CD8 T Cells Control Cytomegalovirus Latency by ... - Journal of Virology

5 downloads 3213 Views 1MB Size Report
Jun 14, 2006 - the upper and lower 95% confidence limit values of slope a (determined ..... .ucla.edu/Applets.dir/WilcoxonRankSumTable.html (Ivo Dinov, Statistics On- ...... Color codes are explained by the insert legend. ...... and L. J. Picker.
JOURNAL OF VIROLOGY, Nov. 2006, p. 10436–10456 0022-538X/06/$08.00⫹0 doi:10.1128/JVI.01248-06 Copyright © 2006, American Society for Microbiology. All Rights Reserved.

Vol. 80, No. 21

CD8 T Cells Control Cytomegalovirus Latency by Epitope-Specific Sensing of Transcriptional Reactivation䌤 Christian O. Simon,1 Rafaela Holtappels,1 Hanna-Mari Tervo,1 Verena Bo ¨hm,1 Torsten Da¨ubner,1 1 1 1 Silke A. Oehrlein-Karpi, Birgit Ku ¨hnapfel, Ange´lique Renzaho, Dennis Strand,2 1 Ju ¨rgen Podlech, Matthias J. Reddehase,1* and Natascha K. A. Grzimek1 Institute for Virology, Johannes Gutenberg-University, Mainz, Germany,1 and I. Department of Internal Medicine, Medical Centre Mainz, Mainz, Germany2 Received 14 June 2006/Accepted 10 August 2006

During murine cytomegalovirus (mCMV) latency in the lungs, most of the viral genomes are transcriptionally silent at the major immediate-early locus, but rare and stochastic episodes of desilencing lead to the expression of IE1 transcripts. This low-frequency but perpetual expression is accompanied by an activation of lung-resident effector-memory CD8 T cells specific for the antigenic peptide 168-YPHFMPTNL-176, which is derived from the IE1 protein. These molecular and immunological findings were combined in the “silencing/ desilencing and immune sensing hypothesis” of cytomegalovirus latency and reactivation. This hypothesis proposes that IE1 gene expression proceeds to cell surface presentation of the IE1 peptide by the major histocompatibility complex (MHC) class I molecule Ld and that its recognition by CD8 T cells terminates virus reactivation. Here we provide experimental evidence in support of this hypothesis. We generated mutant virus mCMV-IE1-L176A, in which the antigenic IE1 peptide is functionally deleted by a point mutation of the C-terminal MHC class I anchor residue Leu into Ala. Two revertant viruses, mCMV-IE1-A176L and the wobble nucleotide-marked mCMV-IE1-A176L*, in which Leu is restored by back-mutation of Ala codon GCA into Leu codons CTA and CTT, respectively, were constructed. Pulmonary latency of the mutant virus was found to be associated with an increased prevalence of IE1 transcription and with events of IE3 transactivator splicing. In conclusion, IE1-specific CD8 T cells recognize and terminate virus reactivation in vivo at the first opportunity in the reactivated gene expression program. The perpetual gene expression and antigen presentation might represent the driving molecular force in CMV-associated immunosenescence. After resolution of productive primary infection, in particular by CD8 T cells, cytomegaloviruses (CMVs) establish lifelong latent infections in their respective hosts (for reviews, see references 29, 31, 32, 52, 75, 83–85, and 87). Reactivation of latent human CMV (hCMV) to productive, cytopathogenic infection is still a health risk in immunocompromised patients (9, 57). Hematoablative therapy of leukemias, followed by bone marrow transplantation (BMT) or hematopoietic stem cell transplantation, is associated with a risk of CMV disease resulting from reactivation of latent donor and/or recipient CMV (15, 23). Among the manifestations of CMV disease in humans, interstitial pneumonia is the most dreaded because of its high fatality rate (79). Lungs were also identified as a major organ site of murine CMV (mCMV) disease, latency, and recurrence (4, 43, 70, 78). Studies in the BALB/c mouse model of CMV infection in the BMT recipient have focused on the lungs for investigating mechanisms of immune control, latency, and reactivation (reviewed in references 25, 75, and 83). In this model, control of productive lung infection and prevention of disseminated viral pneumonia proved to be critically dependent upon the efficient reconstitution of CD8 T cells that infiltrated the lungs, confined infection to inflammatory foci, and eventually resolved * Corresponding author. Mailing address: Institute for Virology, Johannes Gutenberg-University, Hochhaus am Augustusplatz, 55101 Mainz, Germany. Phone: 49-6131-39-33650. Fax: 49-6131-39-35604. E-mail: [email protected]. 䌤 Published ahead of print on 23 August 2006.

productive infection (27, 61, 62). Viral genomes, however, were maintained in the lungs, and disseminated pulmonary CD8 T-cell infiltrates persisted after clearance of productive infection (61). That viral latency is not a static state but involves a permanent immune sensing of reactivation attempts was first suggested by the finding that a high fraction of pulmonary CD8 T cells displayed an activated effector-memory T-cell (TEM) phenotype characterized by low-to-absent cell surface expression of L-selectin CD62L and by effector function (61). Enrichment of CD8 T cells specific for the IE1 protein-derived major histocompatibility complex (MHC) class I Ld-restricted antigenic peptide 168-YPHFMPTNL-176 (76) in the CD62Llow fraction predicted expression of the ie1 gene and presentation of the IE1 peptide by latently infected lung cells (26). Studies on viral transcription in latently infected lungs have indeed revealed transcriptional activity at the major immediate-early (MIE) locus from the enhancer-flanking transcription units ie1/3 (m123/M122) and ie2 (m128) giving rise to spliced IE1 and IE2 transcripts but, notably, not to spliced IE3 transcripts (21, 41). Thus, the IE3 protein, the essential transactivator of viral Early (E) genes downstream in the productive cycle (1), was not expressed. Accordingly, the essential gene M55 (gB) was also not expressed, and latency was maintained despite MIE locus activity. MIE locus gene expression, however, is not latency associated in the sense that it is inherent to the latent state. In fact, a great majority of the latent viral genomes are silenced at the MIE locus at any moment. As

10436

VOL. 80, 2006

IMMUNE SENSING OF CYTOMEGALOVIRUS REACTIVATION

shown by a statistical analysis of transcriptional events in the lungs (21, 82), MIE locus transcription is a rare event of variegated gene expression, also known as mosaic expression (17). This expression reflects MIE locus desilencing, putatively linked to local opening of the proposed higher-order chromatin-like structure of the latent viral episome (references 5, 47, and 54; for reviews, see references 3, 83, and 84). To give an idea of the “point prevalence,” that is, the proportion of viral genomes expressing MIE genes at any moment, a statistical estimate gave ⬃20 events of MIE locus activity per 106 latent viral genomes (82, 83). Activation of the MIE enhancer through the tumor necrosis factor alpha/NF-␬B/AP-1 signaling pathways (29, 30, 82) led to an ⬃10-fold increase in the frequency of IE1 transcription and to IE3 splicing. However, as it was indicated by the absence of gB transcripts, gene expression did not proceed to complete reactivation and virus recurrence (82). Combined, these immunological and molecular findings led to the silencing/desilencing and immune sensing hypothesis of CMV latency and reactivation (83). According to this hypothesis, MIE gene expression is sensed and terminated by patrolling CD8 T cells. Although the events are of low frequency at any single time point, the cumulative incidence over a period of months can lead to a substantial activation and expansion of the CD8 T-cell pool. Furthermore, the incidence of MIE gene expression during latency may so far have been underestimated as a consequence of CD8 T-cell function having already terminated MIE gene expression in most cases. As far as we know today, IE2 and IE3 do not contain MHC class I H-2drestricted antigenic peptides; therefore, the immunodominant IE1 peptide is the only candidate for CD8 T-cell surveillance of MIE gene reactivation in the BALB/c mouse model. In support of this hypothesis, we provide here experimental evidence to conclude that MIE gene-expressing cells in latently infected lungs indeed present the IE1 peptide for CD8 T-cell sensing of viral reactivation and for intervention at the first opportunity in the viral gene expression program. MATERIALS AND METHODS Construction of plasmids. Recombinant plasmids were constructed according to established procedures, and enzyme reactions were performed as recommended by the manufacturers. (i) Shuttle plasmid for mutagenesis. pST76KIE1Ala was constructed to introduce a point mutation Ala (codon GCA) in place of Leu codon CTA at the C-terminal MHC class I anchor residue position of the IE1 peptide 168-YPHFMPTNL-176. Briefly, plasmid pUCAMB (21) containing the DNA sequence of the IE1 peptide from nucleotide positions 181,020 to 181,046 (n181,020 to 181,046) of the mCMV genome (GenBank accession no. NC_004065 [complete genome]) (68) served to generate plasmid pIE1_168176L3A. This plasmid was derived from pUCAMB by excision of a 1,005-bp HindIII/BclI fragment containing the DNA sequence corresponding to the IE1 peptide and replacement by a 1,005-bp HindIII/BclI PCR fragment containing the mutated codon GCA, specifying Ala. This fragment was generated with plasmid pUCAMB as template DNA via site-directed mutagenesis by overlap extension using PCR (24) with primers Nona-Ala-f (5⬘-CTTCATGCCCACTA ATGCAGGG-3⬘ [underlining indicates nucleotides implicated in mutagenesis]) (n181,038 to 181,017) and IE1ex4 (5⬘-ACTGCCTTAGCCAGATTCTCC-3⬘) (n180,697 to 180,717) as well as IE1ex2 (5⬘-TTTTTAGAGAGATGGAGCCCG C-3⬘) (n181,777 to 181,756) and Nona-Ala-rev (5⬘-CCCTGCATTAGTGGGCAT GAAG-3⬘) (n181,017 to 181,038). In the subsequent fusion reaction, primers IE1ex2 and IE1ex4 were used. PCR was performed with the following cycler conditions: an initial step for 5 min at 95°C for activation of ProofStart Taq DNA polymerase (catalog no. 202205; QIAGEN) followed by 30 cycles for 45 s at 94°C, 60 s at 55°C (or 60°C in the fusion PCR), and 60 s at 72°C. Finally, pIE1_168-

10437

176L3A was cleaved with BssHII, HpaI, and FspI, and a 6,523-bp BssHII/HpaI fragment was filled up by Klenow DNA polymerase and ligated blunt end into the SmaI-cleaved shuttle plasmid pST76-KSR (8). This plasmid is a derivative of shuttle plasmid pST76K (65) and contains the Bacillus subtilis sacB gene and the Escherichia coli recA gene. (ii) Shuttle plasmids for revertants. For construction of shuttle plasmid pST76KIE1Leu, pUCAMB25 was cleaved with BssHII, HpaI, and FspI, and a 6,523-bp BssHII/HpaI fragment, encompassing the IE1 peptide-encoding sequence, was filled up by Klenow DNA polymerase and ligated into the SmaIcleaved vector pST76-KSR. Shuttle plasmid pST76KIE1Leu* was constructed as follows. A 1,471-bp ApaLI fragment carrying a point mutation of nucleotide A3T at the wobble position of the Leu codon CTA was generated by mutagenesis PCR with pST76IE1Ala as template DNA and with primers rIE1-revert (5⬘-ATCTCCTGCTGCTGTTGCTGTTCTTC-3⬘) (n180,313 to 180,338) and Nona-Leu-f* (5⬘-CTTCATGCCCACTAATCTTGGG-3⬘) (n181,038 to 181,017) as well as Nona-Leu-r* (5⬘-CCCAAGATTAGTGGGCATGAAG-3⬘) (n181,017 to 181,038) and fIE1-revert (5⬘-CACAGAGGATTCTGTCTGTGTCAAGG-3⬘) (n181,968 to 181,943). The fusion reaction was performed with primers rIE1revert and fIE1-revert. pST76KIE1Ala was digested with ApaLI, and the 1,471-bp ApaLI fragment was replaced by the PCR-mutated 1,471-bp ApaLI fragment. Throughout, the fidelity of PCR-based cloning steps and cloning crossings was verified by sequencing (automated DNA sequencing system, model 4000; LI-COR Inc., Lincoln, Nebraska). (iii) Plasmids for reporter gene assays. Recombinant plasmid pIE1-L176A (for use in firefly luciferase assays; see below) was constructed as follows. Plasmid pST76KIE1Ala was first cleaved with XbaI and Eco47III and then digested with BsmI to eliminate vector sequences. The resulting 1,809-bp fragment, containing the mutated IE1 peptide L176A, was cloned into the XbaI- and Eco47IIIdigested plasmid pIE100/1 (49). This plasmid contains a genomic fragment of mCMV that encodes the authentic IE1 protein (36, 49). Recombinant plasmid pIE1-A176L* was constructed in an analogous manner, except that pST76KIE 1Leu* provided the insert sequence. Plasmids pTLG (2) and pGL3R2 1.5 (12) contain the mouse thymidylate synthase promoter and the mouse ribonucleotide reductase 2 promoter, respectively, each linked to an intronless luciferase reporter gene. Plasmid pRL-TK encoding Renilla luciferase (GenBank accession no. AF025846; catalog no. E2241; Promega) was used to standardize for transfection efficacy. (iv) Plasmid pDrive-e1 for the synthesis of E1 in vitro transcripts. A cDNA fragment of the e1 gene (11, 13, 68) was amplified by reverse transcription-PCR (RT-PCR) from RNA derived from mouse embryo fibroblasts (MEFs) infected with mCMV-WT.Smith. Oligonucleotides Early1-for1 (5⬘-GACGACGTTACTT CACCTTCCG-3⬘) and Early1-rev1 (5⬘-GAACACATTGTCCAAGTCGACC3⬘) served as forward and reverse primers, respectively. The 1,509-bp amplification product, representing the intronless mCMV e1 sequence from map positions 203 to 2,130 (GenBank accession no. M35146) (11), was cloned into plasmid pDrive by means of UA-based ligation (catalog no. 231122; QIAGEN, Hilden, Germany). For use as a template in the in vitro transcription, plasmid pDrive-e1 was linearized by digestion with SphI. Enzyme reactions were performed as recommended by the manufacturers. Synthetic transcripts were prepared according to the instructions given in the Ambion MEGAscript technical manual no. 1330. BAC mutagenesis. Mutagenesis of full-length mCMV BAC (bacterial artificial chromosome) plasmid pSM3fr (92) was performed with E. coli strain DH10B (Invitrogen) by using a two-step replacement method (6, 50, 55) with the modifications described previously by Wagner et al. (92) and Borst et al. (8). Shuttle plasmid pST76KIE1Ala was used to generate the BAC plasmid C3XIE1Ala, which contains the mutated codon GCA corresponding to the amino acid point mutation L176A in the IE1 protein and IE1 antigenic peptide sequence. To restore Leu in position 176 of IE1, E. coli DH10B carrying BAC plasmid C3XIE1Ala was used for recombination with shuttle plasmids pST76KIE1Leu and pST76IE1Leu* encompassing the Leu codons CTA and CTT, respectively. Integrity and sequence analysis of recombinant mCMV BAC plasmids. BAC plasmid DNA was isolated from small-scale cultures (catalog no. BMAX044; Epicenter) and purified (NucleoBond PC 500, catalog no. 740574.25; MachereyNagel). The overall integrity of the recombinant BAC plasmids was tested by standard methods of restriction enzyme cleavage, agarose (0.7% [wt/vol]) gel electrophoresis, and ethidium bromide staining. The point mutations in recombinant mCMV BAC plasmids C3XIE1Ala, C3XIE1Leu, and C3XIE1Leu* were verified by sequencing (model 4000; LI-COR) using the Thermo Sequenase fluorescent labeled primer cycle sequencing (7-Deaza-dGPT) kit (catalog no. RPN 2438; Amersham Bioscience). Purified BAC DNA (see above) served as the template and was sequenced in both directions with primers IE1-ex4F1-IR

10438

SIMON ET AL.

(5⬘-GAGCGTTCTGTTGTCCTGTAAG-3⬘) (n181,245 to 181,224) and IE1ex4R1-IR (5⬘-ACTGCCTTAGCCAGATTCTCCC-3⬘) (n180,697 to 180,718). Reconstitution of BAC-derived recombinant viruses. Purified DNA of the respective BAC plasmids (see above) was transfected into MEFs by using PolyFect transfection reagent (catalog no. 301107; QIAGEN). To eliminate BAC vector sequences that could attenuate viruses for growth in vivo (92), BACderived viruses were subjected to five rounds of passaging in MEF cultures followed by two rounds of plaque purification. Verified BAC vector-free virus clones (see below) were used to prepare high-titered stocks of sucrose gradientpurified viruses (43, 60) mCMV-IE1-L176A (2.3 ⫻ 108 PFU/ml), mCMV-IE1A176L (2.1 ⫻ 108 PFU/ml), and mCMV-IE1-A176L* (6.5 ⫻ 108 PFU/ml). Other viruses used in this study include wild-type mCMV-WT.Smith (ATCC VR194/1981, recently reaccessioned as VR-1399), BAC-derived mCMVWT.BAC MW97.01 (92), and the ie1 gene deletion mutant mCMV-⌬ie1 (19). Verification of the absence of BAC vector sequences. Total DNA derived from infected cells after the second round of plaque purification was prepared by using a High Pure viral nucleic acid kit (catalog no. 11858874001; Roche). To examine whether a correct excision of the BAC vector sequences from the recombinant mCMV genomes had occurred, PCRs were performed as described by Ghazal et al. (18, 19). Amplification products were analyzed by agarose gel electrophoresis, Southern blotting, and hybridization with the ␥-32P-end-labeled BAC vector sequence-specific probe 5⬘-GGATACTCAGCGGCAGTTTGC-3⬘ to detect even traces of amplification products. The mutations were verified again by sequencing as described above for BAC plasmids. Transient transfections and reporter gene assays. Plasmids (see above) were purified with the QIAGEN plasmid Maxi kit (catalog no. 12163; QIAGEN). A dual-luciferase reporter (DLR) assay (catalog no.1910; Promega Corp.) was employed to standardize for transfection efficacy. NIH 3T3 cells were plated in 60-mm-diameter dishes. On the following day, the cells (ca. 2 ⫻ 105 cells per dish) were cotransfected with 20 ng of Renilla luciferase-encoding control plasmid pRL-TK, 2.5 ␮g of firefly luciferase-encoding reporter plasmid (pTLG or pGL3R2), and 2.5 ␮g of one of the mCMV IE1 protein-encoding transactivator donor plasmids (pIE100/1, pIE1-L176A, or pIE1-A176L*) or of pUC19 as the negative control. Transfection was performed by using 15 ␮l PolyFect transfection reagent (catalog no. 301107; QIAGEN). The transactivating effects of the authentic and mutated IE1 proteins on the mouse ribonucleotide reductase 2 promoter activity (45) as well as on the thymidylate synthase promoter activity (20) was tested as follows. The transfected cells were first incubated in Dulbecco’s minimal essential medium (DMEM) (5% fetal calf serum) for 5 to 6 h, were then washed twice, and finally were incubated again in DMEM until the next day. For growth arrest, NIH 3T3 cells were then washed three times with low-serum starving medium (DMEM with 0.5% newborn calf serum) followed by an incubation in this starving medium for exactly 48 h. After this starvation period, luciferase activity was measured by the DLR assay as recommended by the manufacturer. The assays were performed with 20-␮l samples of cleared cell lysate. Luminescence, expressed as relative luminescence units (RLU), was measured with a single-sample luminometer (Lumat LB 9507; Berthold, Bad Wildbad, Germany). According to the read-inject-read format of the DLR assay, signals from firefly and Renilla luciferase were detected sequentially for each individual sample, with firefly luminescence being measured first. Linear ranges and assay backgrounds for the two luciferases were determined as suggested by the manufacturer. Background RLU were subtracted from assay RLU. Experimental BMT and establishment of latent mCMV infection. Syngeneic BMT with female BALB/cJ (H-2d haplotype) mice as bone marrow cell donors and recipients was performed as described previously (60). In brief, hematoablative conditioning of 8- to 9-week-old recipients was achieved by total-body ␥-irradiation with a single dose of 6.5 Gy. BMT was performed 6 h later by intravenous infusion of 5 ⫻ 106 femoral and tibial donor bone marrow cells. Shortly after BMT, intraplantar infection of recipients was performed at the left hind footpad with 105 PFU of the recombinant mCMVs indicated. Criteria for the definition of latency were specified previously (for a review, see reference 75) and include absence of infectivity in key organs of CMV tropism (liver, spleen, lungs, and salivary glands) as well as PCR-verified absence of viral DNA from blood (⬍1 copy per 104 leukocytes). Clearance of viral DNA from the blood is the criterion that takes longest to be fulfilled in the BMT model, usually 8 to 10 months. Animals were bred and housed under specified-pathogen-free conditions in the Central Laboratory Animal Facility (CLAF) of the Johannes Gutenberg University. Animal experiments were approved according to German federal law under permission number 177-07/021-28. Quantification of productive in vivo infection. Infection of the lungs was assessed by quantification of infectious virus in lung homogenates using a virus plaque assay (PFU assay) on MEFs with the technique of centrifugal enhancement of infectivity as described in greater detail elsewhere (60).

J. VIROL. Infection of the liver was assessed by counting of infected cells identified by immunohistological staining of intranuclear IE1 protein pp76/89 in 10 mm2 of 2-␮m liver tissue sections as described in greater detail previously (60). Growth curves were determined on the basis of three BALB/c mice tested individually per time point after an intraplantar infection. Linear regression lines in a semilogarithmic plot logN(t) ⫽ at ⫹ logN(0), where N(t) is the number of infected cells at time t after infection, a the slope of the regression line, and logN(0) its ordinate intercept, were calculated by using Mathematica Statistics LinearRegression software, version 5.1 (Wolfram Research, Inc., Champaign, Ill.). The doubling time (DT) of the number of infected cells is then log2/a. Accordingly, the upper and lower 95% confidence limit values of slope a (determined from the ellipsoidal parameter confidence region) give the 95% confidence interval of the DT. The time point of first detection of liver infection as well as its 95% confidence interval is revealed from the points of intersection between the calculated regression lines (slope a and confidence limits of a) and the line logN(t) ⫽ 0 (equals one infected cell per detection area). Quantification of viral genomes in host tissues. For molecular and statistical analysis of latency, the lungs were subdivided into 18 bona fide equally sized pieces as follows: 1 to 3, superior lobe; 4 to 6, middle lobe; 7 to 9, inferior lobe; 10 and 11, postcaval lobe; and 12 to 18, left lung. Each piece represents approximately 3 ⫻ 106 to 4 ⫻ 106 lung cells containing approximately 18 to 24 ␮g of DNA. The two pieces of the postcaval lobe were used for determining the load of latent viral DNA in two triplicates for each latently infected mouse. The tissue was minced, and DNA was extracted with a DNeasy tissue kit (catalog no. 69504; QIAGEN) as described previously (82). Latent viral genomes were quantified by real-time PCR using the TaqMan (ABI 7700) system (Applied Biosystems) and the QuantiTect SYBR green PCR kit (catalog. no. 204143; QIAGEN). A 100-ng aliquot of the DNA was added as template DNA to a reaction mixture that included the 2X QuantiTect SYBR green PCR master mix with an initial MgCl2 concentration of 2.5 mM and 1 ␮M of each primer. Primers for amplification of a 135-bp fragment of the viral gene M55 (gB) were LCgB-forw and LCgB-rev, and primers for amplification of a 142-bp fragment of the cellular pthrp gene were LCPTHrP-forw and LCPTHrP-rev, as described in greater detail previously (82). PCR was performed with the following cycler conditions: an initial 15 min at 95°C for HotStarTaq DNA polymerase activation followed by 50 cycles of 15 s at 94°C, 30 s at 62°C, and 45 s at 72°C. Data were obtained during the extension period. After amplification, melting curve analysis of the PCR products was performed by raising the temperature to 95°C, cooling down rapidly to 60°C for 30 s, and then slowly raising the temperature again to 95°C, with the fluorescence being recorded continuously. DNA from each of the two lung pieces 10 and 11 was tested in triplicate 100-ng samples. Standard curves for quantification were established by using graded numbers of linearized plasmid pDrive_gB_PTHrP_Tdy (82) as the template. For quantification of viral replication in the time course of acute infection of footpad and spleen, tissue was processed for DNA extraction essentially as described above for the lungs, except that plantar tissue was homogenized with a QIAGEN MM300 mixer mill at 30 Hz for two periods of 5 min. Real-time quantitative PCR was performed, and log-linear regression lines logN(t) ⫽ at ⫹ logN(0) as well as the DTs of viral genomes were calculated as described above. Isolation of poly(A)ⴙ RNA from the lungs. Isolation of poly(A)⫹ RNA from lung pieces was performed as described in greater detail previously (82). In essence, lung pieces shock-frozen in liquid nitrogen were homogenized and lysed in a QIAGEN MM300 mixer mill with high-salt lysis/binding buffer (␮MACS mRNA isolation kit no. 130-075-201; Miltenyi Biotec, Bergisch Gladbach, Germany). Poly(A)⫹ RNA was isolated as recommended by the manufacturer by binding to paramagnetic oligo(dT)-coated MicroBeads in MACS type ␮ columns (Miltenyi). Prior to the poly(A)⫹ RNA elution step, digestion of contaminating DNA was conducted with DNase (catalog. no. 27-0514-02; Amersham Biosciences). After the reaction was stopped and after a washing step, bound poly(A)⫹ RNA was eluted. The first drop was discarded, and the second to fourth drops were collected. Collected samples were adjusted to a volume of 75 ␮l with RNase-free water and were stored at ⫺70°C. Analysis and quantification of viral transcripts. Quantification of the viral transcripts from genes m123 (ie1), M122 (ie3), M112-113 (e1), and M55 (gB) was performed by real-time one-step RT-PCRs with the primers and probes indicated in the following paragraphs. (i) IE1 transcripts. For ie1-specific RT-PCR, probe ie1-taq1 directed against the exon 3/4 splicing junction comprised nucleotides 5⬘-6,338 to 6,328 on exon 4 and 5⬘-6,205 to 6,192 on exon 3 (GenBank accession no. L06816). Oligonucleotide 5⬘-6,393 to 6,367 served as forward primer ie1_taq_forw1, and oligonucleotide 5⬘-6,139 to 6,156 served as reverse primer ie1_taq_rev1, yielding an amplification product of 133 bp.

VOL. 80, 2006

IMMUNE SENSING OF CYTOMEGALOVIRUS REACTIVATION

(ii) IE3 transcripts. For ie3-specific RT-PCR, probe ie3_taq1 directed against the exon 3/5 splicing junction comprised nucleotides 5⬘-8,079 to 8,061 on exon 5 and 5⬘-6,205 to 6,198 on exon 3. Oligonucleotide 5⬘-6,042 to 6,061 served as forward primer HCH17, and oligonucleotide 5⬘-8,127 to 8,108 served as reverse primer HCH18, yielding an amplification product of 231 bp. (iii) E1 transcripts. For e1-specific RT-PCR, probe early1-taq-P2 directed against exon 2 comprised nucleotides 5⬘-1,070 to 1,091 (11) (GenBank accession no. M35146). Oligonucleotide e1_intron1_for1 served as the forward primer spanning the exon 1/2 splicing junction by comprising nucleotides 5⬘-944 to 955 on exon 1 and 5⬘-1,049 to 1,058 on exon 2. Oligonucleotide e1_intron2_rev1 served as the reverse primer spanning the exon 2/3 splicing junction by comprising nucleotides 5⬘-1,552 to 1,562 on exon 3 and 5⬘-1,217 to 1,225 on exon 2. This resulted in an amplification product of 200 bp. (iv) gB transcripts. For gB-specific RT-PCR, oligonucleotide 5⬘-83,175 to 83,200 (68) (GenBank accession no. U68299; complete genome) was used as probe gB-taq-2RT, oligonucleotide 5⬘-83,137 to 83,156 served as forward primer gB-taq-for2, and oligonucleotide 5⬘-83,227 to 83,207 served as reverse primer gB_taq_rev2, yielding an amplification product of 91 bp. Quantifications were performed on an ABI Prism 7700 sequence detection system (Applied Biosystems) by using dual-labeled probes containing fluorescent reporter at the 5⬘ end and a quencher at the 3⬘ end (6-carboxyfluorescein reporter and 6-carboxytetramethylrhodamine quencher system; Operon Biotechnologies Inc., Huntsville, Alabama). The corresponding in vitro transcripts IE1, IE3, and gB were generated as described previously (41) and were used as standards for the quantifications. Standard titrations ranged from 106 to 101 in vitro transcripts and were measured in duplicate. The amount of RNA in the standard titrations was adjusted to 4 ␮g with synthetic poly(A) as the carrier. Quantification of IE1, IE3, and gB transcripts in the lungs was performed for 1/10 aliquots of the yield of poly(A)⫹ RNA from a lung piece (see above). Reactions for IE1 transcripts were performed in a total volume of 25 ␮l, containing 5 ␮l of 5⫻ QIAGEN OneStep RT-PCR buffer, 1 ␮l of QIAGEN OneStep RT-PCR enzyme mix, 668 ␮M of each deoxynucleoside triphosphate, 1 ␮M of each primer, 0.26 ␮M probe, 1.5 mM additional MgCl2, and 1.32 ␮M ROX (5-carboxy-X-rhodamine) as the passive reference. The reaction mixtures for gB, IE3, and E1 transcripts were modified in that primer concentrations were 0.64 ␮M and, in the case of gB transcripts, the additional MgCl2 concentration was 3 mM. Reverse transcription was performed at 50°C for 30 min. The cycle protocol for cDNA amplification started with an activation step at 95°C for 15 min followed by 45 cycles of denaturation for 15 s at 94°C and a combined primer annealing/extension step for 1 min at 60°C. The efficiency of the RT-PCRs was ⬎90% throughout. The sensitivities of the RT-PCRs were determined by limiting dilution analysis (44) performed with graded numbers of synthetic transcripts. Samples were scored as being negative if the cycle threshold value was not ⬍50, that is, if no signal exceeding that of the water control was obtained after 50 amplification cycles. Based on the Poisson distribution function, the maximum likelihood method (16) was used to determine the most probable number (MPN) value and its 95% confidence interval from the fractions of negative samples. Confocal laser scanning analysis. MEFs were grown for 24 h on acetonecleaned glass coverslips in 24-well plates at a density of ⬃8 ⫻ 104 cells per coverslip. Centrifugal infection with mCMV was performed at a multiplicity of infection (MOI) of 4 (0.2 PFU per cell ⫻ enhancement factor of 20). After 4 h of incubation, infected MEFs were washed and fixed in 70% methanol (stored at ⫺20°C) for 90 min at 4°C. Fixed cells were stored in phosphate-buffered saline (PBS). Before use, fixed cells were incubated with 50 ␮l of blocking buffer (PBS supplemented with 0.3% Triton X-100 and 15% fetal calf serum) for 30 min at room temperature. For the double-immunofluorescence studies, each coverslip was incubated overnight in a humidity chamber with 50 ␮l of blocking buffer containing one of the following primary antibodies: mouse anti-IE1 monoclonal antibody (CROMA 101) (1:200), mouse anti-E1 monoclonal antibody (CROMA 103) (1:100), or affinity-purified rabbit anti-promyelocytic leukemia (PML) polyclonal antibody H-238 (catalog no. sc-5621, Santa Cruz, Biotech. Inc.) (1:200). After being washed with PBS, each coverslip was incubated for 1 h with the appropriate secondary antibody diluted in blocking buffer. Secondary antibodies used were an Alexa Fluor 488-conjugated goat anti-mouse immunoglobulin G (heavy plus light chains) antibody (catalog no. A11001; Molecular Probes) and an Alexa Fluor 546-conjugated goat anti-rabbit antibody (catalog no. A11010; Molecular Probes). The incubations for staining and all subsequent steps were performed in the dark. Coverslips were washed five times in PBS before they were mounted in mounting solution (catalog no. 153-6153; Panbio, Inc.) and stored at 4°C in the dark until the measurements were performed. Throughout, immunofluorescence analyses were followed by staining of the cell nuclei for 5 min at room temperature with 4⬘-6-diamidino-2-phenylindole (DAPI; Hoechst

10439

333242 dissolved in PBS). Images were acquired using a Zeiss LSM-510 laser scanning microscope and Zeiss software. For the counting of intranuclear PML bodies and exclusion of extranuclear signals, three-dimensional scanning of nuclei was performed with a 100⫻ lens and a Z-step size of 0.3 ␮m. Immunological assays. An enzyme-linked immunospot (ELISPOT) assay was used to detect the sensitization, and consequent gamma interferon (IFN-␥) secretion, of CD8 T cells by viral epitopes. Epitopes were presented after exogenous loading of stimulator cells (P815 mastocytoma cells, H-2d haplotype) either with the indicated concentrations of synthetic peptides or with naturally processed peptides contained in high-performance liquid chromatography (HPLC) fractions of lysates of infected cells. Alternatively, epitopes were presented after endogenous antigen processing in stimulator cells (BALB/c MEFs, H-2d haplotype) infected with the indicated recombinant viruses. Second-passage MEFs were infected under conditions of centrifugal enhancement of infectivity with 0.2 PFU per cell, which corresponds to an MOI of 4 (60). To optimize the presentation of the IE1 peptide, viral gene expression was arrested in the IE phase, and MIE gene expression was enhanced by infection in the presence of cycloheximide (50 ␮g/ml) for 3 h that was then replaced by actinomycin D (5 ␮g/ml) (60). Custom peptide synthesis in a 1-mg scale and purification to ⬎75% was performed by JERINI Bio Tools GmbH (Berlin, Germany). For a list with the amino acid sequences of the here tested and currently known MHC class I H-2d-restricted antigenic peptides, see reference 25 or reference 69. HPLC fractionation of lysates from productively infected MEFs was performed as described elsewhere (reference 28 and references therein). The ELISPOT assay was performed as described previously (references 28 and 56 and references therein) with 105 stimulator cells per assay culture and with graded numbers of effector cells seeded in triplicate, except for the analysis of HPLC fractions that was performed with a constant number of 4,000 effector cells. Specifically, effector cells were either cells of an IE1 epitope-specific cytolytic T-lymphocyte (CTL) line, referred to as IE1-CTLL and characterized as shown previously (56), or immunomagnetically purified memory CD8 T cells derived from the spleens or lungs (26, 27, 60). After 18 h of cocultivation, plates were developed and spots were counted. Frequencies of IFN-␥ secreting and spot-forming effector cells and the corresponding 95% confidence intervals were calculated by interceptfree linear regression analysis as described in greater detail previously (56). Cytolytic activity of cells from CTL lines IE1-CTLL and m164-CTLL (28, 56) was measured in triplicates in a standard 4-h 51Cr release assay at an effectorto-target cell ratio of 15,000 effector cells to 1,000 51Cr-labeled target cells. Target cells were P815 mastocytoma cells pulsed with graded volumes of HPLCfractionated cell lysate containing naturally processed peptides. Frequency estimation of transcriptional events in lung tissue. Lungs were subdivided into 18 pieces (see above). Pieces 1 to 9 of the three lobes of the right lung and pieces 12 to 18 of the left lung, altogether 16 lung tissue pieces per mouse, were tested piece by piece with quantitative RT-PCRs for the presence and quantities of IE1, IE3, and gB transcripts. Based on the Poisson distribution function (44), frequencies of transcriptional events, also referred to as foci of transcription, were estimated by the maximum likelihood method (16) from the fractions of lung pieces that were negative for the respective type of transcript. For details of the calculations, see the methodological appendix of reference 21. Significance analysis. Two independent sets of data with sample sizes n1 and n2, for instance loads of latent virus genomes in two groups of mice infected with different viruses (nmut, sample size for mutant virus; nrev, sample size for revertent virus), were compared by using distribution-free Wilcoxon-Mann-Whitney (rank sum) statistics. A calculator is provided on the Web site http://www.socr .ucla.edu/Applets.dir/WilcoxonRankSumTable.html (Ivo Dinov, Statistics Online Computational Resources, UCLA Statistics, Los Angeles, California). Samples are not significantly different if the P value is ⬎0.05 (two-tailed test).

RESULTS Working hypothesis and rationale of the experimental approach. According to the silencing/desilencing and immune sensing hypothesis of CMV latency and reactivation (see the introduction), the presentation of the IE1 peptide by latently infected lung cells that have entered transcriptional reactivation restimulates patrolling IE1 epitope-specific pulmonary memory CD8 T cells, which thereby acquire an activated TEM state and interrupt the reactivation. It is important to note that all attempts to detect the IE1 protein or the IE1 peptide in latently infected lungs have failed because of the very low

10440

SIMON ET AL.

J. VIROL.

FIG. 1. Construction and verification of recombinant viruses. (A) Maps of the mutagenesis, drawn to scale. The HindIII physical map of the mCMV Smith strain genome is shown at the top with the ie1/3 transcription unit expanded to reveal the location of the authentic antigenic IE1 peptide 168-YPHFMPTNL-176 coding region (WT/Revertant). By means of site-directed mutagenesis, the C-terminal MHC class I anchor residue Leu of the IE1 peptide was point mutated to Ala by mutating the Leu codon CTA into the Ala codon GCA (Mutant). Authentic revertant and wobble revertant (*) were generated by back-mutation of Ala codon GCA into Leu and Leu* codons CTA and CTT. The 6,523-bp BssHII/HpaI fragment comprises the ie1/3 transcription unit with the authentic or mutated IE1 peptide-coding region used in the shuttle plasmids for recombination. (B) Structural analysis of BAC plasmids. Purified DNA of BAC plasmids pSM3fr (lanes 1), C3XIE1Leu (lanes 2), C3XIE1Ala (lanes 3), and C3XIE1Leu* (lanes 4) was subjected to cleavage by EcoRI, HindIII, and XbaI, and fragments were analyzed by agarose gel electrophoresis and staining with ethidium bromide. Lanes M show the size markers. (C) Sequence analysis of mutated BAC plasmids. The fidelity of the sequences is shown between n181,011 and 181,028 for the BAC plasmids C3XIE1Ala (mutant L176A), C3XIE1Leu (revertant A176L), and C3XIE1Leu* (wobble revertant A176L*).

frequency of MIE transcription of ⬃10 in the whole organ consisting of ⬃60 million cells. Therefore, a genetic approach was required. The hypothesis postulates that functional deletion of the antigenic IE1 peptide abolishes the immune sensing

of MIE gene reactivation. As a consequence, the transcriptional image of latently infected lungs should be altered to more cells expressing MIE genes and, possibly, to a progression of viral gene expression to beyond that of the MIE genes.

VOL. 80, 2006

IMMUNE SENSING OF CYTOMEGALOVIRUS REACTIVATION

10441

We have chosen the reverse genetics approach of constructing a recombinant virus that does not encode the antigenic IE1 peptide. The idea to use a virus with a knockout of the whole exon 4 (m123) of the ie1/3 transcription unit (exons 2, 3, and 4 specifying the IE1 protein and exons 2, 3, and 5 specifying the IE3 protein; see references 36 and 49) was immediately dismissed, because the IE1 protein is known to be involved in the disruption of PML nuclear bodies (19, 90) and in virulence in vivo (19). The aim, therefore, was to selectively eliminate IE1 antigenicity with a minimized influence on IE1 protein structure and function. Soon after the identification of the IE1 peptide sequence 168-YPHFMPTNL-176 (76), systematic replacements of amino acid residues by Ala had identified Pro169 in position 2 of the peptide and Leu176 at its Cterminal position 9 as anchor residues (74) required for efficient binding into hydrophobic pockets of the presenting MHC class I molecule Ld (for a review of MHC binding motifs, see reference 67). Specifically, when measured in a cytolytic assay with the prototype CTL clone IE1 (71), the substitution of Ala for Leu176 led to a reduction of peptide antigenicity by 6 log10 of molar peptide concentration (74). The substitution of other residues for Leu176 showed that Phe can replace Leu without loss of function; that Met, Tyr, and Ile reduce antigenicity by 1 log10 (Met) to 4 log10 (Tyr and Ile); and that Val, Asn, and Trp destroy antigenicity by ⬎6 log10 (80). Notably, removing the branching of the side chain at ␥C by replacing the ␤C isopropyl group in Leu by a ␤C propyl group in the synthetic Leu isomer Nleu was found to enhance antigenicity by 1 log10 (72). Among the tested substitutions that caused a loss of function (Ala, Val, Asn, and Trp), Ala, in which the hydrophobic side chain is shortened by just one isopropyl group, appeared to be the most conservative modification. The alternative strategy of deleting the core epitope 170-HFMPT-174, which interacts with the T-cell receptor (TCR) (74, 76), a strategy shown previously to destroy IE1 antigenicity and immunogenicity in the respective vaccinia virus recombinant VacV-mCMV-IE1-⌬HFMPT (14), was dismissed, because the deletion of this core epitope or its replacement by a penta-Ala string creates more of a risk of affecting protein function than a point mutation does. The virus mutant mCMV-IE1-L176A and two revertant viruses were generated by BAC mutagenesis using the two-step replacement method (for a review, see reference 10). As outlined in Fig. 1A, the replacement of Leu by Ala was achieved by mutating codon CTA into GCA. Revertant virus mCMVIE1-A176L was generated by back-mutation of codon GCA into CTA. Finally, revertant virus mCMV-IE1-A176L* was generated by mutation of codon GCA into CTT, leaving a single nucleotide exchange in the wobble position of the Leu codon as a genetic marker. The three viruses were tested for genomic structural integrity by restriction enzyme cleavage FIG. 2. Influence of the Leu codon mutation on viral transcription. Throughout, MEFs were infected in six-well culture dishes at an MOI of 4 (0.2 PFU/cell ⫻ 20; centrifugal enhancement of infectivity) for transcriptional analysis. Closed and open symbols indicate infection with revertant virus mCMV-IE1-A176L and with mutant virus mCMVIE1-L176A, respectively. Ordinate values represent copies per 50 ng of total RNA. Shown are the data from triplicate cultures. The schedules for inhibitor treatment (CH, 50 ␮g of cycloheximide per ml; ActD, 5 ␮g of actinomycin D per ml) and virus (V) infection are indicated. Time zero is defined as the end of the 30-min period of centrifugal infection. (A) MIE transcription in the absence of protein synthesis. MEFs were infected in the continuous presence of CH, and spliced IE1

(circles) and IE3 (squares) transcripts were quantified by real-timeRTPCRs at the indicated time points after infection. (B) Stability of IE1 transcripts. Infection was performed for 3 h in the presence of CH. At the indicated time points after the replacement of CH by ActD, IE1 transcripts were quantified by real time RT-PCR. (C) Viral transcription in absence of metabolic inhibitors. At the indicated time points after infection, IE1 (circles), IE3 (squares), and E1 (diamonds) transcripts were quantified by real-time RT-PCRs.

10442

SIMON ET AL.

FIG. 3. Influence of the L176A mutation on the IE1-dependent transactivation of cellular promoters. Transactivation of reporter genes by the authentic IE1 protein or the mutated IE1 protein was measured by a firefly-Renilla DLR assay. NIH 3T3 cells were cotransfected with reference plasmid pRL-TK encoding Renilla luciferase for standardization of transfection efficacy (not depicted), IE1 proteinencoding donor plasmid, and a reporter plasmid carrying a promoter of interest driving the expression of firefly luciferase. Donor plasmids were pIE100/1, specifying the authentic IE1 protein; pIE1-L176A, specifying the mutated IE1 protein; and pIE1-A176L*, specifying the rescued IE1 protein. Cotransfection with pUC19 in place of the donor plasmid served as a vector control. (A) Transactivation of the ribonucleotide reductase 2 promoter (PR2) by the IE1 protein. Plasmid pGL3R2 1.5 served as the reporter plasmid. (B) Transactivation of the thymidylate synthase promoter (PTS) by the IE1 protein. Plasmid pTLG served as the reporter plasmid. Plasmid maps (not drawn toscale) are illustrated. MIEPE, major immediate-early promoter and

J. VIROL.

patterns (Fig. 1B), and the fidelity of the mutations was confirmed by sequencing (Fig. 1C). The nucleotide mutations in the MHC anchor residue codon have no influence on MIE transcription rate, MIE transcript splicing, or stability of IE1 transcripts. A theoretical possibility to be considered is an influence of the nucleotide exchanges in the Leu codon CT/UA of ie1/3 transcription unit exon 4 leading to the Ala codon GCA on the transcription rate, differential splicing of the precursor transcript to IE1 and IE3 transcripts, and stability of the IE1 transcript, for instance through conformational alteration. To exclude trans-effects of the autostimulatory and autorepressive IE1 and IE3 proteins, respectively, MEFs were infected in the presence of the protein synthesis inhibitor cycloheximide with mutant virus mCMV-IE1-L176A and revertant virus mCMV-IE1-A176L, and spliced transcripts were quantified by specific real-time RT-PCRs. While the amount of IE1 transcripts exceeded the amount of IE3 transcripts, there was no significant difference between the two viruses in this respect (Fig. 2A). Thus, the mutation in ie1 exon 4 did not affect transcription rate and differential splicing. For comparing the stabilities of the authentic and mutated IE1 transcripts, the declines in their amounts were monitored in a time course after replacing cycloheximide with the transcription inhibitor actinomycin D to avoid replenishment through new transcription (Fig. 2B). Although in this particular experiment the amount of IE1 transcripts accumulated in the presence of cycloheximide was a bit higher for the revertant virus, probably due to minor experimental variance in the MOI, the decay rates of the IE1 transcripts were apparently not different for the two viruses. Functional integrity of the mutated IE1 protein. (i) The mutation L176A does not affect the regulatory function of the IE1 protein in viral gene expression. Functions of the mCMV IE1 protein that are pertinent to viral transcription are the autoactivation of the MIE promoter and the cotransactivation, in cooperation with the IE3 protein, of the e1 promoter (49). We have therefore compared the amounts of IE1, IE3, and E1 transcripts generated after infection of MEFs with mutant virus mCMV-IE1-L176A and revertant virus mCMV-IE1A176L in the absence of metabolic inhibitor. As shown in Fig. 2C, the mutation L176A had no noticeable influence on the expression kinetics and amounts of these three viral transcripts. (ii) The mutation L176A does not affect transactivation of cellular promoters. While the IE3 protein is the essential transactivator of mCMV E gene expression (1, 49), IE1 was found to be relevant for the efficacy of virus replication in host tissues (19), a finding that possibly is related to the intrinsic property of IE1 to transactivate promoters of cellular genes of

enhancer. Ex, exon. Arrows indicate the direction of transcription. Arrowheads mark the location of the mutation in exon 4. Luminescence data are expressed as RLU and are normalized for transfection efficacy by forming the quotient of firefly luciferase activity (RLUFL) and Renilla luciferase activity (RLURL). Gray-shaded bars represent the median values of triplicate assay cultures, and the error bars indicate the range.

VOL. 80, 2006

IMMUNE SENSING OF CYTOMEGALOVIRUS REACTIVATION

the nucleotide metabolism, such as the ribonucleotide reductase gene (45) and the thymidylate synthase gene (20), in resting cells. Luciferase reporter gene assays with donor plasmids encoding the nonmutated (pIE100/1), the mutated (pIE1-L176A), or the back-mutated (pIE1-A176L*) IE1 protein and with reporter plasmids, in which firefly luciferase gene expression is driven by promoters PR2 of the ribonucleotide reductase gene (Fig. 3A) and PTS of the thymidylate synthase gene (Fig. 3B), were employed to test the functional integrity of the mutated IE1 protein. In essence, the transactivating activity of the IE1 protein was not affected by the L176A mutation. (iii) The mutation L176A does not affect the dispersion of nuclear domains ND10. Another noted function of the IE1 protein is its capacity to disperse nuclear domains ND10, also known as PML protein-associated nuclear bodies or, briefly, PML bodies. The involvement of these structures in the cell’s epigenetic defense against infection by silencing the MIE promoter has been discussed previously (for a review, see reference 91). In accordance with data obtained by Ghazal et al. (19), deletion of gene ie1 in mutant virus mCMV-⌬ie1 prevented the disruption of PML bodies in infected cell nuclei (Fig. 4A, panels a to d) but did not prevent the progression of viral gene expression; this became obvious from the detection of the E-phase proteins E1 (M112-113) (11, 13) in the cell nuclei (Fig. 4A, panel c). In contrast, PML bodies were found to be dispersed in the nuclei of cells infected with mutant virus mCMV-IE1-L176A (Fig. 4A, panels e to h). As indicated by intranuclear staining of the mutated IE1 protein, the mutation did not interfere with the nuclear localization of IE1 (37) and did not affect the CROMA 101 antibody epitope (Fig. 4A, panel g). Quantification of intranuclear PML bodies did not reveal any significant differences between mCMV-WT.BAC, mutant mCMV-IE1-L176A, and revertant mCMV-IE1-A176L (Fig. 4B). In accordance with the data obtained by Ghazal et al. for NIH 3T3 cells (19), the number of intranuclear PML bodies in uninfected MEFs and that in MEFs infected with mCMV-⌬ie1 (n1 ⫽ 30, n2 ⫽ 30; P ⫽ 0.81; Wilcoxon-Mann-Whitney test; two-tailed) were found to be identical at an early stage of infection. We certainly have taken into consideration that these assays may not represent all functions performed by the IE1 protein, but the combined results give reasonable evidence to suggest that the IE1 protein is not severely altered in its key functions by the L176A mutation. The mutation L176A does not affect virus replication and dissemination in the immunocompromised host. In accordance with the functional integrity of the IE1 protein and the unaltered transactivation of E-phase transcription shown above for the example of E1, virus replication in cell culture was not affected by the mutation. Specifically, the genome-toinfectivity ratios, which for sucrose gradient-purified monoand multicapsid virions of mCMV is ⬃2-fold the particle-toinfectivity ratio (43), were found to be identical for mutant virus mCMV-IE1-L176A and revertant virus mCMV-IE1A176L (317 to 683 genomes/PFU and 320 to 555 genomes/ PFU, respectively; nmut ⫽ 12, nrev ⫽ 11; P ⫽ 0.32; WilcoxonMann-Whitney test; two-tailed). To test whether this applies also to virus replication in host tissues under conditions of immunoablation by a 7-Gy total-

10443

body ␥-irradiation, the virus growth curves were determined for the local site of subcutaneous infection, i.e., the heterogeneous tissue of the footpad, as well as for distant organ sites, exemplified for the spleen and the liver (Fig. 5). The virus growth rates, expressed as doubling times of viral genomes in the case of plantar tissue and spleen and as doubling times of infected cells, mainly hepatocytes, in the liver, revealed no significant differences between the wild-type virus, the mutant virus, and the two revertant viruses. It is interesting to note that with values in the ranges of 22 to 46 h (plantar tissue), 10 to 15 h (spleen), and 13 to 23 h (liver), the doubling times were much longer than one would expect from the virus productivity of an infected cell. These findings suggest that cell-to-cell spread in tissue is a rate-limiting step. All four viruses initiated liver infection between days 3 and 4. Thus, the mutation had no significant influence on the capacity of the virus to disseminate from the local site of intraplantar infection to the liver. The mutation L176A results in the intended immunological loss-of-function phenotype. As explained above, the rationale for the L176A mutation was based on the antigenicity of synthetic IE1 peptide analogs with C-terminal amino acid substitutions, tested previously with the IE1-CTL prototype clone (71, 80). It was also known, however, that terminally truncated IE1 peptide fragments that lack one or even both anchor residues bind to the MHC class I molecule Ld and trigger IE1-CTL function if exogenously loaded as synthetic peptides at very high molar concentrations (76). Accordingly, when the antigenicity of synthetic peptide 168-YPHFMPTNA-176 was retested in an IFN-␥-based ELISPOT assay with a still-polyclonal and highly affine IE1-CTL line that comprised TCRs with broad V␤ usage (56), the antigenicity was found to be reduced by ⬃5 log10 of molar peptide concentration, but at concentrations of ⬎10⫺8 M the epitope was sufficiently presented and high-affinity IE1-CTLs in the polyclonal population were sensitized (Fig. 6A). However, exogenous peptide loading onto MHC class I molecules does not predict the presentation of the corresponding naturally processed peptide. Presentation of peptides in infected cells depends on the efficacies of proteasomal processing, precursor peptide transport into the endoplasmic reticulum, N-terminal trimming, peptide loading on nascent MHC class I molecules, and transport of the peptideloaded complexes to the cell surface. In addition, the latter step is also modulated by immunoevasins, viral negative regulators of antigen presentation (for reviews, see references 22, 51, 59, 69, and 77). Thus, a presentation that is equivalent to exogenous loading with ⬎10⫺8 M of peptide, as required for recognition of the IE1-L176A peptide analog, may never occur in infected cells. In addition, it is questionable whether the proteasome cleaves at all after position Ala176 to generate the mutated peptide. We used two approaches to verify an immunological loss-of-function phenotype of the L176A mutation under conditions of infection, namely, a test for IE1-specific antigenicity (Fig. 6B) and a test for IE1-specific immunogenicity (Fig. 6C) of mutant virus mCMV-IE1-L176A. For testing antigenicity (Fig. 6B), MEFs were infected under conditions of enhanced and arrested expression of the MIE genes achieved by the classical cycloheximide and actinomycin D metabolic inhibitor protocol. Thus, inhibitory functions of E-phase immunoevasins on peptide presentation were avoided. A highly affine, polyclonal IE1 epitope-specific CTL

10444

SIMON ET AL.

J. VIROL.

FIG. 4. Dispersion of nuclear domains ND10 by the mutated IE1 protein. MEFs were left uninfected (Ø) or were infected at an MOI of 4 (0.2 PFU/cell ⫻ 20; centrifugal enhancement of infection) with the BAC-derived mCMVs indicated. The analysis was performed at 4 h after infection corresponding to an early stage in the E phase. (A) Confocal laser scanning images. (a and e) Green staining for PML protein. Note that red Alexa Fluor 546 fluorescence was electronically converted into green for better contrast. (b and f) Blue counterstaining of DNA in the cell nucleus with DAPI. (c and g) Red staining for intranuclear mCMV proteins E1 and IE1, respectively. Note that green Alexa Fluor 488 fluorescence was electronically converted into red. (d and h) Merge of blue DAPI and green PML staining. Panels a to d and e to h each show an individual, representative cell after double labeling. Note the presence of green-stained intranuclear PML bodies in panels a and d in a cell infected with the ie1 gene deletion mutant and their absence in panels e and h in a cell infected with mutant mCMV-IE1-L176A. The bar marker represents 20 ␮m. (B) Quantification of intranuclear PML bodies. For statistical significance analysis, intranuclear PML bodies were counted for 30 infected cell nuclei per group. Dots represent the number of PML bodies in individual cell nuclei. The median values are marked by horizontal bars. P values are indicated for comparisons of major interest.

line, the same line as used in the experiment shown in Fig. 6A, served for monitoring the antigenicity of the infected MEFs. The results are absolutely clear in showing that cells infected with the mutant virus are not recognized, whereas the IE1 epitope is presented on cells infected with the two wild-type viruses mCMV-WT.Smith and the BAC-cloned mCMVWT.BAC as well as on those infected with the two revertant viruses mCMV-IE1-A176L and mCMV-IE1-A176L*. For testing immunogenicity (Fig. 6C), BALB/c mice were

primed by subcutaneous, intraplantar infection with revertant virus mCMV-IE1-A176L* and with mutant virus mCMV-IE1L176A. The capacities of the two viruses to prime an IE1 epitope-specific CD8 T-cell response with subsequent generation and maintenance of a memory CD8 T-cell pool were assessed by determining the frequencies of spleen-derived memory CD8 T cells specific for the IE1 epitope. As shown in the top panel of Fig. 6C, ⬃1% of the total CD8 T cells were specific for the IE1 epitope after priming with the revertant

VOL. 80, 2006

IMMUNE SENSING OF CYTOMEGALOVIRUS REACTIVATION

10445

FIG. 5. Virus replication and dissemination in vivo. Virus growth curves are shown as log-linear plots of infection load on the ordinate and time after infection on the abscissa. Infection loads in plantar tissue, spleen, and liver were determined at the indicated time points after subcutaneous, intraplantar infection of 7-Gy total-body ␥-irradiated BALB/c recipients with 105 PFU (corresponds to ⬃5 ⫻ 107 viral genomes) of the BAC-derived mCMVs indicated. In plantar tissue and in the spleen, the viral loads per 106 tissue cells were quantified by real-time PCR specific for viral gene M55 (gB) normalized to the cellular gene pthrp. Infection of the liver was quantified by counting infected liver cells, which are mostly hepatocytes and some endothelial cells, in representative 10-mm2 areas of tissue sections. Infection of cells was identified by immunohistological staining of intranuclear IE1 protein. Dots represent data from three individual mice per time point, with median values marked. The DTs (95% confidence intervals of DTs are in parentheses) were calculated by log-linear regression analysis. Note that DNA load present on day 1 in plantar tissue (arrows) mostly represents DNA of the virus inoculum. Accordingly, data for day 1 in plantar tissue were excluded from the regression analysis of virus replication. The time point of first detection of infected liver cells is revealed by the intersection between the calculated regression line and ordinate 0 (corresponds to one infected cell per test area), which was between days 3 and 4 for all viruses tested. The horizontal bars indicate the corresponding 95% confidence intervals.

virus. This is revealed by their sensitization through stimulator cells presenting exogenously loaded synthetic IE1 peptide 168YPHFMPTNL-176 or its functional analog, 168-YPHFMPTN F-176. By contrast, as shown in the bottom panel of Fig. 6C, infection with the mutant virus failed to generate IE1 epitopespecific CD8 T cells. That this failure does not reflect a generally low priming efficacy of the mutant virus becomes evident from the successful generation of memory CD8 T cells specific for the coimmunodominant MHC class I Dd-restricted epitope 257-AGPPRYSRI-265 that is derived from the ORFm164 protein (28), meanwhile characterized as glycoprotein gp38/50 expressed in the E phase (T. Da¨ubner and S. A. OehrleinKarpi; manuscript in preparation). The observed frequency of m164 epitope-specific memory CD8 T cells of ⬃2%, compared to ⬃1% with the revertant virus, suggests that the m164-specific CD8 T-cell response profits from the absence of IE1specific priming. In conclusion, in accordance with the mutagenesis rationale, the point mutation of the C-terminal MHC class I anchor residue of the IE1 peptide in mutant virus mCMV-

IE1-L176A has abolished both IE1-specific antigenicity and immunogenicity. Functional deletion of IE1 antigenicity and immunogenicity has no influence on clearance of productive infection in the lungs during hematopoietic reconstitution. As shown above, a role for the IE1 epitope in the immunological control of mutant virus mCMV-IE1-L176A is precluded for two reasons. First, host cells infected with this virus fail to present the IE1 epitope, and second, this virus fails in priming an IE1-specific CD8 T-cell response. It was clear, however, that this deletion of IE1 antigenicity and immunogenicity would not abolish CD8 T-cell control of virus replication, because security backup is provided by the coimmunodominant m164 peptide as well as by a series of subdominant antigenic peptides known to elicit protective CD8 T cells (for a review, see reference 25). Yet, it remained a reasonable prediction that functional deletion of one out of two immunodominant epitopes will at least reduce the efficacy of the immune control, in particular under conditions of hematoablative treatment and BMT, where the control of productive infection in the BMT recipients by hematopoi-

10446

SIMON ET AL.

J. VIROL.

etic reconstitution of antiviral CD8 T cells is a race against time (25). Admittedly, we were therefore somewhat surprised by the result that peak replication levels as well as the time courses of clearance of productive infection in the lungs of BMT recipients were essentially alike for mutant virus mCMVIE1-L176A and revertant virus mCMV-IE1-A176L (Fig. 7A). From the virus titer data, this conclusion becomes obvious at a glance, and it is also substantiated by Wilcoxon-Mann-Whitney statistics confirming that virus titers in the lungs were not significantly different between the two groups of BMT recipients at individual time points (except for a marginal significance at 4 weeks) as well as during the whole time course (nrev ⫽ 50; nmut ⫽ 50; P ⫽ 0.23; two-tailed). For both viruses, productive infection was found to be resolved at ⬃8 months after BMT and primary infection. The mutation L176A in the IE1 protein has no impact on the load of latent viral genomes in the lungs. As we have shown in previous reports, the level and duration of productive infection determine the load of latent viral genomes in the case of mCMV-WT.Smith (70, 86). Accordingly, the identical time courses of and magnitudes of productive infection by mutant and revertant viruses predicted similar genome loads in latency. On the other hand, the proposed differential control of latency of the two viruses by IE1-specific CD8 T cells could have an impact upon the load in the long term. We determined the viral genome loads for both viruses during pulmonary latency at 1 year after BMT and infection (Fig. 7B). Although there existed load variances between individual mice and even within individual mice between two neighboring pieces of the postcaval lung lobe used for load determination, the loads were of the order of magnitude of 104 viral genomes per 106 lung cells for both viruses. Wilcoxon-Mann-Whitney statistics substantiate the conclusion that the viral genome loads during latency were not significantly different between the two viruses (nrev ⫽ 29; nmut ⫽ 30; P ⫽ 0.49; two-tailed). Frequencies of epitope-specific CD8 T cells in latently infected lungs. The immunological situations in latently infected lungs were compared for the two viruses at 1 year after BMT and infection (Fig. 8). During latency of the revertant virus mCMV-IE1-A176L, the mCMV-specific fraction of the pulmonary CD8 T-cell pool was dominated by the epitopes IE1 and m164 (Fig. 8A). This was concluded from the frequencies of pulmonary CD8 T cells specific for all currently known

FIG. 6. Immunological loss-of-function phenotype of the L176A mutation. Throughout, gray-shaded bars represent the frequencies of CD8 T cells that were successfully sensitized for IFN-␥ secretion in ELISPOT assays. MPN values were determined by intercept-free linear regression analysis of data (spot counts) obtained for graded numbers of effector cells, each assayed in triplicate cultures. Error bars represent the corresponding 95% confidence intervals. (A) Reduction of MHC binding affinity of the IE1 peptide by the mutation. For their use as stimulator cells, P815 (H-2d haplotype) mastocytoma cells were exogenously loaded with synthetic peptides YPHFMPTNL and YPHF MPTNA at the concentrations indicated. Ø, no peptide added. Effector

cells were CTLs (200 and 100 cells seeded) of a polyclonal IE1 epitopespecific CTL line (IE1-CTLL) representing a broad spectrum of TCR affinities. (B) Presentation of naturally processed IE1 peptides. Effector cells were CTLs (300, 200, 100, and 50 cells seeded) of the polyclonal IE1-CTLL. Stimulator cells were MEFs that were either left uninfected (n.i.) or were infected under conditions of selective and enhanced MIE gene expression (MOI, 4; cycloheximide replaced after 3 h by actinomycin D) with the viruses indicated. (C) Effect of the mutation on IE1 epitope-specific CD8 T-cell memory. Effector cells were CD8 T cells (104, 5 ⫻ 103, and 103 cells seeded) isolated from the pooled spleens of three BALB/c mice at 5 months after intraplantar infection with 105 PFU of wobble revertant virus (top panel) or mutant virus (bottom panel). Target cells were P815 cells exogenously loaded with the indicated synthetic peptides at concentrations of 10⫺8 M, except for peptide YPHFMPTNA, which was used at 10⫺6 M. Ø, no peptide added.

VOL. 80, 2006

IMMUNE SENSING OF CYTOMEGALOVIRUS REACTIVATION

10447

FIG. 7. Resolution of productive infection of the lungs and establishment of latency after bone marrow transplantation. (A) Time course of productive infection of the lungs. Infectious virus per lung was monitored by a PFU assay at the indicated time points after BMT and intraplantar infection with 105 PFU of revertant virus mCMV-IE1-A176L (closed circles) or mutant virus mCMV-IE1-L176A (open circles). Symbols represent data for individual BMT recipients with the median values indicated. P values compare mutant virus and revertant virus at each time point of analysis by the distribution-free Wilcoxon-Mann-Whitney rank sum test. p.i., postinfection. (B) Viral DNA load during pulmonary latency. Viral genomes were quantified at 1 year after BMT and infection in the two neighboring lung tissue pieces (10 and 11) of the postcaval lobe for five individual BMT recipients per group. Symbols (closed circles, revertant virus; open circles, mutant virus) represent triplicate real-time PCR data for each lung DNA preparation. Median values are indicated.

H-2d-restricted antigenic peptides of mCMV tested as synthetic peptides (Fig. 8A, left panel), as well as from the recognition of HPLC-separated naturally processed peptides that might encompass also unidentified antigenic peptides (Fig. 8A, right panel). Indeed, while HPLC fractions 24 and 29 contained the m164 peptide and the IE1 peptide, respectively (Fig. 8A, miniature inserts), unknown minor activities were detected in fractions 25 and 30. During latency of the mutant virus mCMV-IE1-L176A, the mCMV-specific fraction of the pulmonary CD8 T-cell pool consisted almost exclusively of cells specific for the m164 peptide (Fig. 8B). While the frequency of m164-specific CD8 T cells was found to be increased twofold to ⬃3% of all pulmonary CD8 T cells, subdominant epitopes apparently did not profit from the functional deletion of the IE1 peptide. Detection limit of IE1-specific quantitative RT-PCR. While the functional deletion of the IE1 peptide led to the intended immunological phenotype, the data have so far not revealed any other phenotype of the L176A point mutation in IE1. Specifically, virus replication in the lungs and the establishment of pulmonary latency were in no way affected. Yet, unless our theory of latency control by immune sensing is to be abandoned, we should see a phenotype of the mutation with regard to the transcriptional activity in latently infected lungs. To test this, it was important first to define the detection limit. The sensitivity of real-time quantitative RT-PCR specific for the spliced IE1 transcript (36) was calibrated with synthetic polyadenylated transcripts by a limiting dilution assay. As shown in Fig. 9A, eight molecules of synthetic IE1 RNA were detected after 33 to 41 (median value of 35) amplification cycles in all 48 replicates tested, whereas replicates that remained at the level of the water control after 50 amplification

cycles occurred at four molecules seeded. There was a clear distinction between negative replicates (n ⫽ 6; ⬎50 cycles) and positive replicates (n ⫽ 42; 34 to 41 cycles with a median value of 38 cycles). At one molecule seeded per replicate, 20 out of 48 replicates gave an amplification product after 35 to 48 cycles, with a median value of 40 cycles. The fractions of negative replicates exactly followed a Poisson distribution, with an MPN of 1.8 (95% confidence interval, 1.5 to 2.3) molecules required for detection (Fig. 9B). We defined the detection limit of the assay as the next integer number above the upper 95% confidence limit, that is, three mRNA molecules. In the same way, the MPNs of molecules required for detection of IE3 and M55 (gB) mRNA were found to be 1.4 (0.8 to 2.5) and 0.8 (0.5 to 1.3), respectively (data not shown). As transcription assays for latently infected lungs were performed with 1/10 aliquots of the poly(A)⫹ RNA yields of lung pieces, the detection limit for a whole piece was defined as 30 molecules in the case of IE1 as well as IE3 transcripts and as 20 molecules in the case of M55 (gB) transcripts. The mutation L176A in the IE1 protein leads to a significant increase in the point prevalence of IE1 transcription in latently infected lungs. Inherent to the impossibility of monitoring viral transcription continuously in a living being, we can view transcriptional activity only in a “snapshot” after shockfreezing of tissue. An assay for transcripts thus reveals the point prevalence of transcriptionally positive lung pieces, that is, the transcriptional activity detected at a certain moment. It is the implicit understanding that viral genomes that are silenced at the MIE locus at the time of measurement could have been active at an earlier moment or could have become active at a later moment. In any case, a significant difference in

10448

SIMON ET AL.

J. VIROL.

FIG. 8. Frequency and virus epitope specificity of memory CD8 T cells present in latently infected lungs. Pulmonary CD8 T cells were isolated at 1 year after infection of BMT recipients from pools of six lungs and were tested as effector cells in IFN-␥-based ELISPOT assays. (A) BMT and infection with revertant virus mCMV-IE1-A176L. (B) BMT and infection with mutant virus mCMV-IE1-L176A. (Left panels) Stimulator cells were P815 mastocytoma cells exogenously loaded with the indicated synthetic peptides at concentrations of 10⫺8 M. Ø, no peptide added. Bars represent MPN values determined by intercept-free linear regression analysis of data (spot counts) obtained for graded numbers (8 ⫻ 103, 4 ⫻ 103, 1 ⫻ 103) of effector cells, each assayed in triplicate cultures. Error bars represent the corresponding 95% confidence intervals. (Right panels) Stimulator cells were P815 mastocytoma cells pulsed with 60 ␮l of 800-␮l HPLC fractions containing naturally processed peptides derived from MEFs at 24 h after infection with revertant virus (for testing group A) and mutant virus (for testing group B). Bars represent the median values of triplicate data obtained with a constant number of 4,000 effector cells for testing each HPLC fraction. Error bars indicate the range. Arrowheads mark the HPLC fractions in which the immunodominant peptides m164 and IE1 eluted. Miniature inserts document the identification of the HPLC fractions that contained peptides m164 and IE1. Cytolytic activities of lines m164-CTLL and IE1-CTLL were determined at an effector-to-target cell ratio of 15:1 with 1,000 P815 target cells pulsed with 10 ␮l of 800-␮l HPLC fractions (highest concentration) or with 1:6, 1:36, or 1:216 dilutions thereof. Data represent the mean values of triplicate determinations.

point prevalence undoubtedly indicates a difference in transcriptional activities. In a first approach (Fig. 10A), the influence of the L176A mutation on transcriptional activity during latency in the lungs was tested at 1 year after primary infection for mice from the BMT experiment that was characterized in detail as shown above. Although the mutation had not resulted in any phenotype with regard to clearance of productive infection (Fig. 7A) and latent viral genome load in the lungs (Fig. 7B), its impact on MIE gene transcription during latency was striking at a glance. Whereas during latency of the revertant virus only 14 out of 80 lung pieces from five mice tested contained IE1 transcripts, 51 out of 80 lung pieces were positive during latency of the mutant. In addition, splicing to IE3 transcripts was found in 7 out of 80 pieces selectively in lungs that were latently infected with the mutant. In this group, there existed a

single piece that contained also M55 (gB) transcripts. As the amounts of IE1 and IE3 transcripts found in this particular piece were clearly far beyond the range of those found in all the other transcriptionally active pieces, this likely represented a solitary case of either persistent or reactivated productive infection. The result was essentially reproduced in an independent, second BMT experiment in which one group of BMT recipients were infected with the mutant virus and the other one with revertant virus mCMV-IE1-A176L* carrying a single nucleotide point mutation at the wobble position of the Leu codon (Fig. 10B). In this experiment, the load of latent viral genomes in the lungs was again in the order of magnitude of 104 per 106 lung cells (not shown) with no significant difference between the two viruses (nrev* ⫽ 29; nmut ⫽ 29; P ⫽ 0.52; two-tailed; Wilcoxon-Mann-Whitney statistics). As in the first BMT ex-

VOL. 80, 2006

IMMUNE SENSING OF CYTOMEGALOVIRUS REACTIVATION

FIG. 9. Detection limit of the IE1-specific real-time quantitative RT-PCR. (A) Graded numbers of synthetic polyadenylated IE1 transcripts in 48 replicates were amplified by real-time RT-PCR. Dots represent the numbers of cDNA amplification cycles (cycle threshold [Ct] values) required for detection, with the median values marked by horizontal bars. The dotted line indicates the cutoff value defined by the water control. (B) Limiting dilution (Poisson distribution) analysis based on the experimentally determined fraction of negative replicates (see panel A). The log-linear plot shows the Poisson distribution graph (calculated with the maximum-likelihood method) with its 95% confidence region shaded. The MPN (the reciprocal of the Poisson distribution parameter ␭) is revealed as the abscissa coordinate (dashed arrow) of the point of intersection between 1/e and the regression line; it gives us the number of transcripts that need to be seeded for detection. 95% CI, 95% confidence interval of MPN; P, probability value indicating the goodness of fit.

periment, the point prevalence of IE1 transcription during latency was low for the revertant virus (11 out of 96 lung pieces of six mice tested) and high for the mutant virus (52 out of 96 lung pieces). Only three lung pieces of a single mouse latently infected with the mutant virus contained IE3 transcripts, and M55 (gB) transcripts were undetectable throughout. As we have dealt with in greater detail in previous reports (21, 41, 42, 82) and reviewed recently (83), the on-or-off patterns of variegated MIE gene expression during latency follow Poisson distributions. Whereas, by definition, a negative piece does not contain a transcriptional event, positive pieces contain at least one event but may contain also more than one event. As a consequence, the frequency of transcriptional

10449

events or episodes of MIE locus desilencing is somewhat higher than the experimentally observed point prevalence of positive pieces. The frequencies can be calculated from the fractions of negative pieces by employing the Poisson distribution function (16, 21, 44). The results of both BMT experiments are summarized in Fig. 11A by showing the calculated frequencies and their corresponding 95% confidence intervals normalized to the lungs of 10 mice, that is, extrapolated to 180 pieces, including the pieces of the postcaval lobe used for determination of the DNA load. In essence, the mutation L176A resulted in a five- to sixfold increase in detectable transcriptional events in latently infected lungs. According to the hypothesis, the mutation L176A in IE1 should increase the frequency of detectable transcriptional events, i.e., the point prevalence of transcription, not by increasing the rate and incidence of transcription but indirectly, by precluding the IE1-specific CD8 T-cell-mediated extinction of MIE transcription. Quantitative real-time RT-PCR in combination with the Poisson statistics allowed us to validate the prediction by determining the steady-state amount of transcripts per MIE locus desilencing event. In this context it is important to recall our finding that IE1 transcript stability is not altered—and certainly not increased—by the mutation on the codon level (Fig. 2B). The experimentally determined numbers of transcripts (Fig. 10) were extrapolated to 180 lung pieces of 10 mice and divided by the similarly extrapolated number of transcriptional events. The result confirmed the prediction of the hypothesis, in that the amount of transcripts per event was not significantly altered by the mutation. Specifically, and in agreement with the unaltered transcription rate during productive infection (Fig. 2C), the mutation L176A did not enhance the rate of MIE gene transcription during latency (Fig. 11B). It is interesting to note that while IE3 splicing occurred with a frequency much lower than that of IE1 splicing, the numbers of IE1 and IE3 transcripts per positive event were comparable (with a slight difference seen for BMT experiment 1 but no difference seen for BMT experiment 2). This strengthens the conclusion that the observed frequencies are not biased by assay sensitivity or different amounts of transcripts. Furthermore, the numerical values were strikingly reproducible, although the two BMT experiments were performed independently in an interval of ⬃6 months and with different revertant viruses. This suggests that the number of transcripts released per MIE desilencing event is an almost constant value. DISCUSSION The concept of immune surveillance of CMV latency by IE antigen-specific CTLs was originally raised by Reddehase and Koszinowski in 1984 on the occasion of the first description of IE protein antigenicity and immunogenicity for CD8 T cells (73). We started the present work with the hypothesis that IE1 epitope-specific CD8 T cells patrolling in latently infected lungs contribute to the maintenance of viral latency by sensing and terminating viral transcriptional reactivation before the viral replication cycle is completed and recurrence of infectious virus takes place (83). Hints in support of this hypothesis were provided by the previous findings that depletion of CD8 T cells favors virus recurrence (63) and that activated IE1-

10450

SIMON ET AL.

J. VIROL.

FIG. 10. Contextual analysis of transcription in latently infected lungs. (A) First BMT experiment performed with mutant mCMV-IE1-L176A and the authentic revertant mCMV-IE1-A176L. The analysis was performed at 1 year after BMT and infection. (B) Second BMT experiment performed with mutant mCMV-IE1-L176A and the wobble revertant mCMV-IE1-A176L*. The analysis was performed at 8 months after BMT and infection. Transcripts were quantified for each of the illustrated lung pieces, for seven pieces of each left lung, and for nine pieces of each right lung, comprising superior lobe, middle lobe, and inferior lobe. Color codes are explained by the insert legend. Numbers at the pieces give the numbers of IE1/IE3/gB transcripts determined by the respective real-time quantitative RT-PCRs for 1/10 aliquots of the yield (per piece) of poly(A)⫹ RNA.

VOL. 80, 2006

IMMUNE SENSING OF CYTOMEGALOVIRUS REACTIVATION

FIG. 11. Summary of data for transcription in latently infected lungs. (A) Frequencies (point prevalences) of transcription. Based on the fraction of pieces determined to be negative for a particular type of transcript (see Fig. 10, i.e., uncolored pieces for the absence of IE1 transcripts and uncolored pieces plus green pieces for the absence of IE3 transcripts), the Poisson distribution function was employed to estimate the total number of transcriptional events. Black (for IE1) and gray-shaded (for IE3) bars represent the frequencies (MPN values and their 95% confidence intervals) extrapolated to 180 pieces of 10 lungs. (B) Numbers of transcripts present per transcriptional event. The experimentally determined numbers of transcripts in lung tissue pieces (see Fig. 10; the singular gB-positive piece shown in Fig. 10A was excluded from the calculation) were added up, multiplied with the yield factor of 10, extrapolated to 180 pieces of 10 lungs, and divided by the number of transcriptional events per 10 lungs (see panel A). Black (for IE1) and gray-shaded (for IE3) bars represent the MPNs of transcripts and their 95% confidence intervals.

specific CD8⫹ CD62Llow TEM persist and accumulate in the lungs of BMT recipients after clearance of productive infection (26, 61). This expansion of the pool of IE1-specific TEM during latency suggested a perpetual epitope-specific restimulation by IE1 peptide presentation driven by the latency-associated transcriptional activity at the MIE locus (21, 26, 41). In support of these previous data and conclusions and in agreement with mCMV latency in multiple organs (reference 70, reviewed in reference 75), Karrer and colleagues have documented a continuous expansion and increasing oligoclonality of IE1 epitopespecific TEM in various organs after infection of immunocompetent mice (34). Of interest is the fact that TEM specific for unrelated epitopes were found to expand if and only if expressed in recombinant mCMVs under the control of the MIE

10451

locus (35). The continuous restimulation and oligoclonal expansion of CMV-specific TEM during latency implies that these cells undergo many cell cycles with the consequence of telomere shortening and gradual loss of function as well as an impoverishment of the host’s memory TCR repertoire. This “chronic antigenic stress” links the immunosurveillance of CMV latency to the phenomenon of immunosenescence (for a recent review, see reference 58). While recurrence models based on the detection of infectious virus cannot easily discriminate between molecular control of latency in the latently infected cell and control of the spread of recurrent virus, a transcriptional analysis should give direct evidence for a possible immunological surveillance of the latently infected cell. Sensing of preinfectious stages of reactivated viral gene expression requires the desilencing of genes that encode antigenic proteins, the processing of these proteins, and the presentation of the corresponding antigenic peptides by MHC class I molecules at the cell surface. Linked to the still incompletely understood phenomenon of “immunodominance” (94), not all viral proteins qualify as antigens that yield immunogenic peptides for MHC class I presentation in a particular MHC haplotype. Specifically, as far as it is known today, the memory CD8 T-cell response to mCMV in the mouse strain BALB/c (H-2d haplotype) is largely focused on two dominant epitopes derived from the proteins IE1 (pp76/89) expressed in the IE phase and m164 (gp38/50) expressed in the E phase (28), whereas it is much less focused in the mouse strain C57BL/6 (H-2b haplotype) (53). Likewise, by testing the response of 33 latently infected, healthy volunteers covering a wide range of MHC (HLA) alleles, a genome-wide analysis of antigenic open reading frames (ORFs) of hCMV has shown great variance between individuals, ranging from 1 ORF to 32 ORFs recognized by CD8 T cells of an individual (89). Thus, as in the two mouse models, the response in humans can be fairly focused (as in the BALB/c mouse) or less focused (as in the C57BL/6 mouse), depending on the MHC (HLA) alleles. With relevance to the understanding of the immune sensing hypothesis, these facts imply that the proposed recognition of CMV reactivation by CD8 T cells must necessarily occur at different stages in the reactivated gene expression program depending on the individual host’s MHC genetics, which determine when reactivated gene expression reaches an antigenic ORF. The BALB/c mouse is a particularly convenient model for testing the hypothesis because of its relatively focused memory CD8 T-cell response to mCMV and because an immunodominant epitope is present already in a MIE protein. As noted above, the enhancer-regulated MIE locus must be desilenced to kick-start reactivation by the transactivator proteins IE1 and IE3. The MIE locus in hCMV plays an analogous role (for a review, see reference 48), and, notably, the hCMV IE1 protein contains CD8 T-cell epitopes for many human MHC (HLA) alleles that were represented by more than half of the study subjects in work by Sylwester and colleagues (89). Moreover, an above-average role of MIE gene products in the CD8 T-cell response to hCMV is indicated by a recognition that was found to be threefold over their coding space representation (89). With relevance to latency, MIE region transcripts were detected in CD33⫹ CD15⫹ granulocyte-macrophage progenitors latently infected with hCMV (reference 40, reviewed in reference 52). Although these hCMV latency-associated transcripts,

10452

SIMON ET AL.

briefly known as CLTs, are deviant sense and antisense MIE region transcripts with a role that is unknown to date, hCMV reactivation is likely to involve the productive-phase IE1 transcript. Thus, the BALB/c mouse may give us clues to principles of the CD8 T-cell surveillance of hCMV latency and reactivation sensing. In accordance with the hypothesis, our data have shown that functional deletion of the mCMV IE1 epitope by the amino acid point mutation L176A increases the frequency of cells in which MIE transcripts can be detected. We see two possible explanations for this phenotype of the L176A mutation. First, one might theoretically argue that the corresponding codon mutation GCA has a direct influence on the probability of desilencing of the MIE locus. However, there exists no hint in favor of such a speculation, and our experiments have not revealed any nonimmunological phenotype of this mutation. Furthermore, the remaining codon mutation CTT in the A176L* revertant virus did not interfere with the revertant phenotype of low-frequency MIE gene expression during latency. A straightforward interpretation of the data is that the absence of IE1 epitope-specific CD8 T-cell control of the mutant—and this conclusion is evidence-based (Fig. 8)—leads to a higher number of cells that remain detectable in the process of MIE transcription. According to this interpretation, the reduced frequencies of MIE gene expression during latency observed here with the two revertant viruses, as well as in previous studies with mCMV-WT.Smith (21, 41, 82), have reflected IE1-expressing cells that were not yet hit by the IE1 epitope-specific TEM at the time of shock-freezing of the tissue for analysis. Accordingly, even though the higher frequencies revealed by the mutant give us a closer estimate of the true incidence of MIE locus desilencing, this estimate may still be an underestimate because of a possible extinction of transcriptional reactivation at further immunological checkpoints kinetically located beyond the presentation of the IE1 epitope. In principle, CD8⫹ TEM could physically eliminate cells in which reactivation takes place through cytolytic activity or through death ligands inducing apoptosis. Alternatively, they could terminate viral transcription through the secretion of cytokines inhibiting viral gene expression without harming cell viability. As we have shown previously, pulmonary CD8⫹ TEM isolated during latent infection possess cytolytic activity that can be triggered through the TCR-CD3 complex (61), and their capacity to secrete IFN-␥ upon recognition of MHCpresented mCMV peptides was actually the basis of the ELISPOT assays shown here (Fig. 8). Although the effector mechanism of the CD8 T-cell control of latency is not of critical importance for the hypothesis as such and was not specifically addressed by our experiments, it must be discussed in the context of our finding that the L176A mutation in the IE1 protein was not associated with a higher load of latent viral genomes (Fig. 7B). The lack of a latent viral load phenotype of the L176A mutation is compatible with the idea of a cytokine-mediated inhibition of reactivated viral gene expression. This mechanism is discussed elsewhere for the dynamic CD8 T-cell control of herpes simplex virus neuronal latency in sensory ganglia (33), which is plausible since cytolysis of irreplaceable neurons would otherwise lead to neurological deficits. In fact, neurons appear to be actively protected against cytolysis by expression

J. VIROL.

of ligand Qa-1b engaging the inhibitory receptor CD94/ NKG2A expressed on activated TEM (88). In fibroblasts infected with mCMV, viral replication can be blocked by synergistic effects of IFN-␥ and tumor necrosis factor alpha, but this mechanism does not affect IE1 expression and instead operates during nucleocapsid assembly (46). However, unlike the situation in fibroblasts, IFN-␥ reduces MIE gene expression in bone marrow-derived macrophages (66). In this context it is important to note that macrophages are among the proposed cellular sites of mCMV latency (39, 64). Maintenance of latent viral load, however, is not a valid argument against cytolysis. A decline in the number of latent viral genomes, which one would expect as a consequence of cytolysis occurring over a period of many months during latency, could possibly be compensated for by perpetual replenishment of latently infected cells and latent viral episomes through cell division and viral genome replication. That such a mechanism is in principle possible has been documented for Epstein-Barr virus latency, where the EBNA1 protein facilitates the replication and partitioning of latent viral genomes during mitosis by tethering them to cellular chromosomes (81). Pertinent to this discussion is that our data have shown an impact of the mutation on the frequency of transcriptional events but a constant number of transcripts per transcriptional event (Fig. 11). This argues against a gradual inhibition of viral transcription and rather suggests an all-or-nothing phenomenon, as one would expect from the elimination of epitopepresenting cells. Regardless of which mechanism applies, our data have shown that termination of reactivated viral transcription by CD8 T cells involves an epitope-specific sensing of the presentation of the cognate antigenic peptide for local delivery of an effector function. As has been reported previously by Kurz and Reddehase (42), a general ablation of cellular immunity by total-body ␥-irradiation leads to a recurrence of infectious virus, but this recurrence starts at very few reactivation foci in the lungs. In fact, in that previous study, most viral genomes remained silent even under such extreme conditions of immunosuppression. Many foci of MIE reactivation still stopped at IE1 splicing despite the fact that sensing by IE1-specific CD8⫹ TEM was abolished. Some foci proceeded to IE3 splicing, only few proceeded to the expression of M55 (gB), and even fewer proceeded to the production of infectious virus. These findings clearly imply that reactivation is restricted also—and actually primarily—by mechanisms not involving CD8 T cells. Rather, latency and reactivation appear to be determined by viral gene silencing and desilencing involving viral chromatin remodeling (reviewed in references 3, 83, and 84). Obviously, complete reactivation to virus recurrence requires desilencing of all viral genes that are essential for the generation of virus progeny and, specifically, the origin of lytic replication, oriLyt, needs to be derepressed or activated for replication of the viral genome (7, 38, 93). The fulfillment of these multiple conditions appears to be an extremely rare event. It was therefore not unexpected that the singular L176A mutation in IE1 did not generally break viral latency. We propose that the role of CD8 T cells is to reduce the probability of virus recurrence by interruption of transcriptional reactivation before viral DNA replication is initiated and infectious virus is assembled. As we have shown here, selective

VOL. 80, 2006

IMMUNE SENSING OF CYTOMEGALOVIRUS REACTIVATION

10453

FIG. 12. Refined model of the silencing/desilencing and immune sensing hypothesis of CMV latency and reactivation. (A) Viral latency after infection of BALB/c (H-2d haplotype) mice with mCMV-WT or mCMV-IE1-A176L (or L*) revertants. Most latent viral genomes (symbolized as episomes associated with histones) are silenced at the MIE locus (MIE locus latency). A low incidence of local desilencing (forward arrow) leads to MIE transcription (MIE locus reactivation), splicing of IE1 transcripts (green wavy line), antigenic processing of IE1 protein, presentation of IE1 peptide 168-YPHFMPTNL-176 (green triangle) by MHC class I molecule Ld at the cell surface, and recognition by IE1 epitope-specific effector-memory CD8 T cells (IE1-TEM). Due to the restimulation, these IE1-TEM undergo clonal expansion, and delivery of effector function(s) terminates the viral reactivation. It is still open to question (see Discussion) whether IE1-expressing cells are eliminated or whether desilenced viral genomes fall back into the pool of silenced/latent viral genomes (backward arrow). As a consequence of IE1-TEM effector function at this first immunological checkpoint, the experimentally determined point prevalence of IE1 transcription (Fig. 10 and 11) necessarily underestimates the incidence of MIE locus desilencing. NC, nuclear compartment; CYC, cytoplasmic compartment; ECC, extracellular compartment. Arrowhead, position of gene m01. (B) Viral latency after infection of BALB/c mice with mutant virus mCMV-IE1-L176A. MIE locus desilencing is proposed to occur with the same incidence (forward arrow) as that during latency of mCMV-WT. However, as the mutated IE1 RNA (green wavy line with red dot) does not encode a functional IE1 peptide and thus revokes the first immunological checkpoint, MIE locus reactivation is not recognized and therefore not terminated, which results in an increased point prevalence of IE1 transcription (Fig. 10 and 11). With a reduced incidence (smaller forward arrow), transcriptional reactivation proceeds to IE3 splicing (yellow wavy line), but IE3 does not specify an MHC class I-restricted antigenic peptide for the H-2d haplotype. Accordingly, it is not recognized by CD8 T cells in the BALB/c model. It is proposed that transcriptional reactivation further proceeds with an unknown incidence (question mark symbol; smallest forward arrow) to a second immunological checkpoint located in the E phase. The m164 transcript (blue wavy line) specifies the Dd-restricted antigenic peptide 257-AGPPRYSRI-265 (blue triangle). Its recognition leads to clonal expansion of m164-TEM and to the termination of transcriptional reactivation. As a consequence of m164-TEM effector function at this second immunological checkpoint, the incidences of m164 transcription as well as of all preceding transcriptions are necessarily underestimated (backward arrow) in experiments measuring the respective point prevalences. In an extrapolation of the model, further immunological checkpoints might exist downstream of m164, although at present no corresponding peptides that would cause significant CD8 T-cell expansions during latency are known for the H-2d haplotype.

elimination of the IE1 epitope prevents interception of reactivation at the stage of IE1 expression and also allows progression to IE3 splicing in a lower frequency of MIE gene expression episodes, but progression to M55 (gB) expression does not

regularly occur. This may be explained by the existence of a second immunological checkpoint kinetically located between MIE and M55 (gB) expression. As suggested by the specificities and frequencies of CD8⫹ TEM present in the lungs during

10454

SIMON ET AL.

latent infection with the L176A mutant and the A176L revertant (Fig. 8), the E-phase glycoprotein m164 is a promising candidate. The increased frequency of m164-specific TEM observed during latency of the mutant virus is compatible with a more frequent restimulation due to a higher number of cells in which transcriptional reactivation gets the chance to proceed to m164 gene expression. However, evidence for m164 transcription in latently infected lungs is so far missing. A reason for this may be a very low, and thus not easily detected, frequency due to an elimination of most transcriptional events by m164-TEM effector function. A dual-mutant virus with Cterminal anchor residue mutations in the antigenic peptides IE1 and m164 is under construction for future investigation of this predicted second immunological checkpoint of mCMV latency. As an extension of our recent review article on the silencing/desilencing and immune sensing hypothesis (83), a refined model of CMV latency and reactivation based on the new results presented here is illustrated and explained in Fig. 12. In conclusion, our data have shown a phenotype of an MHC class I epitope deletion in CMV latency and thus provide direct evidence for an involvement of CD8 T cells in the immunosurveillance of latency by sensing of reactivated gene expression. We predict that there exist multiple immunological checkpoints of CMV latency in each individual, and that their number as well as the corresponding CMV ORFs differs between individuals. Differential ORF usage is a consequence of the MHC-determined selection of antigenic epitopes from the CMV proteome in reflection of MHC (HLA) polymorphism in the population. The perpetual gene expression and presentation of antigenic peptides during latency is proposed to be a driving force in CMV-associated immunosenescence.

ACKNOWLEDGMENTS We are indebted to previous members of the group of Ulrich H. Koszinowski, Munich, Germany, in particular to Eva Borst and Martin Messerle, for helping us with establishing the two-step replacement method of BAC mutagenesis in our laboratory. We thank Martin Messerle, Hannover, Germany, Ana Angulo, Barcelona, Spain, and Peter Ghazal, Edinburgh, United Kingdom, for the IE1 deletion mutant of mCMV. Reporter gene plasmids for the dual-luciferase assays were kindly provided by Santo Landolfo, Turin, Italy, with permission from Lars Thelander, Umea, Sweden, and Lee Johnson, Columbus, Ohio. Monoclonal antibodies CROMA 101 and CROMA 103 were generously supplied by Stipan Jonjic, Rijeka, Croatia. We thank the Confocal Laser Scanning Microscopy Core Facility of the Immunology Cluster of Excellence for assistance with image collection. This work was supported by the Deutsche Forschungsgemeinschaft, SFB 490, individual projects E2 (C.O.S., H.-M.T., B.K., A.R., M.J.R. and N.K.A.G.), E3 (R.H. and V.B.), and E4 (T.D., S.A.O.-K., D.S., and M.J.R.), as well as SFB 432, individual project A10 (J.P. and M.J.R). Further support was provided by the Ministry of Science of the Federal State Rheinland-Pfalz, Immunology Cluster of Excellence “Immunointervention.”

ADDENDUM In transactivator protein IE3 of mCMV, recent work by the group of A. B. Hill has identified two Kb-restricted antigenic peptides that cause CD8 T-cell expansion in latently infected C57BL/6 mice, for which no antigenic IE1 peptide of mCMV is currently known (53a). Combined with our data, this novel

J. VIROL.

finding predicts that the first immunological checkpoint of mCMV latency in the H-2b haplotype is IE3 in place of IE1. REFERENCES 1. Angulo, A., P. Ghazal, and M. Messerle. 2000. The major immediate-early gene ie3 of mouse cytomegalovirus is essential for viral growth. J. Virol. 74:11129–11136. 2. Ash, J., W.-C. Liao, Y. Ke, and L. F. Johnson. 1995. Regulation of mouse thymidylate synthase gene expression in growth-stimulated cells: upstream S phase control elements are indistinguishable from essential promoter elements. Nucleic Acids Res. 23:4649–4656. 3. Bain, M., M. Reeves, and J. Sinclair. 2006. Regulation of human cytomegalovirus gene expression by chromatin remodeling, p. 167–183. In M. J. Reddehase (ed.), Cytomegaloviruses: molecular biology and immunology. Caister Academic Press, Wymondham, Norfolk, United Kingdom. 4. Balthesen, M., M. Messerle, and M. J. Reddehase. 1993. Lungs are a major organ site of cytomegalovirus latency and recurrence. J. Virol. 67:5360–5366. 5. Bolovan-Fritts, C. A., E. S. Mocarski, and J. A. Wiedeman. 1999. Peripheral blood CD14⫹ cells from healthy subjects carry a circular conformation of latent cytomegalovirus genome. Blood 93:394–398. 6. Borst, E. M., G. Hahn, U. H. Koszinowski, and M. Messerle. 1999. Cloning of the human cytomegalovirus (HCMV) genome as an infectious bacterial artificial chromosome in Escherichia coli: a new approach for construction of HCMV mutants. J. Virol. 73:8320–8329. 7. Borst, E. M., and M. Messerle. 2005. Analysis of human cytomegalovirus oriLyt sequence requirements in the context of the viral genome. J. Virol. 79:3615–3626. 8. Borst, E. M., G. Posfai, F. Pogoda, and M. Messerle. 2004. Mutagenesis of herpesvirus BACs by allele replacement. Methods Mol. Biol. 256:269–279. 9. Britt, W. 2006. Human cytomegalovirus infections and mechanisms of disease, p. 1–28. In M. J. Reddehase (ed.), Cytomegaloviruses: molecular biology and immunology. Caister Academic Press, Wymondham, Norfolk, United Kingdom. 10. Brune, W., M. Wagner, and M. Messerle. 2006. Manipulating cytomegalovirus genomes by BAC mutagenesis: strategies and applications, p. 63–89. In M. J. Reddehase (ed.), Cytomegaloviruses: molecular biology and immunology. Caister Academic Press, Wymondham, Norfolk, United Kingdom. 11. Bu ¨hler, B., G. M. Keil, F. Weiland, and U. H. Koszinowski. 1990. Characterization of the murine cytomegalovirus early transcription unit e1 that is induced by immediate-early proteins. J. Virol. 64:1907–1919. 12. Chabes, A., and L. Thelander. 2000. Controlled protein degradation regulates ribonucleotide reductase activity in proliferating mammalian cells during the normal cell cycle and in response to DNA damage and replication blocks. J. Biol. Chem. 275:17747–17753. 13. Ciocco-Schmitt, G. M., Z. Karabekian, E. W. Godfrey, R. M. Stenberg, A. E. Campbell, and J. A. Kerry. 2002. Identification and characterization of novel murine cytomegalovirus M112-113 (e1) gene products. Virology 294:199– 208. 14. Del Val, M., H.-J. Schlicht, H. Volkmer, M. Messerle, M. J. Reddehase, and U. H. Koszinowski. 1991. Protection against lethal cytomegalovirus infection by a recombinant vaccine containing a single nonameric T-cell epitope. J. Virol. 65:3641–3646. 15. Emery, V. C. 1998. Relative importance of cytomegalovirus load as a risk factor for cytomegalovirus disease in the immunocompromised host. Monogr. Virol. 21:288–301. 16. Fazekas de St. Groth, S. 1982. The evaluation of limiting dilution assays. J. Immunol. Methods 49:R11–R23. 17. Fiering, S., E. Whitelaw, and D. I. K. Martin. 2000. To be or not to be active: the stochastic nature of enhancer action. Bioessays 22:381–387. 18. Ghazal, P., M. Messerle, K. Osborn, and A. Angulo. 2003. An essential role of the enhancer for murine cytomegalovirus in vivo growth and pathogenesis. J. Virol. 77:3217–3228. 19. Ghazal, P., A. E. Visser, M. Gustems, R. Garcia, E. M. Borst, K. Sullivan, M. Messerle, and A. Angulo. 2005. Elimination of ie1 significantly attenuates murine cytomegalovirus virulence but does not alter replicative capacity in cell culture. J. Virol. 79:7182–7194. 20. Gribaudo, G., L. Riera, D. Lembo, M. De Andrea, M. Gariglio, T. L. Rudge, L. F. Johnson, and S. Landolfo. 2000. Murine cytomegalovirus stimulates cellular thymidylate synthase gene expression in quiescent cells and requires the enzyme for replication. J. Virol. 74:4979–4987. 21. Grzimek, N. K. A., D. Dreis, S. Schmalz, and M. J. Reddehase. 2001. Random, asynchronous, and asymmetric transcriptional activity of enhancerflanking major immediate-early genes ie1/3 and ie2 during murine cytomegalovirus latency in the lungs. J. Virol. 75:2692–2705. 22. Hengel, H., W. Brune, and U. H. Koszinowski. 1998. Immune evasion by cytomegalovirus—survival strategies of a highly adapted opportunist. Trends Microbiol. 6:190–197. 23. Ho, M. 1991. Observations from transplantation contributing to the understanding of pathogenesis of CMV infection. Transplant. Proc. 23(Suppl. 3):104–109. 24. Ho, S. N., H. D. Hunt, R. M. Horton, J. K. Pullen, and L. R. Pease. 1989.

VOL. 80, 2006

25.

26.

27.

28.

29. 30.

31. 32. 33.

34.

35.

36. 37. 38. 39. 40. 41. 42. 43. 44. 45.

46.

47.

IMMUNE SENSING OF CYTOMEGALOVIRUS REACTIVATION

Site-directed mutagenesis by overlap extension using the polymerase chain reaction. Gene 77:51–59. Holtappels, R., M. W. Munks, J. Podlech, and M. J. Reddehase. 2006. CD8 T-cell-based immunotherapy of cytomegalovirus disease in the mouse model of the immunocompromised bone marrow transplantation recipient, p. 383– 418. In M. J. Reddehase (ed.), Cytomegaloviruses: molecular biology and immunology. Caister Academic Press, Wymondham, Norfolk, United Kingdom. Holtappels, R., M.-F. Pahl-Seibert, D. Thomas, and M. J. Reddehase. 2000. Enrichment of immediate-early 1 (m123/pp89) peptide-specific CD8 T cells in a pulmonary CD62Llo memory-effector cell pool during latent murine cytomegalovirus infection of the lungs. J. Virol. 74:11495–11503. Holtappels, R., J. Podlech, G. Geginat, H.-P. Steffens, D. Thomas, and M. J. Reddehase. 1998. Control of murine cytomegalovirus in the lungs: relative but not absolute immunodominance of the immediate-early 1 nonapeptide during the antiviral cytolytic T-lymphocyte response in pulmonary infiltrates. J. Virol. 72:7201–7212. Holtappels, R., D. Thomas, J. Podlech, and M. J. Reddehase. 2002. Two antigenic peptides from genes m123 and m164 of murine cytomegalovirus quantitatively dominate CD8 T-cell memory in the H-2d haplotype. J. Virol. 76:151–164. Hummel, M., and M. I. Abecassis. 2002. A model for reactivation of CMV from latency. J. Clin. Virol. 25:S123–S136. Hummel, M., Z. Zhang, S. Yan, I. Deplaen, P. Golia, T. Varghese, G. Thomas, and M. I. Abecassis. 2001. Allogeneic transplantation induces expression of cytomegalovirus immediate-early genes in vivo: a model for reactivation from latency. J. Virol. 75:4814–4822. Jarvis, M. A., and J. A. Nelson. 2002. Human cytomegalovirus persistence and latency in endothelial cells and macrophages. Curr. Opin. Microbiol. 5:403–407. Jordan, M. C. 1983. Latent infection and the elusive cytomegalovirus. Rev. Infect. Dis. 5:205–215. Kanna, K. M., R. H. Bonneau, P. R. Kinchington, and R. L. Hendricks. 2003. Herpes simplex virus-specific memory CD8⫹ T cells are selectively activated and retained in latently infected sensory ganglia. Immunity 18:593–603. (Comment, 18:583–584.) Karrer, U., S. Sierro, M. Wagner, A. Oxenius, H. Hengel, U. H. Koszinowski, R. E. Philipps, and P. Klenerman. 2003. Memory inflation: continuous accumulation of antiviral CD8⫹ T cells over time. J. Immunol. 170:2022– 2029. (Erratum, 170:3895.) Karrer, U., M. Wagner, S. Sierro, A. Oxenius, H. Hengel, T. Dumrese, S. Freigang, U. H. Koszinowski, R. E. Philipps, and P. Klenerman. 2004. Expansion of protective CD8⫹ T-cell responses driven by recombinant cytomegaloviruses. J. Virol. 78:2255–2264. Keil, G. M., A. Ebeling-Keil, and U. H. Koszinowski. 1987. Sequence and structural organization of murine cytomegalovirus immediate-early gene 1. J. Virol. 61:1901–1908. Keil, G. M., M. R. Fibi, and U. H. Koszinowski. 1985. Characterization of the major immediate-early polypeptides encoded by murine cytomegalovirus. J. Virol. 54:422–428. Kiehl, A., L. Huang, D. Franchi, and D. G. Anders. 2003. Multiple 5⬘ ends of human cytomegalovirus UL57 transcripts identify a complex, cycloheximideresistant promoter region that activates oriLyt. Virology 314:410–422. Koffron, A. J., M. Hummel, B. K. Patterson, S. Yan, D. B. Kaufman, J. P. Fryer, F. P. Stuart, and M. I. Abecassis. 1998. Cellular localization of latent murine cytomegalovirus. J. Virol. 72:95–103. Kondo, K., J. Xu, and E. S. Mocarski. 1996. Human cytomegalovirus latent gene expression in granulocyte-macrophage progenitors in culture and in seropositive individuals. Proc. Natl. Acad. Sci. USA 93:11137–11142. Kurz, S. K., M. Rapp, H.-P. Steffens, N. K. A. Grzimek, S. Schmalz, and M. J. Reddehase. 1999. Focal transcriptional activity of murine cytomegalovirus during latency in the lungs. J. Virol. 73:482–494. Kurz, S. K., and M. J. Reddehase. 1999. Patchwork pattern of transcriptional reactivation in the lungs indicates sequential checkpoints in the transition from murine cytomegalovirus latency to recurrence. J. Virol. 73:8612–8622. Kurz, S. K., H.-P. Steffens, A. Mayer, J. R. Harris, and M. J. Reddehase. 1997. Latency versus persistence or intermittent recurrences: evidence for a latent state of murine cytomegalovirus in the lungs. J. Virol. 71:2980–2987. Lefkovits, I., and H. Waldman. 1979. Limiting dilution analysis of cells in the immune system, p. 38–82. Cambridge University Press, Cambridge, England. Lembo, D., G. Gribaudo, A. Hofer, L. Riera, M. Cornaglia, A. Mondo, A. Angeretti, M. Gariglio, L. Thelander, and S. Landolfo. 2000. Expression of an altered ribonucleotide reductase activity associated with the replication of murine cytomegalovirus in quiescent fibroblasts. J. Virol. 74:11557–11565. Lucin, P., S. Jonjic, M. Messerle, B. Polic, H. Hengel, and U. H. Koszinowski. 1994. Late phase inhibition of murine cytomegalovirus replication by synergistic action of interferon-gamma and tumor necrosis factor. J. Gen. Virol. 75:101–110. Meier, J. L. 2001. Reactivation of the human cytomegalovirus major immediate-early regulatory region and viral replication in embryonal NTera2 cells: role of trichostatin A, retinoic acid, and deletion of the 21-base-pair repeats and modulator. J. Virol. 75:1581–1593.

10455

48. Meier, J. L., and M. F. Stinski. 2006. Major immediate-early enhancer and its gene products, p. 151–166. In M. J. Reddehase (ed.), Cytomegaloviruses: molecular biology and immunology. Caister Academic Press, Wymondham, Norfolk, United Kingdom. 49. Messerle, M., B. Bu ¨hler, G. M. Keil, and U. H. Koszinowski. 1992. Structural organization, expression, and functional characterization of the murine cytomegalovirus immediate-early gene 3. J. Virol. 66:27–36. 50. Messerle, M., I. Crnkovic, W. Hammerschmidt, H. Ziegler, and U. H. Koszinowski. 1997. Cloning and mutagenesis of a herpesvirus genome as an infectious bacterial artificial chromosome. Proc. Natl. Acad. Sci. USA 94: 14759–14763. 51. Mocarski, E. S., Jr. 2004. Immune escape and exploitation strategies of cytomegaloviruses: impact on and imitation of the major histocompatibility system. Cell. Microbiol. 6:707–717. 52. Mocarski, E. S., Jr., G. Hahn, K. Lofgren White, J. Xu, B. Slobedman, L. Hertel, S. A. Aguirre, and S. Noda. 2006. Myeloid cell recruitment and function in pathogenesis and latency, p. 465–481. In M. J. Reddehase (ed.), Cytomegaloviruses: molecular biology and immunology. Caister Academic Press, Wymondham, Norfolk, United Kingdom. 53. Munks, M. W., M. C. Gold, A. L. Zajac, C. M. Doom, C. S. Morello, D. H. Spector, and A. B. Hill. 2006. Genome-wide analysis reveals a highly diverse CD8 T cell response to murine cytomegalovirus. J. Immunol. 176:3760–3766. 53a.Munks, M. W., K. S. Cho, A. K. Pinto, S. Sierro, P. Klenerman, and A. B. Hill. 2006. Four distinct patterns of memory CD8 T cell responses to chronic murine cytomegalovirus infection. J. Immunol. 177:450–458. 54. Murphy, J. C., W. Fischle, E. Verdin, and J. H. Sinclair. 2002. Control of cytomegalovirus lytic gene expression by histone acetylation. EMBO J. 21: 1112–1120. 55. O’Connor, M., M. Peifer, and W. Bender. 1989. Construction of large DNA segments in Escherichia coli. Science 244:1307–1312. 56. Pahl-Seibert, M.-F., M. Juelch, J. Podlech, D. Thomas, P. Deegen, M. J. Reddehase, and R. Holtappels. 2005. Highly protective in vivo function of cytomegalovirus IE1 epitope-specific memory CD8 T cells purified by T-cell receptor-based cell sorting. J. Virol. 79:5400–5413. 57. Pass, R. F. 2001. Cytomegalovirus, p. 2675–2705. In D. M. Knipe and P. M. Howley (ed.), Fields virology, 4th ed. Lippincott Williams & Wilkins Publishers, Philadelphia, Pa. 58. Pawelec, G., A. Akbar, C. Caruso, R. Effros, B. Grubeck-Loebenstein, and A. Wikby. 2004. Is immunosenescence infectious? Trends Immunol. 25:406– 410. 59. Pinto, A. K., and A. B. Hill. 2005. Viral interference with antigen presentation to CD8 T cells: lessons from cytomegalovirus. Viral Immunol. 18:434– 444. 60. Podlech, J., R. Holtappels, N. K. A. Grzimek, and M. J. Reddehase. 2002. Animal models: murine cytomegalovirus. Methods Microbiol. 32:493–525. 61. Podlech, J., R. Holtappels, M.-F. Pahl-Seibert, H.-P. Steffens, and M. J. Reddehase. 2000. Murine model of interstitial cytomegalovirus pneumonia in syngeneic bone marrow transplantation: persistence of protective pulmonary CD8-T-cell infiltrates after clearance of acute infection. J. Virol. 74: 7496–7507. 62. Podlech, J., R. Holtappels, N. Wirtz, H.-P. Steffens, and M. J. Reddehase. 1998. Reconstitution of CD8 T cells is essential for the prevention of multiple-organ cytomegalovirus histopathology after bone marrow transplantation. J. Gen. Virol. 79:2099–2104. 63. Polic, B., H. Hengel, A. Krmpotic, J. Trgovcich, I. Pavic, P. Lucin, S. Jonjic, and U. H. Koszinowski. 1998. Hierarchical and redundant lymphocyte subset control precludes cytomegalovirus replication during latent infection. J. Exp. Med. 21:1047–1054. 64. Pollock, J. L., R. M. Presti, S. Paetzold, and H. W. Virgin, IV. 1997. Latent murine cytomegalovirus infection in macrophages. Virology 227:168–179. 65. Posfai, G., M. D. Koob, H. A. Kirkpatrick, and F. R. Blattner. 1997. Versatile insertion plasmids for targeted genome manipulations in bacteria: isolation, deletion, and rescue of the pathogenicity island LEE of the Escherichia coli O157:H7 genome. J. Bacteriol. 179:4426–4428. 66. Presti, R. M., D. L. Popkin, M. Connick, S. Paetzold, and H. W. Virgin, IV. 2001. Novel cell type-specific antiviral mechanism of interferon gamma action in macrophages. J. Exp. Med. 193:483–496. 67. Rammensee, H.-G., J. Bachmann, and S. Stevanovic. 1997. MHC ligands and peptide motifs. Molecular Biology Intelligence Unit, Landes Bioscience, Austin, Tex. 68. Rawlinson, W. D., H. E. Farrell, and B. G. Barrell. 1996. Analysis of the complete DNA sequence of murine cytomegalovirus. J. Virol. 70:8833–8849. 69. Reddehase, M. J. 2002. Antigens and immunoevasins: opponents in cytomegalovirus immune surveillance. Nat. Rev. Immunol. 2:831–844. 70. Reddehase, M. J., M. Balthesen, M. Rapp, S. Jonjic, I. Pavic, and U. H. Koszinowski. 1994. The conditions of primary infection define the load of latent viral genome in organs and the risk of recurrent cytomegalovirus disease. J. Exp. Med. 179:185–193. 71. Reddehase, M. J., H.-J. Bu ¨hring, and U. H. Koszinowski. 1986. Cloned long-term cytolytic T-lymphocyte line with specificity for an immediate-early membrane antigen of murine cytomegalovirus. J. Virol. 57:408–412. 72. Reddehase, M. J., and H. Kalbacher. 1989. Role of individual amino acid

10456

73. 74.

75. 76. 77. 78.

79. 80.

81.

82. 83.

SIMON ET AL.

side chains in the binding of antigenic peptides to their MHC class I glycoprotein receptor, p. 45–49. In M. Braxenthaler, K.-U. Fro ¨hlich, B. Jopski, R. Kamarek, and R. Schubert (ed.), Proceedings of MoBBEL, 4th International Autumn Meeting, Tannheim, Austria. Molekularbiologische und Biotechnologische Entwicklungsliga (MoBBEL e.V.), Tuebingen, Germany. Reddehase, M. J., and U. H. Koszinowski. 1984. Significance of herpesvirus immediate early gene expression in cellular immunity to cytomegalovirus infection. Nature (London) 312:369–371. Reddehase, M. J., and U. H. Koszinowski. 1991. Redistribution of critical major histocompatibility complex and T cell receptor-binding functions of residues in an antigenic sequence after biterminal substitution. Eur. J. Immunol. 21:1697–1701. Reddehase, M. J., J. Podlech, and N. K. A. Grzimek. 2002. Mouse models of cytomegalovirus latency: overview. J. Clin. Virol. 25:S23–S36. Reddehase, M. J., J. B. Rothbard, and U. H. Koszinowski. 1989. A pentapeptide as minimal antigenic determinant for MHC class I-restricted T lymphocytes. Nature (London) 337:651–653. Reddehase, M. J., C. O. Simon, J. Podlech, and R. Holtappels. 2004. Stalemating a clever opportunist: lessons from murine cytomegalovirus. Hum. Immunol. 65:446–455. Reddehase, M. J., F. Weiland, K. Mu ¨nch, S. Jonjic, A. Lu ¨ske, and U. H. Koszinowski. 1985. Interstitial murine cytomegalovirus pneumonia after irradiation: characterization of cells that limit viral replication during established infection of the lungs. J. Virol. 55:264–273. Riddell, S. R. 1995. Pathogenesis of cytomegalovirus pneumonia in immunocompromised hosts. Semin. Respir. Infect. 10:199–208. Rognan, D., M. J. Reddehase, U. H. Koszinowski, and G. Folkers. 1992. Molecular modeling of an antigenic complex between a viral peptide and a class I major histocompatibility glycoprotein. Proteins Struct. Funct. Genet. 13:70–85. Sears, J., M. Ujihara, S. Wong, C. Ott, J. Middeldorp, and A. Aiyar. 2004. The amino terminus of Epstein-Barr virus (EBV) nuclear antigen 1 contains AT hooks that facilitate the replication and partitioning of latent EBV genomes by tethering them to cellular chromosomes. J. Virol. 78:11487– 11505. Simon, C. O., C. K. Seckert, D. Dreis, M. J. Reddehase, and N. K. Grzimek. 2005. Role for tumor necrosis factor alpha in murine cytomegalovirus transcriptional reactivation in latently infected lungs. J. Virol. 79:326–340. Simon, C. O., C. K. Seckert, N. K. A. Grzimek, and M. J. Reddehase. 2006. Murine model of cytomegalovirus latency and reactivation: the silencing/

J. VIROL.

84. 85. 86.

87. 88.

89.

90. 91.

92. 93.

94.

desilencing and immune sensing hypothesis, p. 483–500. In M. J. Reddehase (ed.), Cytomegaloviruses: molecular biology and immunology. Caister Academic Press, Wymondham, Norfolk, United Kingdom. Sinclair, J., and P. Sissons. 2006. Latency and reactivation of human cytomegalovirus. J. Gen. Virol. 87:1763–1779. Soderberg-Naucler, C., and J. A. Nelson. 1999. Human cytomegalovirus latency and reactivation—a delicate balance between the virus and its host’s immune system. Intervirology 42:314–321. Steffens, H.-P., S. Kurz, R. Holtappels, and M. J. Reddehase. 1998. Preemptive CD8 T-cell immunotherapy of acute cytomegalovirus infection prevents lethal disease, limits the burden of latent viral genomes, and reduces the risk of virus recurrence. J. Virol. 72:1797–1804. Streblow, D. N., and J. A. Nelson. 2003. Models of HCMV latency and reactivation. Trends Microbiol. 11:293–295. Suvas, S., A. K. Azkur, and B. T. Rouse. 2006. Qa-1b and CD94-NKG2a interaction regulate cytolytic activity of herpes simplex virus-specific memory CD8⫹ T cells in the latently infected trigeminal ganglia. J. Immunol. 176: 1703–1711. Sylwester, A. W., B. L. Mitchell, J. B. Edgar, C. Taormina, C. Pelte, F. Ruchti, P. R. Sleath, K. H. Grabstein, N. A. Hosken, F. Kern, J. A. Nelson, and L. J. Picker. 2005. Broadly targeted human cytomegalovirus-specific CD4⫹ and CD8⫹ T cells dominate the memory compartments of exposed subjects. J. Exp. Med. 202:673–685. Tang, Q., and G. G. Maul. 2003. Mouse cytomegalovirus immediate-early protein 1 binds with host cell repressors to relieve suppressive effects on viral transcription and replication during lytic infection. J. Virol. 77:1357–1367. Tang, Q., and G. G. Maul. 2006. Immediate-early interactions and epigenetic defense mechanisms, p. 131–149. In M. J. Reddehase (ed.), Cytomegaloviruses: molecular biology and immunology. Caister Academic Press, Wymondham, Norfolk, United Kingdom. Wagner, M., S. Jonjic, U. H. Koszinowski, and M. Messerle. 1999. Systematic excision of vector sequences from the BAC-cloned herpesvirus genome during virus reconstitution. J. Virol. 73:7056–7060. Xu, Y., S. A. Cei, A. Rodriguez Huete, K. S. Celletti, and G. S. Pari. 2004. Human cytomegalovirus DNA replication requires transcriptional activation via an IE2- and UL84-responsive bidirectional promoter element within oriLyt. J. Virol. 78:11664–11677. Yewdell, J. W., and M. Del Val. 2004. Immunodominance in TCD8⫹ responses to viruses: cell biology, cellular immunology, and mathematical models. Immunity 21:149–153.