Phytophthora rot of asparagus, strain CH98ASP059 (Phytophthora sp.), a PEV2 and PEV3 co-staining strain isolated in Toyama Prefecture, Japan, was incubated at 25°C on V8A plate medium. The strains were then cultured in V8 liquid medium at 25°C for 1 week for total RNA and dsRNA extraction.

The yeast strain BY4742 (MATα, his3Δ, leu2Δ, lys2Δ, ura3Δ) provided by National BioResource in Japan, was used for the construction of the full-length cDNA clone. YPAD (1% yeast extract, 2% polypeptone, 2% glucose, 0.04% adenine sulfate) was used as nutrient-rich medium, and standard yeast medium (Sherman, 2002) was used as dropout medium. Synthetic complete medium without leucine (SC-Leu) and Synthetic complete medium without uracil (SC-Ura) were used as nutrient dropout media according to standard yeast media (Sherman, 2002). For galactose induction experiments, 2% galactose (SGal) was used instead of 2% glucose (SC).

Mycelia of strain CH98ASP059 were collected from liquid culture, dried using Miracloth (Merck, Darmstadt, Germany) and sterile filter paper, and stored at −80°C. According to the manufacturer’s recommendation, total RNA was extracted from 0.1 g of mycelia using the RNeasy Plant Mini Kit (QIAGEN, Hilden, Germany). For dsRNA purification, a spin column containing cellulose D (Advantec, Tokyo, Japan) was used (Okada et al., 2015). Mycelia (0.1 g dry weight) were ground in 500 μL of extraction buffer (100 mM NaCl, 10 mM Tris–HCl pH 8.0, 1 mM EDTA, 1% SDS, and 0.1% (v/v) β-mercaptoethanol), then mixed with equal volumes of phenol-chloroform-isoamyl alcohol (25:24:1). The aqueous phase containing total nucleic acids was mixed with ethanol (final concentration 16%), and dsRNA was selectively recovered by spin column method. Finally, dsRNA was precipitated by ethanol and stored at −80°C. Quantification of total RNA and dsRNA was confirmed by agarose gel electrophoresis.

Yeast transformants were incubated in 50 mL of SC-Leu or SC-Ura at 30°C for about 16 h (h) with shaking (140 rpm) for RNA extraction. Yeast cells were harvested by centrifugation at 1,000x g for 5 min at 4°C. The collected yeast cells were resuspended in 50 mL of ice-cold sterile water. The cells were then centrifuged in a swinging bucket rotor at 1,000× g at 4°C for 5 min. This step was repeated several times. The resulting yeast cells were crushed with liquid nitrogen, and total RNA was extracted using the RNeasy Plant Mini Kit (QIAGEN, Hilden, Germany) according to the manufacturer’s recommendation. The obtained total RNA was concentrated by ethanol precipitation and stored at −80°C.

Rapid amplification of cDNA ends (RACE) was performed to identify the 3′ end of PEV2 and PEV3 nicks. Total RNA extracted and purified from strain CH98ASP059 was polyadenylated by poly (A) tailing reaction before use. The polyadenylated total RNA was used as a template and reacted with the SMARTer® RACE 5′/3’ Kit (Takara Bio USA, Inc., Mountain View, CA, US) according to the manufacturer’s protocol. After amplifying the 3′ end region of the PEV2 and PEV3 nick sequences, they were TA cloned into the pGEM-T easy vector (Promega, Wisconsin, USA) and sequenced by the Sanger method. The primer pairs used are listed in Supplementary Table S1.

For northern blotting analysis, 250 ng of dsRNA and 20 μg of total RNA extracted from strain CH98ASP059 or yeast transformants were heat-denatured, separated on 0.8% agarose 3-(N-morpholino) propanesulfonic acid (MOPS) gel containing 6% formaldehyde, and transferred by capillary blotting onto a positively-charged nylon membrane (Roche, Basel, Switzerland) (Moriyama et al., 1999). After UV cross-linking, prehybridization and hybridization were performed in hybrisolution (5× SSPE, 5× Denhardt’s, 0.5% SDS, 0.1 mg/mL denatured salmon testis DNA). A total of 8 probes were used to detect PEV2 and PEV3 genomes: (+) and (−) PEV2-5′ (nt 144–823) and PEV2-3′ probes (nt 13,280–14,032) (Figure 1A), as well as (+) and (−) PEV3-5′ (nt 66–892) and PEV3-3′ probes (nt 12,830–13,610) (Figure 2A) that were prepared to specifically detect the positive and negative strands, respectively, in the genomic regions of each endornavirus. The probe sequences used for hybridization (PEV2-5′ and PEV2-3′ probes (Figure 1A), and PEV3-5′ and PEV3-3′ probes) were amplified by reverse transcription PCR (RT-PCR) using total RNA extracted from CH98ASP059 as a template, and with Superscript IV reverse transcriptase (Invitrogen, Massachusetts, USA) according to the manufacturer’s instructions. PCR using Go-Taq (Promega, Wisconsin, USA) was then performed under the following conditions with primers that are listed in Supplementary Table S1: denaturation for 1 min at 94°C, and 35 cycles of PCR (98°C for 10 s, 55°C for 30 s, and 72°C for 1 min) followed by a final extension step at 72°C for 5 min. The amplified PCR products were gel purified and cloned into the pGEMT-easy vector (Promega).

According to the manufacturer’s protocol, the riboprobes were prepared using the DIG RNA Labeling Kit (SP6/T7) (Roche, Basel, Switzerland). After the plasmids were linearized with the appropriate restriction enzymes (T7 promoter: NcoI, SP6: SalI), each riboprobe was obtained by run-off transcription. Then, DNA probes were prepared by amplification of DIG-labeled PCR products by PCR using the plasmids mentioned above as templates according to the protocol provided with the PCR DIG Labeling Mix (Roche, Basel, Switzerland). The hybridization was performed in a solution containing specific DIG-labeled RNA or DNA probes for PEV2 or PEV3 sequences for 16 h at 65°C.

Detection of bands using the DIG system was performed using a modification of the manufacturer’s recommended protocol. After hybridization, the membrane was washed twice at 65°C for 5 min in low stringency buffer (2× SSC, 0.1% SDS) and twice for 15 min in high stringency buffer (0.1% SSC, 0.1% SDS). For detection, the membrane was washed in maleic acid wash buffer (1× maleic acid, 0.3% Tween20, pH 7.0) for 2 min, followed by blocking in DIG blocking buffer (Roche, Basel, Switzerland) for 1.5 h. Anti-Digoxigenin-AP, Fab fragments from sheep (Roche, Basel, Switzerland) diluted 5,000-fold in DIG blocking buffer was added and shaken for another 30 min. The membrane was washed again with maleic acid wash buffer for 15 min and repeated twice. The membrane was then incubated in DIG detection buffer (Tris–HCl, pH 9.5, 0.1 M NaCl) for 2 min. Then, ready-to-use CDP-Star^® (Roche, Basel, Switzerland) was added directly to the membrane and allowed to stand in the dark for 5 min, and the signal was detected by Ez-Capture MG (ATTO, Tokyo, Japan).

DsRNA and total RNA were dissolved in 2.0× or 0.2× SSC, and ssRNA was selectively digested by S1 nuclease (10 U) for 15 min at room temperature.

First, five and six fragments were amplified by RT-PCR using the PEV2 and PEV3 RNA genomes as templates, respectively, with the high-fidelity PrimeSTAR^® Max DNA Polymerase (TaKaRa, Shiga, Japan). The following conditions and primers that are listed in Supplementary Table S1 were used: denaturation for 1 min at 94°C, and 35 cycles of PCR (98°C for 10 s, 55°C for 10 s, and 68°C for 2 min) followed by a final extension step at 68°C for 5 min. The PCR products were then purified by the gel extraction method. Then, the fragments were assembled and cloned in the correct order into a shuttle vector by DNA homologous recombination events that occur between overlapping sequence regions of each fragment, by taking advantage of the high DNA recombination activity of yeast cells using lithium acetate method (Gietz et al., 1995). The yeast strain BY4742 (MATα, his3Δ, leu2Δ, lys2Δ, ura3Δ) was cultured in YPAD liquid medium, and the recovered yeast cells were transformed with 25 ng EcoRI- and SmaI-linearized shuttle vectors (pRST425 or pRST426) and 250 ng each of five PEV2 fragments (A ~ E) or six PEV3 fragments (A ~ F) that were previously amplified by RT-PCR.

The transformed yeast were streaked on drop-out medium plates (SC-Leu or SC-Ura), then single colonies were picked for total nucleic acid extraction, and PCR was performed using primers that specifically amplify PEV2 or PEV3.

We compared the growth of BY4742 cells transformed with vectors harboring the galactose-inducible promoter fused with either PEV2 or PEV3 genome (pYES2-PEV2 and pYES2-PEV3, respectively) with that of control (pYES2-Empty). For the dilution drop tests, yeast cells were cultured first on SC-Ura agar medium at 30°C for 2 days, then inoculated into 2 mL of SC-Ura liquid medium and precultured at 30°C with shaking overnight at 140 rpm. The precultured yeast cells were inoculated into 50 mL of SGal-Ura liquid medium at OD₆₀₀ = 0.1 and then incubated at 30°C with shaking at 140 rpm until OD₆₀₀ = 1.0. The cells were subsequently serially diluted ten times in distilled water (DW), and diluted cells were dropped at a volume of 10 μL each onto SGal-Ura (containing 2% galactose) agar medium, which induces heterologous expression, and SC-Ura (containing 2% glucose) agar medium as control. Then their growth was examined after incubation at 30°C for 2 days. One mL of the culture suspension was collected every 4 h from 0 to 12 h after inoculation, and every 12 h thereafter. The number of viable cells, turbidity (OD₆₀₀), and cell size (after 24 h of incubation) were measured by Countstar (ABER, Aberystwyth, UK).

Plasmids expressing PEV2- or PEV3-EGFP fusion protein in S. cerevisiae strain BY4742 were constructed based on pRST426, which had a strong TDH3 promoter and terminator cassette as per our previous report (Urayama et al., 2016). The BY4742 strain was transformed with the shuttle vectors pRST426-EGFP, pRST426-PEV2-NEGFP, pRST426-PEV2-CEGFP, pRST426-PEV3-NEGFP or pRST426-PEV3-CEGFP. Transformants were streaked to obtain single colonies and incubated on SC-Ura agar plates at 30°C for 3 days. A portion of the single colonies was suspended in DW, and the cells and GFP fluorescence were observed at 1,000x magnification using an optical microscope (Olympus IX71, Tokyo, Japan) and differential interference contrast (DIC) optics.

The data of yeast cell diameter were subjected to analysis of variance. Statistical significance is considered when p < 0.05. Cell diameter measurements after 24 h of incubation were averaged and performed three times. Differences between means of different treatments were determined using the Tukey–Kramer test at p < 0.01.

Nick structures in the 5′ region of the positive strand of dsRNA replicative intermediate in plant endornaviruses were found by northern blotting (Fukuhara et al., 1995), while those in Phytophthora endornavirus 1 (PEV1) were suggested by RNA ligase-mediated RACE sequencing of the 3′-terminal of the RNA (Hacker et al., 2005). In this paper, we performed northern blotting of total RNA and dsRNA from PEV2 (Figure 1) and PEV3 (Figure 2) infecting Phytophthora rot of asparagus (strain CH98ASP059) to determine whether nick structures were also present.

We used eight probes to detect high molecular weight bands from purified dsRNA and total RNA that presumably corresponded to the whole genome: a 14.3 kb band from PEV2 (Figures 1B,C) and a 13.8 kb band from PEV3 (Figures 2B,C). Moreover, 1 kb bands were also detected in both endornaviruses, which indicated the presence of nick structures (Figures 1B, 2B). However, these 1 kb bands were detected only when 5′ probes were used for both PEV2 (Figure 1B, black arrows) and PEV3 (Figure 2B, black arrows). As in previous reports of plant endornavirus nick structures, a strong signal in the positive strand of the nick structure was detected by the (−) probe, which detects the positive strand near the 5′ end (Fukuhara et al., 1995). However, our results also showed a signal in the negative strand of the nick structure as detected by the (+) probe. When only 3′ probes for both strands were used, no signal was detected in either strand of PEV2 (Figure 1C) and PEV3 (Figure 2C). These results were similar to those previously reported for plant endornaviruses such as OsEV and BPEV (Fukuhara et al., 1995; Okada et al., 2011), indicating site-specific nick structures near the 5′ terminal in PEV2 and PEV3. These nick structures were also detected by DNA probes (Supplementary Figure S1).

Notably, in PEV2, a 1 kb signal was also detected in the negative strand using the PEV2-5′ (+) probe that specifically detects the negative strand near the 5′ terminal (Figure 1B, black arrowheads), which has not been shown thus far in other plant endornaviruses. The intensity of the 1 kb signal of the positive strand was stronger than that of the negative strand in PEV2. In PEV3, a 1 kb band in the negative strand was also observed, albeit at a weaker signal intensity (Figure 2B, black arrowheads).

The specific position of the 3′-nick terminal was determined by the 3′-terminal RACE method. The 3′-terminal of the nicks were found to be located at nt 958 in PEV2 by sequencing four independent clones and at nt 965 in PEV3 by sequencing five independent clones. These results corresponded to the northern blot analyses and showed that the nick structures were detected around 1,000 nt from the 5′-terminal in PEV2 and PEV3, suggesting that PEV2 and PEV3 have nick structures at similar positions as in previously reported endornaviruses. The 5′-terminal RACE method was also attempted as previously shown for OsEV and BPEV, which failed due to the relatively long 3′ fragment of PEV2 and PEV3. It is possible that the different results were obtained because purified dsRNA was used as a template for cloning nicks in OsEV and BPEV, while total RNA was used as a template for PEV2 and PEV3 in this study. However, since full-length genomic RNA was detected by the northern blot analysis using probes that detected the 5′-terminal sequence of PEV2 or PEV3 (Figures 1B, 2B), the 5′-terminal sequence of each RNA genome was assumed to be amplified by the 5′-terminal RACE method and our conjecture on its failure was considered to be reasonable.

To investigate in detail the genomic structure of the nicks in PEV2 and PEV3 we used the probes for detecting the nicks (PEV2-5′ (+, −) and PEV3-5′ (+, −) probes) following S1 nuclease treatment (Figures 3A, 4A).

The S1 nuclease tests were performed under two conditions: a high salt concentration of 2.0x SSC, which strengthens the binding between each strand of dsRNA, and a low salt concentration of 0.2x SSC, which relaxes it (Figures 3B,C, 4B,C).

When the PEV2-5′ (−) probe was used following treatment with S1 nuclease, both bands corresponding to the 14.3 kb PEV2 genome and 1 kb nick fragment were detected in the dsRNA and total RNA (Figure 3B). On the other hand, when the PEV2-5′ (+) probe was used, the 14.3 kb band was exclusively detected in the purified dsRNA samples, and the 1 kb nick fragment was detected only rarely (Figure 3C, left side). However, a 1 kb nick fragment was detected in the total RNA sample (Figure 3C, right side), indicating that negative-stranded 1 kb nick fragments occurred in small amounts. In any case, the presence of 1 kb nick fragments after S1 nuclease treatment and denaturing gel electrophoresis suggested that dsRNA forms of PEV2 have a nick structure in the positive strand (Figure 3A). Similar results were obtained in PEV3; the PEV3-5′ (−) probe detected a band of 13.8 kb that corresponded to the positive strand of the PEV3 genome and a 1 kb nick fragment in the positive strand, although the signal was relatively weak (Figure 4B). The PEV3-5′ (+) probe detected exclusively the 14.3 kb band corresponding to the negative strand of the genome and only a tiny amount of the 1 kb nick fragment of the negative strand (Figure 4C).

Since the previous experiments showed that the nick structure in the positive strand was more strongly detected (higher quantity) than that in genomic RNA by northern blotting (Figures 1, 2), the nick structure in the positive strand may also exist as ssRNA independent of genomic RNA. To clarify this hypothesis, non-denatured RNA was electrophoresed and under conditions allowing ssRNA to be preferentially detected, we investigated the nick structure using 5′-terminal strand-specific riboprobes for PEV2 (Figures 3D,E) and PEV3 (Figures 4D,E). In total RNA from PEV2, the 14.3 kb genomic band and the 1 kb nick fragment band were detected by the PEV2-5′ (−) probe (Figure 3D right side). On the other hand, no band was detected by the PEV2-5′ (+) probe, which specifically detects the negative strand, under non-denaturing condition (Figure 3E, right side). In the case of PEV3, the 13.8 kb positive strand genomic band and the 1 kb nick fragment band were detected only with the PEV3-5′ (−) probe (Figure 4D, right side). This result suggests that the genomic and nick RNA components of PEV2 and PEV3 exist as ssRNA(+) in the host cells, while the negative strand genomic RNA and nick fragments of both viruses exist only as dsRNA. This series of experiments indicated that the positive-stranded 1 kb nick fragment exists independently of dsRNA in host cells, while the negative-stranded 1 kb nick fragment exists only in dsRNA form.

Twelve colonies that were selected for analysis from the drop-out SC-Leu or SC-Ura medium plates harbored the full-length PEV2 and PEV3 cDNAs cloned in pRST425 and pRST426, respectively (Figure 5, Supplementary Figure S2). Less frequently, colonies that did not contain the desired sequence were obtained, which were transformants of empty shuttle vectors possibly due to incomplete restriction enzyme treatment. These results showed that highly efficient constructions of full-length cDNA clones of PEV2 and PEV3 were achieved using DNA homologous recombination in yeast cells.

Using yeast cells transformed with pRST425-PEV2 (Figure 6A) and pRST426-PEV3 (Figure 6B), northern blotting was performed to analyze transcripts of PEV2 (Figure 6C) and PEV3 (Figure 6D) genomic RNA transcribed by RNA polymerase II under the control of the constitutive TDH3 promoter.

In the yeast transformants of pRST425-PEV2, a band of 14.3 kb, which corresponded to the full-length PEV2 genome, was detected with both PEV2-5′ (−) and PEV2-3′ (−) probes (Figure 6C). On the other hand, in the yeast transformants of pRST426-PEV3, a band around 1 kb in size was detected in addition to the full-length band of 13.8 kb when the PEV3-5′ (−) probe was used (Figure 6D, left side). When the PEV3-3′ (−) probe was used, only a band of 13.8 kb in size was detected. The results of a parallel experiment using total RNA from Phytophthora were similar to those shown in Figures 1B,C, 2B,C. To determine the size of nick-like RNA occurring in yeast cells, we performed 3’RACE to PEV3 heterologously expressed in yeast (Supplementary Figure S3). After polyadenylated by poly (A) tailing reaction, an 856 nt amplified band whose size was 109 nt shorter than the PEV3 nick occurred in native host, strain CH98ASP059 (nt 965) was observed in heterologously PEV3 expressed yeast cells (YP3) (Supplementary Figure S3). Without poly (A) tailing reaction, no bands were amplified in the strain CH98ASP059 ASP nor YP3 when PEV3-66 primer located in a 5′ -nick fragment was used (Supplementary Figure S3). Using primer PEV3-12830 which is located near the 3′ terminus, a band of 980 bp was amplified with polyadenylated transcripts of ASP and YP3, whose 3′ terminuses were polyadenylated. This result showed that the RNA of nt 856 generated by YP3 did not have Poly (A) and might not be generated by termination of transcription by RNA Polymerase II.

These results indicate that the TDH3 promoter driven in yeast cells can complete transcription of the PEV2 and PEV3 full-length RNA genomes. Besides, a 1 kb-sized nick product was detected in pRST426-PEV3-transformed cells (Figure 6D, left side), albeit less intense and smaller than that from the natural host Phytophthora spp.

We previously reported that high titer of PEV2 and PEV3 in Phytophthora fungus severely inhibited hyphal growth (Uchida et al., 2021). To investigate the effect of expression of the full-length PEV2 and PEV3 RNA on yeast cell growth, we generated constructs (pYES2-PEV2 and pYES2-PEV3, respectively) in which the full-length cDNA clones of PEV2 or PEV3 were cloned downstream of the galactose-inducible promoter (GAL1) (Figures 7A,B). These constructs were created using multiple homologous recombination as in the cases of pRST425-PEV2 and pRST426-PEV3 (Supplementary Figure S2). A 14.3 kb band from pYES2-PEV2-transformed cells and a 13.8 kb band from pYES2-PEV3-transformed yeast (Figures 7C,D) were detected by northern blot hybridization using the PEV2-5′ (−) and PEV3-5′ (−) probes to assess for PEV2 and PEV3 RNA transcripts transcribed under the control of GAL1. A 1 kb nick fragment was also detected as a distinct band in yeast cells harboring pYES2-PEV3 (Figure 7D).

Colony formation on SC-Ura agar medium as well as viable cell counts and turbidity (OD₆₀₀) over time in SC-Ura liquid medium were compared among the three transformants of BY4742 strain harboring pYES2-PEV2, pYES2-PEV3, and pYES2-Empty (for control). On 2% glucose agar medium, no difference in single colony growth was observed between the PEV2- or the PEV3-expressing transformants and the control transformants (Figures 7E,G). On the other hand, in 2% galactose agar medium where the GAL1 promoter induced downstream gene expression, the PEV2 and the PEV3-expressing transformants had fewer colonies, showed smaller colony size, and had reduced growth as compared to the control transformants (Figures 7F,H).

In addition, we investigated the effects of heterologous expression of PEV2 and PEV3 at low (25°C) and high temperatures (37°C) on yeast cell growth. On SC-Ura agar medium containing 2% glucose, no discernible difference in growth was observed between the three transformants at each temperature (Figures 7I,K,M), but on SC-Ura containing 2% galactose, the growths of the PEV2 and PEV3 transformants were apparently slower than that of the control (Empty) transformants under both temperatures (Figures 7J,L,N).

In growth tests based on OD₆₀₀ values in SC-Ura liquid medium, the turbidity of PEV2 and PEV3 transformants was significantly lower as compared to control transformants (empty) after 12 h of incubation, but subsequently reached almost identical levels to that of control (Figure 8A). Similar results were obtained by counting the number of viable cells in these three transformants (Figure 8B). After 24 h of incubation at 30°C, the sizes of living cells of PEV2 and PEV3 (average diameter = 6.25 μm and 6.05 μm, respectively; Turkey test: p < 0.01) transformants were significantly smaller as compared to the control strain (6.8 μm) (Figure 8C). The observed cell sizes of the empty appeared to be larger than those of PEV2 and PEV3. PEV2, PEV3, and empty cells were also observed under an optical microscope after 24 h of incubation at 30°C. PEV2 and PEV3 cells were apparently uniform in size rather than the empty cells which tended to show varied cell sizes (Figure 8D). These results suggested that the heterologous expression of PEV2 and PEV3 did not induce cell death toxicity in yeast cells, but had inhibitory effects on the growth rates of the cells.

To visually observe the expression of the endornavirus ORF proteins, EGFP fusion expression vectors were constructed by fusing EGFP with the N-terminus of the PEV2 or PEV3 ORF (PEV2-NEGFP and PEV3-NEGFP, respectively), or their C-terminus (PEV2-CEGFP and PEV3-CEGFP, respectively) (Figures 9A–D). These vectors were constructed by multiple homologous recombination of the full-length ORF of PEV2 or PEV3 in yeast cells using pRST426-EGFP, in which EGFP was incorporated under the TDH3 promoter. In this study, we failed to obtain PEV3-NEGFP constructs despite several attempts (data not shown), which may be due to the toxicity of the fused protein caused by structural perturbation.

Yeast cells carrying the above constructs, which expressed the PEV2- or PEV3-EGFP fusion protein, were cultured for 20 h, and EGFP fluorescence was observed under a fluorescence microscope. Yeast cells that were transformed with the control pRST426-EGFP vector and expressed free EGFP exhibited strong green fluorescence throughout the cytoplasm (Figure 9A). On the other hand, very weak green fluorescence was observed in yeast cells expressing PEV2-NEGFP, possibly due to the very low expression level of the PEV2-NEGFP fusion protein (Figure 9B). In contrast, PEV2-CEGFP showed some bodies in cells (Figure 9C). PEV3-CEGFP also exhibited a conspicuous localized dot-like green fluorescence (Figure 9D).

To examine the level of transcripts produced from the EGFP fusion constructs, northern blotting using the EGFP sequence as a probe was performed (Figures 9E,F). In cells transformed with control pRST426-EGFP, a strong band of 714 bp was detected, whereas weak bands of around 14.5 kb and 714 bp were detected in cells transformed with pRST426-PEV2-NEGFP and pRST426-PEV2-CEGFP (Figure 9E). Further, the intensity of a 1.5 kb band was much weaker in PEV2-NEGFP-expressing cells as compared to that in cells with PEV2-CEGFP (Figure 9E), in accordance with the observed difference in fluorescence intensity (Figures 9B,C). Northern blot analysis for PEV3-CEGFP showed that the intensity of the signal was comparable to that of PEV2-CEGFP (Figure 9F), consistent with the results of fluorescent microscopy (Figure 9D).