Introduction
A new coronavirus (CoV) emerged during the summer of 2012 in Saudi Arabia (
1–3). This virus has been named the Middle East respiratory syndrome coronavirus (MERS-CoV) by the International Coronavirus Study Group of the International Committee on Taxonomy of Viruses (ICTV) (
4). Based on its genome sequence, the virus has been classified within lineage C of the genus
Betacoronavirus. MERS-CoV is closely related to
Tylonycteris bat CoV HKU4 and
Pipistrellus bat CoV HKU5, the two prototype species in lineage C
Betacoronavirus. MERS-CoV is even closer in sequence to two CoVs found in bats circulating in the Netherlands and Spain (
5–7). Therefore, reemergence of the virus in these and other European and Middle Eastern countries is a realistic possibility. Using a nonhuman primate disease model (rhesus macaques), it has been shown that Koch’s postulates can be fulfilled (
8).
MERS-CoV has been disseminated into 10 countries of the Middle East, North Africa and Europe. The virus may have crossed from bats to a domesticated or agricultural animal species and subsequently spread from there into humans. To date, the identities of both reservoir and intermediate hosts remain unknown, complicating virus control efforts. Human-to-human transmission has been documented in at least four independent hospital settings (
9–12). As of 7 August 2013, a total of 94 cases have been confirmed, of which around 50% were fatal (
http://www.cdc.gov/coronavirus/mers/index.html). This might be an overestimation of the case fatality risk, as many infected patients likely have not required hospital assistance. In fact, recent data suggest that mild respiratory illness might also be part of the clinical spectrum of MERS-CoV infection (
3). In addition to mild or acute respiratory illness, other reported clinical symptoms are abdominal pain and diarrhea, fever, and in some cases, renal failure (
9). Many hospitalized cases occurred in persons with chronic underlying medical conditions or immunosuppression (
3,
13). The virus loads are highest in lower respiratory tract samples, although low concentrations of viral RNA can also be found in stool, urine, and blood samples (
12).
The genome of MERS-CoV includes more than 30,100 nucleotides and contains at least 10 predicted open reading frames (ORF1a, ORF1b, S, 3, 4a, 4b, 5, E, M, and N), 9 of which seem to be expressed from a nested set of eight mRNAs (
14,
15). Interestingly, the partial genome sequences of three independent MERS-CoV isolates reveal that it evolved following a strict molecular clock model (
6).
A functional receptor of MERS-CoV is dipeptidyl peptidase 4 (DPP-4) from both human and bat (
16). This receptor binds to a 231-residue region in the spike (S) protein of MERS-CoV (
17,
18), a domain different from the receptor-binding site of other
Betacoronaviruses (
18). Infection of human airways by MERS-CoV prevents the induction of interferon-regulating factor 3 (IRF-3)-mediated antiviral alpha/beta interferon (IFN-α/β) responses. However, MERS-CoV was markedly more sensitive to the antiviral state established by ectopic IFN than severe acute respiratory syndrome CoV (SARS-CoV) (
14,
19,
20).
Soon after MERS-CoV emergence, a diagnostic assay was designed (
21). Similarly, antivirals inhibiting virus replication, such as cyclosporine A, IFN-α, or ribavirin, have been described (
14,
22,
23). In contrast, reliable vaccines have not yet been developed, although the S protein and the receptor-binding site within this protein induce neutralizing antibodies and, in principle, could serve as a subunit vaccine (
17). CoVs infect respiratory and enteric mucosal areas, and thus, induction of mucosal immunity is necessary to protect these tissues from infection. Live attenuated viruses are expected to elicit mucosal immunity more efficiently than nonreplicating antigens, which elicit reduced secretory immune responses. Live attenuated viruses can be generated by the deletion of genes conferring virulence, a procedure that requires the availability of a reverse genetics system for MERS-CoV. In this article, we describe the construction of an infectious cDNA clone of MERS-CoV using a bacterial artificial chromosome (BAC). Using this clone, recombinant MERS-CoV (rMERS-CoV) deletion mutants were constructed lacking genes nonessential for virus replication. In addition, we deleted the structural envelope (E) protein gene, because in previous work from our laboratory, deletion of the E gene in two other CoVs led to mutants that were either replication-competent, propagation-defective viruses or attenuated viruses (
24–26). All deletion mutants efficiently replicated and spread in cell cultures except the one in which the E gene was deleted, which was replication competent but propagation defective. This virus was propagated in cells by providing E protein in
trans. Therefore, this deletion mutant missing the E gene can serve as the basis for a safe vaccine candidate.
DISCUSSION
The emergence of MERS-CoV represents a public health threat that requires further research to understand the virus biology and provide the basis for the development of control strategies. This paper describes for the first time the construction of a reverse genetics system for MERS-CoV, using BACs as vectors. The recombinant viruses described in this work were rescued using a combination of synthetic biology and reverse genetics techniques, as previously described for other CoVs (
33). This system constitutes a valuable molecular tool for the rational design and launching of attenuated viruses that may serve as efficient and safe vaccine candidates. In addition, the infectious cDNA clone will be useful to study the role of specific viral genes in virus-host interactions in the context of the complete viral cycle.
A full-length cDNA copy of MERS-CoV-EMC12 was generated from synthetic fragments cloned downstream from the CMV promoter in a BAC. The BAC-based strategy allows the efficient and reproducible intracellular production of viral RNA, since it is first synthesized in the nucleus by the cellular RNA polymerase II (Pol II) and then amplified in the cytoplasm by the viral replicase encoded in the RNA itself (
27,
28). The MERS-CoV infectious cDNA was stably maintained in bacteria for more than 200 generations, allowing the easy and direct manipulation of the viral cDNA for molecular studies. In addition, this BAC-based system allows for the generation of viral replicons that may be used for the screening of drugs affecting viral RNA synthesis (
27,
34).
The full-length sequence of rMERS-CoV, recovered from the infectious cDNA clone, was completely identical to that published for the original MERS-CoV-EMC12 isolate (
15), except for the silent point mutation introduced as a genetic marker. Therefore, both viruses should have the same biological properties. In fact, the growth kinetics and CPE caused by both viruses were similar when the same multiplicity of infection (MOI) was used to infect Huh-7 cells (
14).
Interestingly, the rMERS-CoV sequence seemed more stable in Huh-7 cells than in Vero A66 cells. However, the data presented in this article are statistically very limited to definitively conclude that virus genome stability depends on the cell type used. Based on the preliminary data presented here, it would be interesting to analyze the evolution of the MERS-CoV sequence in different cell types, including human respiratory epithelial cells.
The 3′ third of the MERS-CoV genome contains a set of accessory genes encoding proteins with no similarity to other viral or mammalian known proteins (
35). In general, CoV accessory genes are not essential for virus growth
in vitro (
36–39). The reverse genetics system described in this article was used to study the importance of these proteins in cell culture. MERS-CoV genes 3, 4a, 4b, and 5 were each found to be dispensable for virus replication in tissue cultures. Interestingly, some of the rMERS-CoV viruses recovered from Vero A66 cells contained mutations in the accessory gene genome region that would prevent the expression of any of these genes. Similar results were previously reported for the original MERS-CoV-EMC12 isolate after passage in Vero cells (
15). These data suggested an apparent lack of selection pressure on MERS-CoV accessory genes during passages in cell culture and reinforced the dispensability of these genes for virus growth
in vitro.
Although not essential in tissue culture, these MERS-CoV accessory genes could have an important role in virus-host interaction
in vivo, leading to attenuated phenotypes. CoV accessory genes have been associated with the modulation of viral virulence (
40). Among all CoVs, SARS-CoV contains the largest number of accessory genes, and it has been proposed that these genes may have important contributions to its high virulence (
26,
39). To date, mouse hepatitis virus (MHV) ns2 and 5a, TGEV 7, and SARS-CoV 3b and 6 proteins have been implicated in the modulation of innate immune responses, using different mechanisms to influence virus virulence (
36,
41–44).
rMERS-CoV-ΔE was a replication-competent, propagation-deficient virus and was only efficiently disseminated in cells expressing the E protein in
trans. In the presence of transiently expressed E protein, rMERS-CoV-ΔE yielded maximum progeny viral titers of around 10
3 TCID
50/ml. This modest yield could be improved by the generation of cell lines stably expressing the E protein. In fact, a direct relation between viral titers and the amount of E protein expressed was previously observed for TGEV (
45). However, high expression levels of E protein could induce apoptosis, as described for MHV E protein expression (
46). To overcome this potential adverse effect in the case of MERS-CoV, an inducible system for E protein expression would have to be established.
rMERS-CoV-ΔE did not spread in cells in the absence of E protein, thus constituting a single-cycle replicative virus. However, infected cells produced syncytia, which suggests good expression of viral S protein. In addition, high levels of N protein were observed by immunofluorescence. These data suggested that the high expression levels of viral proteins might serve as potent immunogens to elicit a protective immune response. In the case of SARS-CoV, it has been shown that nonreplicating SARS-CoV-like particles bearing the E, S, and membrane (M) proteins induced immune responses that were protective against SARS in mice (
47,
48). In addition, SARS-CoV inactivated viruses induced adaptive immunity that protected against challenge (
49–52). The potential of rMERS-CoV-ΔE as a vaccine candidate is reinforced by previous observations indicating that a SARS-CoV lacking the E gene (SARS-CoV-ΔE) is attenuated and induces protection in hamsters, transgenic mice, and conventional aged mice (
53–55).
MERS-CoV infects mucosal areas in the lungs and, probably, the enteric tract. Mucosal immunity in a specific tissue, such as in lung infections with MERS-CoV, is optimally induced by local stimulation. Therefore, immunization with live attenuated forms of rMERS-CoV-ΔE virus grown in a packaging cell line providing the E protein in trans may be a convenient option, particularly in comparison with purified MERS-CoV antigens, such as the S protein, that could serve as a subunit vaccine.
Vaccines based on live attenuated viruses may present biosafety problems associated with the possibility of reversion to virulent phenotypes or causing disease in immunocompromised individuals. In this sense, the use of rMERS-CoV-ΔE would be a safer option, as it does not propagate in the absence of E protein expression, preventing straightforward reversion to virulence. To increase the biosafety of a rMERS-CoV-ΔE-based vaccine, additional safety guards could be included, such as the previously described attenuating mutations in distant genomic locations, like those encoding the nsp1 (
56,
57) or nsp14 (
58) replicase proteins, or by introducing genomic rearrangements (
59). Overall, we consider rMERS-CoV-ΔE a promising vaccine candidate that should be further developed.
rMERS-CoV-ΔE could also be used as the starting point to generate an inactivated vaccine in case of an urgent need to control the disease. In order to guarantee the absence of virulent viruses after an incomplete chemical inactivation due to clump formation, potential noninactivated viruses would be propagation defective and, therefore, attenuated.
rMERS-CoV-ΔE, in which one of the nonessential accessory proteins was deleted, could be considered a marker vaccine, as it will allow the sera of field-infected patients to be distinguished from sera of vaccinated patients, based on the lack of antibodies specific for nonessential viral proteins (
60).
The rMERS-CoV-ΔE could also be considered a viral replicon, as its genome self amplifies in infected cells but infection is not efficiently spread from cell to cell. The construction of a minimal replicon is also possible with reduced effort using the infectious clone, a project that is currently in progress in our laboratory. Therefore, the introduction of a reporter gene, such as green fluorescent protein, within this replicon could easily generate a useful tool for MERS-CoV antiviral drug screening.
In this paper, we describe for the first time a reverse genetics system of MERS-CoV engineered on BACs, which has allowed the generation of the first modified live vaccine candidate to protect against MERS-CoV. Furthermore, this reverse genetics system is a useful tool for the identification of viral genes involved in pathogenesis and the associated signaling pathways. Drugs inhibiting these pathways would be potential antivirals.
MATERIALS AND METHODS
Cells and viruses.
Baby hamster kidney cells (BHK-21) were obtained from American Type Culture Collection (ATCC CCL-10). Human liver-derived Huh-7 cells were kindly provided by R. Bartenschlager (University of Heidelberg, Germany). African green monkey kidney-derived Vero A66 cells were kindly provided by A. Carvajal (University of Leon, Spain). In all cases, cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 25 mM HEPES, 1% nonessential amino acids (Sigma), and 10% fetal bovine serum (FBS) (BioWhittaker). Virus titrations were performed on Vero A66 or Huh-7 cells following standard procedures and using closed flasks or plates sealed in plastic bags. For plaque assays, infected cells were overlaid with DMEM containing 0.6% low-melting agarose and 2% FBS, and at 72 h.p.i., cells were fixed with 10% formaldehyde and stained with crystal violet. For 50% tissue culture infectious dose (TCID50) assays, CPE was recorded at 72 h.p.i. All work with infectious virus was performed in biosafety level 3 facilities by personnel wearing positive-pressure air-purifying respirators (high-efficiency particulate Air-Mate).
Plasmids and bacteria strains.
Plasmid pBeloBAC11 (
61), kindly provided by H. Shizuya (California Institute of Technology, Pasadena, CA), was used to assemble the MERS-CoV infectious cDNA clone. This plasmid is a low-copy-number plasmid (one to two copies per cell) based on the
E. coli F factor (
62) that allows the stable maintenance of large DNA fragments in bacteria.
E. coli DH10B (Gibco/BRL) cells were transformed by electroporation using a MicroPulser unit (Bio-Rad) according to the manufacturer’s instructions. BAC plasmid and recombinant BACs were isolated and purified using a large-construct kit (Qiagen), following the manufacturer’s specifications.
Construction of a full-length cDNA clone of MERS-CoV.
Based on the data of the full-length sequence of the MERS-CoV-EMC12 strain (GenBank accession number JX869059) (
15), a MERS-CoV infectious cDNA clone was assembled in BAC using a three-step strategy. In the first step, the restriction sites BamHI (genomic position 806), StuI (genomic positions 7,620 and 9,072), SwaI (genomic position 20,898), and PacI (genomic position 25,836), present in the viral genome, were selected (
Fig. 1A). Second, the intermediate plasmid pBAC-MERS-5′3′ was constructed as the backbone for assembly of the full-length cDNA clone (
Fig. 1B). To generate this plasmid, two DNA fragments were generated by chemical synthesis (Bio Basic, Inc.). The first fragment contained the CMV promoter fused to the first 811 nucleotides of the viral genome flanked by SfoI and BamHI sites, and the other one contained a multicloning site with the restriction sites selected in the first step (BamHI, StuI, SwaI, and PacI) followed by the last 4,272 nucleotides of the viral genome joined to a 25-nt poly(A), HDV ribozyme, and BGH termination and polyadenylation sequences. The first DNA fragment was cloned into pBeloBAC11
−StuI (a pBeloBAC without the StuI restriction site) and digested with SfoI and BamHI to generate the plasmid pBAC-MERS-5′, and then the plasmid pBAC-MERS-5′3′ was generated by cloning the second DNA fragment, digested with BamHI and SfiI, into pBAC-MERS-5′ digested with the same restriction enzymes. Finally, the third step was the assembly of the full-length cDNA clone (pBAC-MERS
FL) by sequential cloning of four overlapping DNA fragments (MERS-1 to MERS-4) into the multicloning site of the intermediate plasmid pBAC-MERS-5′3′ (
Fig. 1C). The overlapping DNA fragments flanked by the appropriate restriction sites were generated by chemical synthesis (Bio Basic, Inc.). In the case of fragment MERS-3, a silent mutation (T to C) was introduced at position 20,761 in order to eliminate the SwaI restriction site at position 20,760 and to use it as a genetic marker. The genetic integrity of the cloned DNAs was verified throughout the assembly process by extensive restriction analysis and sequencing.
Construction of MERS-CoV cDNA clones lacking accessory genes 3, 4a, 4b, and 5.
The deletion of gene 3 was generated by PCR-directed mutagenesis using the plasmid pUC-MERS-1 (a pUC plasmid containing the MERS-1 fragment spanning nucleotides 20,898 to 25,836 of the MERS-CoV genome) as the template and the oligonucleotides MERS-S-Tth111I-VS (5′ TGCTATTT
GACAAAGTCACTATAGCTGATC 3′, where the restriction site Tth111I is underlined) and MERS-S-PacI-RS (5′ CCC
TTAATTAACTGAGTAACCAACGTCAAAAAGATTCACACTATTAGTGAACATGAACCTTATGCGGCTCGAGGTCGTATTCC 3′, where the restriction site PacI is underlined). The PCR product, including the deletion (from nucleotides 25,518 to 25,803), was digested with Tth111I and PacI and cloned into the same sites of pUC-MERS-1, leading to pUC-MERS-1-∆3. To generate pBAC-MERS-∆3, the SwaI-PacI digestion product from pUC-MERS-1-∆3 was cloned into the same restriction sites of pBAC-MERS
FL (
Fig. 3A).
The deletions of genes 4a, 4b, and 5 were introduced by PCR-directed mutagenesis, using as a template the plasmid pBAC-MERS-3′ (a BAC plasmid containing the MERS-3′ fragment spanning nucleotides 25,836 to 30,107 of the MERS-CoV genome). For deletion of genes 4a and 4b, overlapping PCR fragments were amplified using oligonucleotides del4ab-VS (5′ GAACTCTATGGATTACGGTTGTCTCCATACGGTC 3′) and del4ab-RS (5′ GACCGTATGGAGACAACCGTAATCCATAGAGTT 3′). The final PCR product was amplified with outer oligonucleotides T7 and SA27201RS (5′ CAAACAGTGGAATGTAGG 3′), digested with PacI and NheI, and cloned into the same restriction sites of pBAC-MERS-3′, leading to pBAC-MERS-3′-∆4ab that contains a deletion spanning nucleotides 25,862 to 26,751 of the MERS-CoV genome. For gene 5 deletion, a PCR fragment lacking gene 5 (nucleotides 26,835 to 27,513) was amplified using oligonucleotides SA25834VS (5′ G
TTAATTAACGAACTCTATGGATTACG 3′, where the restriction site PacI is underlined) and del5-SanDI-RS (5′ CAC
GGGACCCATAGTAGCGCAGAGCTGCTGTTAAAATCCTGGATG 3′, where the restriction site SanDI is underlined), digested with PacI and SanDI, and cloned in the same sites of pBAC-MERS-3′, leading to pBAC-MERS-3′-∆5. To generate plasmids pBAC-MERS-∆4ab and pBAC-MERS-∆5, the PacI-RsrII digestion products from plasmids pBAC-MERS-3′-∆4ab and pBAC-MERS-3′-∆5 were cloned in the same sites of pBAC-MERS
FL (
Fig. 3A). All cloning steps were checked by sequencing of the PCR fragments and cloning junctions.
Construction of a MERS-CoV cDNA clone lacking the structural E gene.
The pBAC-MERS-ΔE, encoding a MERS-CoV lacking the E gene, was constructed from the full-length plasmid pBAC-MERS
FL. To this end, the SanDI-RsrII DNA fragment (2,634 bp) from pBAC-MERS
FL was exchanged with a chemically synthesized (Bio Basic, Inc.) SanDI-RsrII DNA fragment with a deletion from nucleotides 27,580 to 27,786 that included the TRS core sequence and the first 197 nucleotides of the E gene (
Fig. 4A). The genetic integrity of the cloned DNA was verified by restriction analysis and sequencing.
Recovery of recombinant viruses from the cDNA clones.
To recover infectious virus, BHK cells were grown to 95% confluence in a 12.5-cm2 flask and transfected with 6 µg of the infectious cDNA clone using 18 µg of Lipofectamine 2000 (Invitrogen) according to the manufacturer’s specifications. At 6 h.p.t., cells were trypsinized, plated over a confluent monolayer of either Vero A66 or Huh-7 cells grown in a 12.5-cm2 flask, and incubated at 37°C for 72 h. The cell supernatants were harvested and passaged once on fresh cells, and the recovered viruses were cloned by three rounds of plaque purification, following standard procedures.
Virus genome sequencing.
The complete genome sequence of each rescued recombinant MERS-CoV was determined by sequencing overlapping RT-PCR fragments of 2.5 kb covering the full-length viral genome. Reverse transcription and PCRs were performed with specific oligonucleotides using ThermoScript reverse transcriptase (Invitrogen) and the Expand high-fidelity PCR system (Roche), respectively, following the manufacturers’ recommendations. The genomic 5′- and 3′-terminal sequences were determined using the 5′/3′ RACE (rapid amplification of cDNA ends) kit (Roche) according to the manufacturer’s specifications. Sequence assembly and comparison with the consensus sequence of the MERS-CoV-EMC12 strain were performed with the SeqMan and MegAlign programs (Lasergene, Madison, WI).
Generation of Huh-7 cells expressing MERS-CoV E protein.
For the generation of Huh-7 cells transiently expressing E protein, cells were nucleofected with the plasmid pcDNA3-E (expressing the MERS-CoV E protein under the CMV promoter) by using a 4D Nucleofector device (Lonza) and the buffer and program recommended by the manufacturer. For the construction of plasmid pcDNA3-E, the E gene was amplified by PCR using pBAC-MERSFL as the template and the specific oligonucleotides E1-EcoRI-VS (5′ GTGCTGGAATTCGCCGCCATGTTACCCTTTGTCCAAGAACGAA 3′, restriction site EcoRI is underlined) and E249-XhoI-RS (5′ CGCCCAGCTCGAGTTAAACCCACTCGTCAGGTGG 3′, restriction site XhoI is underlined) and cloned into the plasmid pcDNA3 (Invitrogen) digested with EcoRI and XhoI.
Analysis of viral RNA synthesis by RT-qPCR.
Total intracellular RNA was extracted from transfected or infected cells with the RNeasy miniprep kit (Qiagen) according to the manufacturer’s specifications. In the case of transfected cells, the residual DNA was removed from samples by treating 7 µg of each RNA with 20 U of DNase I (Roche) in 100 µl for 30 min at 37°C, and DNA-free RNAs were repurified using the RNeasy miniprep kit (Qiagen). Viral RNA synthesis was quantified by RT-qPCR. Total cDNA was synthesized with random hexamers from 100 ng of total RNA using a high-capacity cDNA reverse transcription kit (Invitrogen). Using this cDNA, the viral RNA synthesis was analyzed using two custom TaqMan assays specific for MERS-CoV gRNA (forward primer 5′ GCACATCTGTGGTTCTCCTCTCT 3′, reverse primer 5′ AAGCCCAGGCCCTACTATTAGC 3′, and MGB probe 5′ TGCTCCAACAGTTACAC 3′) and sgmRNA N (forward primer 5′ CTTCCCCTCGTTCTCTTGCA 3′, reverse primer 5′ TCATTGTTATCGGCAAAGGAAA 3′, and MGB probe 5′ CTTTGATTTTAACGAATCTC 3′). Data were acquired with an Applied Biosystems 7500 real-time PCR system and analyzed with ABI PRISM 7500 software, version 2.0.5. The relative quantifications were performed using the cycle threshold (2
−∆∆CT) method (
63). To normalize for differences in RNA sampling, the expression of eukaryotic 18S rRNA was analyzed using a specific TaqMan gene expression assay (Hs99999901_s1; Applied Biosystems).
Generation of polyclonal antisera specific for MERS-CoV N and E proteins.
Rabbit polyclonal antisera (pAb) specific for MERS-CoV N and E proteins were purchased from BioGenes. In brief, peptides NTGRSVYVKFQDSKPPL (corresponding to E protein amino acids 60 to 76) and AAAKNKMRHKRTST (N protein amino acids 244 to 257) were synthesized and used to immunize two rabbits with each peptide according to the company’s standard protocol. The polyclonal antisera obtained were evaluated by enzyme-linked immunosorbent assay (ELISA) using the synthetic peptides, leading to titers ranging from 1:150,000 to 1:200,000 in all cases.
Indirect immunofluorescence assay.
Vero A66 and Huh-7 cells were grown to 80% confluence on glass coverslips and infected with the recombinant MERS-CoVs. At 48 h.p.i., cells were fixed either with 4% paraformaldehyde in phosphate-buffered saline (PBS) at room temperature for 20 min or with methanol at −20°C for 15 min. For N protein immunodetection, paraformaldehyde-fixed cells were permeabilized with 0.2% saponin in PBS containing 10% FBS for 20 min and incubated with MERS-CoV N protein pAb (dilution 1:200) in PBS containing 10% FBS at room temperature for 90 min. For E protein immunodetection, methanol-fixed cells were incubated with MERS-CoV E protein pAb (dilution 1:500) in PBS containing 10% FBS overnight at 4°C. Coverslips were washed 4 times with PBS and incubated at room temperature for 45 min with goat anti-rabbit antibody conjugated to Alexa Fluor 488 (Invitrogen) diluted 1:500 in PBS containing 10% FBS. Nuclei were stained using DAPI (4′,6′-diamidino-2-phenylindole) (1:200, Sigma). To fully inactivate the samples’ infectivity, methanol-fixed cells were treated with 4% paraformaldehyde in PBS as described above. Finally, coverslips were mounted in ProLong Gold antifade reagent (Invitrogen) and analyzed on a Leica SP5 confocal microscope. Images were acquired with the same instrument settings and analyzed with Leica software.