PBMCs with EV71 and assessed the virus-infected cells in relationship to the PSGL-1 expression of pDCs. Our findings indicate that pDCs with higher levels of PSGL-1 expression are more prone to EV71 infection, and the infection rate in PSGL-1⁺ pDCs is dependent on the multiplicity of infection (MOI) (Fig. 5E, 5F). Additionally, the frequency of PSGL-1⁺ pDCs decreases significantly with higher infection doses (Fig. 5G). These results further support the notion that PSGL-1 expression plays a crucial role in EV71 infection of pDCs.

IFN-α is a critical cytokine in the innate immune response against viral infections, with pDCs initially thought to be the primary producers of IFN-α during such infections (22). When testing EV71’s impact on IFN-α production from pDCs, subsets of PBMCs from healthy individuals were isolated and cocultured with EV71. Consistent with the infection results, only pDCs produced IFN-α after EV71 stimulation (Supplemental Fig. 3A), with a higher production observed after 24 h compared with 12 h (Fig. 6A). Additionally, when pure pDCs were cocultured with EV71 at various MOIs, the amount of IFN-α in the culture supernatants was closely related to the MOIs of EV71, reaching a peak when the MOI was <1 and then declining (Fig. 6B). These findings strongly suggest that EV71 induces IFN-α production from pDCs.

In investigating the contribution of both SCARB2 and PSGL-1 to IFN-α production induced by EV71 in pDCs, two inhibition assays were conducted. First, EV71 was preincubated with the SCARB2-Fc fusion protein or PSGL-1–Fc fusion protein before coculturing with pure pDCs. Preincubation of the virus with PSGL-1–Fc fusion protein led to a dose-dependent reduction in IFN-α secretion, whereas SCARB2-Fc fusion protein did not show a significant inhibition effect (Fig. 7A). Second, pure pDCs were preincubated with anti–PSGL-1 or anti-SCARB2 before EV71 infection. It was observed that anti–PSGL-1, compared with the isotype control, inhibited pDC activation by EV71 (Fig. 7B), whereas anti-SCARB2 did not show a significant inhibition effect (Fig. 7C). These results were consistent with the inhibition results obtained with the fusion protein. Furthermore, it was confirmed that a PB virus, but not a non-PB virus, activated pDCs (Fig. 7D, 7E).

In summary, these findings suggest that surface-expressed PSGL-1, rather than SCARB2, is crucial for EV71-induced IFN-α production from pDCs.

PDCs predominantly use endosomal receptors TLR7 and TLR9 for virus detection and subsequent IFN-α production (22, 37). Our study aimed to determine whether productive infection was required for EV71-induced pDC activation. To address this question, we cocultured pure pDCs with UV-inactivated EV71 0804232Y, which had lost its ability to replicate. Live virus was used as a control. Surprisingly, the UV-inactivated virus induced higher levels of IFN-α production in culture supernatants compared with the live virus (Fig. 8A), suggesting that pDC activation by EV71 is not dependent on productive infection. Additionally, when pure pDCs were infected with EV71-GFP, IFN-α–producing cells were not GFP-positive (Fig. 8B), indicating that nonproductive infection can still activate pDCs to produce IFN-α. To investigate the role of PSGL-1 in early viral entry and pDC activation, we performed two inhibition assays. Our results showed that IFN-α secretion from pDCs stimulated by UV-inactivated EV71 was inhibited by the PSGL-1–Fc fusion protein, but not the SCARB2-Fc fusion protein (Fig. 8C). Consistent with these findings, the Ab anti–PSGL-1 inhibited pDC activation by UV-inactivated EV71, but not anti-SCARB2 (Fig. 8D, 8E). Moreover, only UV-inactivated PB virus was able to activate pDCs (Fig. 8F, 8G).

This study demonstrates that surface-expressed PSGL-1 is crucial for EV71 entry into pDCs and subsequent activation of pDCs to produce IFN-α.

The findings also suggest that the presence of IFN-I may have a negative impact on the virus. To explore this possibility, an inhibition assay was performed by treating pDCs with an anti-IFNAR Ab or an isotype control prior to EV71 infection. The results revealed that blocking IFNAR resulted in elevated viral RNA levels in comparison with untreated or isotype control conditions (Supplemental Fig. 3B), indicating that IFN-α plays a protective role in preventing EV71 infection.

In conclusion, the research emphasizes the importance of PSGL-1 expression on human pDCs for attachment of specific EV71 strains and initiation of viral entry, ultimately triggering pDCs to secrete IFN-α and suppress viral infection before replication (Fig. 9).

The current study has shown that PSGL-1 expression is higher in pDCs compared with other immune cells in human peripheral blood. Some EV71 strains were observed to replicate in pDCs and induce the secretion of IFN-α, which was dependent on PSGL-1 binding rather than productive infection. These findings indicate the essential role of PSGL-1 in the infection of EV71 and the activation of pDCs.

Research has emphasized the importance of IFN in combating EV71 infection, revealing how EV71 hinders IFN-I signaling through mechanisms such as cleavage of IFN-I system components by nonstructural proteins 2A and 3C, as well as downregulation of JAK1 protein expression (24–29). Importantly, note that virus replication is necessary for the activation of these inhibitory pathways. pDCs are known to selectively express TLR7 and TLR9 in the endosome, enabling them to detect viral RNA and DNA before replication. Upon recognition of viral nucleic acids, TLR7 and TLR9 stimulate pDCs to produce significant amounts of IFN-I, especially IFN-α, and mature into DCs (22). This suggests a critical role for pDCs in managing EV71 infection.

PSGL-1, a membrane protein highly expressed on pDCs (Fig. 2), has been identified as a key player in binding to EV71 and initiating caveola-dependent endocytosis of the virus (9). This suggests that PSGL-1 acts as a binding receptor that aids in the virus entry into cells. Furthermore, our study has revealed that pDCs express higher levels of SCARB2 compared with other immune cells. While SCARB2 is mainly found in late endosomes/lysosomes (30), it is not necessary for EV71 to activate pDCs but is essential for viral infection (Figs. 4, 7, 8). SCARB2 has been recognized as the primary mediator of EV71 uncoating (38). Notably, our results indicate that increasing the expression of PSGL-1 did not enhance EV71 replication in SCARB2 knockout cells (data not shown), supporting the idea that PSGL-1 is involved in virus binding and entry into pDCs, while SCARB2 plays a crucial role in viral uncoating during infection.

The VP1-A/G/Q variant has been frequently detected in patients with severe symptoms (14–17), showing an infection efficiency >200-fold higher in RD-A cells compared with VP1-145E virus (18). These results suggest that E-to-A/G/Q mutations may confer greater pathogenicity. However, experiments on monkeys have shown that EV71 with an E at the VP1-145 position can replicate in vivo. Intriguingly, the same strain with an E-to-G mutation exhibited minimal replication in monkeys, and most EV71 sequences obtained from infected monkeys had a G-to-E mutation (12). The authors proposed that the presence of HS on vascular epithelial cells, along with low SCARB2 expression, facilitated the absorption of VP1-A/G/Q virus, leading to abortive infection in vivo (19). Additionally, they noted that the VP1-A/G/Q virus was more easily neutralized by neutralizing Abs (21). These observations contrast with molecular epidemiology research on EV71 (14–17).

Our model aims to clarify the interplay between IFN-α secretion and viral replication. In individuals with a healthy immune system, VP1-145-A/G/Q strains can stimulate pDCs to produce IFN-α, triggering strong early antiviral responses that hinder the replication of these variants. In contrast, the VP1-145-E strain is capable of replicating and causing hand, foot, and mouth disease. Conversely, in individuals with weakened immune systems, such as infants or individuals undergoing treatment, VP1-145-A/G/Q strains can replicate effectively due to the lack of sufficient IFN-α production by pDCs, leading to severe symptoms postinfection (39, 40) (Fig. 9). This could explain the higher percentage of VP1-145 (A/G/Q) strains in patients with severe symptoms. Based on our model, blocking IFNAR or depleting pDCs with Abs may exacerbate illness in adult cynomolgus monkeys infected with PB virus. Considering the activation of the immune system by IFN-α, it is suggested that PB viruses, as a vaccine strain, might offer enhanced efficacy compared with non-PB viruses (41). Furthermore, the level of pDC activation in response to the virus is seen as a potential prognostic indicator following EV71 infection.

The authors have no financial conflicts of interest.

We thank Prof. Honglin Xu (National Vaccine and Serum Institute) for providing us with the wild-type EV71 (0804232Y), Prof. Xiangxi Wang (Institute of Biophysics, Chinese Academy of Sciences) for providing us with the wild-type EV71 (FUYANG) and EV71 Ab (C1), Prof. Chunlai Jiang (Jilin University) for providing us with the pcDNA3.1(−)-T7RNAP, and Prof. Hanchun Yang (China Agricultural University) for providing us with the pWSK29 vector.