Previous Article | Next Article ![]()
Eukaryotic Cell, December 2003, p. 1253-1265, Vol. 2, No. 6
1535-9778/03/$08.00+0 DOI: 10.1128/EC.2.6.1253-1265.2003
Copyright © 2003, American Society for Microbiology. All Rights Reserved.
Center for Biosystems Research, University of Maryland Biotechnology Institute, College Park, Maryland 20742-4450
Received 18 June 2003/ Accepted 22 August 2003
|
|
|---|
|
|
|---|
The pleiotropic nature of hypovirus-mediated phenotypic changes suggested the perturbation of one or more key regulatory pathways (13). Subsequent studies implicated two principal signal transduction pathways in the manifestation of hypovirus-induced phenotypic changes: G-protein-linked, cyclic AMP (cAMP)-mediated (5, 7, 20, 37) and calcium/calmodulin/inositol trisphosphate-dependent (31, 32) signaling cascades. However, evidence for hypovirus-mediated alteration of these pathways has relied predominantly on single-gene analyses to monitor pathway activity, e.g., the 13-1 gene for G-protein-linked, cAMP-mediated signaling and the lac-1 (laccase) gene for calcium/calmodulin/inositol trisphosphate-dependent signal transduction. Using differential mRNA display, Chen et al. (5) reported that more than 400 PCR products either increased (n = 296) or decreased (n = 127) in abundance as a result of infection by the prototypic hypovirus CHV1-EP713. Moreover, similar changes in the accumulation of approximately 65% of these PCR products were observed when G-protein/cAMP signaling was altered in the absence of hypovirus infection. Kang et al. (25) subsequently used the more sensitive ordered differential display approach to estimate that 20% of the total C. parasitica genes were modulated following hypovirus CHV1 infection. While mRNA differential display can provide a good indication of the relative extent to which transcriptional profiles change, considerable additional effort is required to determine the identities of differentially expressed genes. We now report the successful use of a spotted expressed sequence tag (EST) array derived from a C. parasitica EST (CEST) library consisting of over 4,200 sequences representing approximately 2,200 unique genes (12) to monitor hypovirus-mediated global changes in host gene expression.
|
|
|---|
Total RNA isolation. Cultures used for RNA isolation were grown on PDA-cellophane for 6 days and harvested by freezing the mycelia in liquid nitrogen and then immediately grinding the mycelia into a fine powder by using a mortar and pestle. The powder was resuspended in RNA extraction buffer (200 mM NaCl, 100 mM Tris-Cl [pH 8.0], 4 mM EDTA [pH 8.0], 2% sodium dodecyl sulfate [SDS], 2 mM dithiothreitol) at a ratio of 4 ml of buffer per g of mycelia. This step was followed sequentially by extraction with equal volumes of water-saturated phenol, acid phenol-chloroform-isoamyl alcohol (125:24:1) (pH 4.5), and chloroform. Single-stranded RNA was precipitated on ice for 2 h by the addition of LiCl to a final concentration of 2 M. The single-stranded RNA precipitate was collected by centrifugation, resuspended in 2 ml of double-distilled H2O, reprecipitated by the addition of 2.5 volumes of ice-cold ethanol and 0.1 volume of 3 M sodium acetate (pH 5.2), and incubated for 30 min at -20°C. The RNA was collected by centrifugation, washed with 2 ml of ice-cold 75% ethanol, dried, and treated with 2 U of RQ1 DNase (Promega) in 0.5 ml of 20 mM Tris (pH 8.0)-20 mM MgCl2 in the presence of 40 U of RNasin (Promega) for 1 h at 37°C. Following phenol-chloroform and chloroform extractions, the RNA was precipitated with ethanol and resuspended in 100 µl of double-distilled H2O.
Microarray slide printing.
An ordered CEST library was constructed from cDNA prepared from a mixed mRNA population isolated from hypovirus CHV1-EP713-infected C. parasitica strain EP155 and uninfected C. parasitica strain EP155 cloned into the
ZipLox vector (Invitrogen) (12). A large majority (95.7%) of the inserts were sized at 500 to 2,000 bp, and sequencing of 4,216 clones from the 5' end yielded an average read length of 608 bp (12). A subset of this CEST library was selected for microarray printing by removal of ESTs corresponding to the hypovirus genome and to the highly abundant hydrophobin gene cryparin. The resulting 3,864 ESTs, representing approximately 2,200 unique genes, were reordered to form the array EST (AEST) library used for printing. The accuracy of library reordering was ensured by sequencing eight AEST clones from each 96-well plate used to organize the library. Correspondence between CEST and AEST library clone designations can be accessed at http://www.umbi.umd.edu/
cbr/aesttocest.pdf. PCR fragments used for printing the microarray chip were amplified from the AEST library by using Sp6 and T7 primer sites flanking the inserts in standard 100-µl PCR mixtures. PCR products were analyzed on gels to confirm the success of the reactions and were subsequently purified by using MultiScreen-PCR plates (Millipore). Purified PCR products were resuspended in 50 µl of 3xSSC (1xSSC is 0.15 M sodium chloride plus 0.015 M sodium citrate [pH 7.0]) for printing. Microarray printing and hybridization were performed at the University of Maryland Biotechnology Institute Center for Biosystems Research DNA Microarray Core Facility, which is equipped with an Affymetrix 417 Arrayer and 418 Scanner (http://www.umbi.umd.edu/
cab/microarraymain.html). Purified PCR products were arrayed in duplicate on poly-L-lysine-coated glass slides with an average spot diameter of 100 µm and a spot spacing of 300 µm.
Following printing, the spotted cDNA was cross-linked to the surface of the slides (at 65 mJ) by using a StrataLinker instrument and washed with a 1% SDS solution to minimize the background. Slides were subsequently placed in a blocking solution containing 0.2 M succinic anhydride and 0.05 M sodium borate prepared in 1-methyl-2-pyrrolidinone for 20 min, washed for 2 min in 95°C water, and rinsed five times in 95% ethanol. Slides were spin dried at 500 rpm for 5 min and stored for future hybridizations.
Fluorescent probe generation for hybridization. Fluorescence-labeled cDNA probes were prepared from total RNA purified from uninfected or CHV1-EP713-infected C. parasitica strain EP155 (60 µg per probe) by direct incorporation of Cy3- or Cy5-labeled dUTP with Superscript II reverse transcriptase (Invitrogen) and 2 µg of oligo(dT) primer according to the manufacturer's instructions. Unincorporated nucleotides were removed by using a Microcon-30 spin column, and the purified probes were combined for further processing immediately prior to hybridization.
Microarray hybridization and scanning. Printed slides were prepared for hybridization by the addition of 30 µl of prehybridization solution, which contained 50% formamide, 6x SSPE (1x SSPE is 0.15 M NaCl, 0.01 M NaH2PO4, and 0.001 M EDTA), 0.5% SDS, 5x Denhardt's solution, and 100 µg of salmon sperm DNA/ml, to the arrayed surface of a glass slide (covered with a coverslip to evenly distribute the prehybridization solution). The slide was incubated for 30 min at 42°C in a hybridization chamber. Fluorescence-labeled probes were dried and resuspended in 20 µl of hybridization solution, which contained 50% formamide, 6x SSPE, 0.5% SDS, 5x Denhardt's solution, blocking solution [2 µg of poly(dA), 4 µg of yeast tRNA, 10 µg of salmon sperm DNA)], 14 µl of master mix solution (70% formamide, 3x Denhardt's solution, 0.7% SDS), and 6 µl of 20x SSPE. The resuspended probes were heated to 100°C for 2 min, vortexed, and collected by a brief spin in a microcentrifuge. The probes were applied to the arrayed surface, covered with a coverslip, and placed in a hybridization chamber overnight at 42°C. Hybridized slides were washed in each of three solutions (solution 1 is 1x SSC- 0.1% SDS, solution 2 is 2x SSC- 0.1% SDS, and solution 3 is 2x SSC) for 10 min, spun dry, scanned in both Cy3 and Cy5 channels with an Affymetrix 418 Scanner at a 10-µm resolution and a 70% photomultiplier tube value, and exported as 16-bit TIFF images for analysis.
Microarray data processing and analysis. Integrated pixel intensity values for each spot were calculated by using TIGR Spotfinder software and saved in tab-delimited format for use by TIGR MIDAS software (The Institure for Genomic Research [TIGR], Rockville, Md.; http://www.tigr.org/software). All hybridization data from three sets of dye-swap experiments were normalized simultaneously in MIDAS to correct for experimental error within a specific hybridization and between repeated hybridizations. With the data processing functions present within MIDAS, intraslide normalization was achieved by applying a locally weighted linear regression (LOWESS) algorithm (smoothing factor, 0.33) on the log ratio-log product plot of the data set for each hybridization (on a block-by-block basis) and adjusting the Cy5 signal for each clone by its calculated LOWESS factor. Interslide normalization was achieved by rescaling the Cy3 and Cy5 signals for each spot on a chip by using the standard deviation of Cy3 and Cy5 signals measured across all hybridizations.
Selection of differentially expressed clones in each hybridization was performed by importing the normalized (and rescaled) Cy3 and Cy5 values calculated in MIDAS into the Functional Genomics module of Spotfire DecisionSite 7.0 (Spotfire, Inc., Somerville, Mass.; www.spotfire.com). The log2 (Cy3/Cy5) ratio for each clone was calculated, and the clones were divided into groups based on the number of standard deviations by which a specific clone log2 ratio varied from the data set average log2 ratio. Clones with log2 ratios equal to or greater than ±2 standard deviations in a minimum of four of six hybridizations were considered differentially expressed. Genes identified as differentially expressed were hierarchically clustered in Spotfire DecisionSite 7.0.
Validation of differentially expressed clones through real-time RT-PCR. Twenty-eight clones predicted to be differentially expressed by microarray analysis were tested by quantitative reverse transcription (RT)-PCR with an Applied Biosystems (Foster City, Calif.) GeneAmp 5700 sequence detection system and an Applied Biosystems TaqMan RT kit. A single cycle consisting of 10 min at 25°C, 40 min at 48°C, and 5 min at 95°C was used to reverse transcribe 100 ng of RQ1-treated RNA in a 100-µl reaction mixture containing 10 mM Tris-Cl (pH 8.3), 50 mM KCl, 5.5 mM MgCl2, 500 µM each deoxynucleoside triphosphate, 2.5 µM random hexamer, 0.4 U of RNase inhibitor, and 1.25 U of Multiscribe reverse transcriptase/ml. The transcript abundance of each clone of interest was then measured by using quantitative PCR with TaqMan master mix reagents. For each real-time PCR, 2.5 µl of cDNA was incubated with 1 µM each forward and reverse primers and 200 nM probe. For each template measured, 18S rRNA (1:1,000 dilution of cDNA) was measured for normalization. Transcript abundance relative to the amount of 18S rRNA in the sample was calculated by using the comparative threshold cycle method (23) with the primers and conditions described by Parsley et al. (37).
|
|
|---|
Additional elements were incorporated into the design of the CEST spotted array to enhance confidence in the microarray results. A number of control cDNAs were spotted on the microarray to guide data analysis software selection. These included CHV1-EP713 coding regions for p29, p48, and several EST clones representing different portions of open reading frame B that would be expected to increase in signal intensity when hybridized with a CHV1-EP713-specific probe and a C. parasitica gene, 13-1, that was recently shown to be CHV1-EP713 inducible (37). Clones containing cDNAs corresponding to ribosomal proteins L8, L3, L15, S10, S27, S8, S25, and S6 and 18S rRNA were included to represent genes not expected to be altered in expression upon hypovirus infection. Clones with no predicted sequence homology to the C. parasitica genome were included to monitor signal background levels; these included Escherichia coli proteins NusA, LacI, AraC, and RecJ. In addition, each of the 3,864 EST clones was spotted in duplicate. Finally, real-time RT-PCR was used to verify differential transcript accumulation for a subset of genes predicted by microarray analysis to be differentially expressed.
Previous differential mRNA display studies (5, 25) suggested the potential for large differences in specific mRNA levels upon virus infection. To avoid data processing problems associated with cDNA spots that may not have been detectable in one channel due to complete gene repression in the presence or absence of hypovirus, we used the MIDAS software package from TIGR (http://www.tigr.org/software). This software uses a LOWESS algorithm to separate a fluorescent signal of biological significance from noise that is characteristic of microarray studies (40). This package identified all control spots in at least five of the six hybridization replicates performed in this study (Table 1) while avoiding the selection of clones with aberrant signals. mRNA quality was also monitored by spotting all four exons of the C. parasitica endothiapepsin (epn-1) gene. These exons showed similar magnitudes of a decrease in transcript abundance following hypovirus infection in at least four of six hybridizations (data not shown), suggesting minimal levels of RNA degradation.
|
View this table: [in a new window] |
TABLE 1. Genes differentially expressed in EP155 and EP155/CHV1-EP713a
|
![]() View larger version (56K): [in a new window] |
FIG. 1. Scatter plot of four independent hybridizations comparing fluorescence-labeled cDNA probes derived from uninfected strain EP155 and from infected strain EP155/CHV1-EP713. Normalized signals for each channel are plotted on a logarithmic scale. Red triangles represent clones for which transcript accumulation increased in EP155/CHV1-EP713 relative to EP155. Green circles represent clones for which transcript accumulation decreased in EP155/CHV1-EP713 relative to EP155. Clones for which transcript levels did not significantly change after hypovirus infection are represented by yellow squares. Hybridizations 1 and 1R are dye-swap experiments done with the same RNA preparations from EP155 and EP155/CHV1-EP713. Hybridizations 2 and 2R are dye-swap experiments done with samples from a second, independent RNA isolation. Shaded red triangles in each data set indicate the magnitude of differential expression for hypovirus-encoded protein p48. Shaded green circles indicate the magnitude of differential expression for clone AEST-05-C-02.
|
|
View this table: [in a new window] |
TABLE 2. Validation of microarray measurements by real-time RT-PCRa
|
![]() View larger version (42K): [in a new window] |
FIG. 2. Hiearchical clustering of 295 differentially expressed clones (see Materials and Methods). Each clone is represented twice on the chip. Red squares indicate clones with increased transcript abundance and green squares indicate clones with decreased abundance after hypovirus infection. Gray blocks indicate missing data; black blocks indicate no differential gene expression. Each row represents six independent transcript abundance measurements for a specific clone. Each column represents a different hybrization experiment. Clones marked with asterisks were confirmed by real-time RT-PCR (Table 2). "A" in clone designations represents "AEST"; +, positive; Meth. Synth., methionine synthetase; Prot., protein; ABC, ATP-binding cassette; reg., regulation.
|
Microarray versus real-time RT-PCR analysis. Although several features had been built into the microarray design to enhance confidence in the microarray results, it was important to validate the differential expression values by independent means. Real-time RT-PCR was adopted because of its high level of sensitivity, its potential speed of validation, and the requirement for low quantities of test material compared to those needed for Northern analysis (3).
A subset of 28 genes predicted by microarray analysis to be differentially expressed was chosen for confirmation based on putative function or magnitude of differential transcript accumulation. Each clone tested by kinetic RT-PCR was measured in triplicate for each of two independent RNA isolations (Table 2). A total of 26 out of 28 clones were confirmed by real-time RT-PCR analysis, indicating a 93% success rate for accurately identifying differentially expressed clones.
|
|
|---|
Microarray and real-time RT-PCR results were in excellent agreement with respect to the direction of changes in transcript accumulation, i.e., up-regulated or down-regulated; however, results from the two methods did differ occasionally in the magnitude of changes. This finding was not surprising, given that cross-hybridization between closely related genes can occur on the surface of the microarray slide. Moreover, the fluorescent signal corresponding to a specific mRNA probe can be titrated out due to the presence of redundant copies of the sequence on the microarray slide. Although the accuracy of microarray data can be problematic (40, 51, 53), the 93% validation rate obtained for the C. parasitica chip (Table 2) provides a high level of confidence in the list of responsive genes presented in Table 1.
It is noteworthy that the microarray results showed a level of inconsistency with published observations for two previously characterized C. parasitica genesthe laccase gene lac-1 and the gene encoding endothiapepsin, epn-1. Based on Northern analysis, lac-1 transcripts were previously reported to be down-regulated after hypovirus infection (8, 31, 41). CHV1-EP713 infection was reported to cause no change in endothiapepsin protein production and secretion (9). Based on microarray analysis, lac-1 transcript levels remained unchanged and epn-1 transcript levels decreased in EP155/CHV1-EP713 colonies, results confirmed by kinetic RT-PCR (data not shown). A major difference between the microarray study and the previous studies with lac-1 and epn-1 responses to CHV1-EP713 infection is that fungal mycelia were grown in liquid media in the previous studies and were grown on PDA-cellophane in the present study. In this regard, the C. parasitica microarray provides a powerful new tool with which to examine the influence of culture conditions on the magnitude and spectrum of hypovirus-induced symptom expression and host gene expression (37).
The microarray analysis presented here has expanded the number of identified C. parasitica genes that respond to hypovirus infection from less than 20 (13, 25) to nearly 300. It is clear from Table 1 that these genes are potentially associated with a wide range of biological processes. A comprehensive understanding of the significance of these cellular responses to hypovirus replication and hypovirus-mediated changes in host phenotype, including hypovirulence, will require additional studies. The 133 responsive genes of unknown function (supplemental data are available at http://www.umbi.umd.edu/
cbr/155_713SUPdata.pdf) represent an additional rich resource for future research. However, several host transcriptional responses that were identified and confirmed in this study merit discussion in terms of their potential relevance to hypovirus-mediated symptom expression and virus replication. Table 3 provides specific identities and associated BLAST information for all clones discussed below.
|
View this table: [in a new window] |
TABLE 3. Characteristics of selected differentially-expressed genes HoxX SAMS HSP70 GST SAHH
|
Although HSPs play a major role in protection against and recovery from thermal stress, several animal and plant viruses also have been shown to regulate HSP synthesis (1, 15, 21, 52). Recruitment of cellular HSP70s to facilitate virion assembly or genome replication has been demonstrated (11, 22, 49). Interestingly, plant closteroviruses actually encode an HSP70 that functions to facilitate cell-to-cell movement (38). It will be of considerable interest to examine whether disruption of the C. parasitica homologue (AEST-10-H-10) of HSP70 effects CHV1-EP713 replication.
GSTs belong to a superfamily of isoenzymes involved in the removal of reactive oxygen species and the conjugation of glutathione with various harmful ligands, including plant phenols and aflatoxins (46). Although fungal GSTs have not been extensively characterized, studies with other organisms have suggested that GSTs actually participate in a wide range of cellular functions not directly associated with detoxification (reviewed in reference 18). GST-mediated glutathione S thiolation of reactive cysteine residues has been implicated in regulating the activity of cellular proteins (27, 43), including kinases, transcription factors, and caspases, that function in apoptosis (17). In this regard, Kampranis et al. recently reported that a GST from tomato suppressed cell death caused by expression of the yeast inducer of apoptosis, Bax (24). In this context, viral induction of GST expression may represent a countermeasure to a cellular apoptosis-mediated defense response (48). It is also worth considering the possibility that increased glutathione S thiolation of cellular transcription factors contributes to virus-mediated changes in gene expression and symptom expression. The constitutive elevation of GST homologue transcript levels certainly warrants an examination of the effect of hypovirus infection on the cellular redox balance.
CHV1-EP713 infection substantially elevates transcript accumulation for homologues of SAMS and SAHH. Kawalleck et al. (6) previously reported the induction of S-adenosyl-L-methionine (SAMe) synthetase (SAMS) and S-adenosyl-L-homocysteine (SAH) hydrolase (SAHH) in plants following treatment with a fungal elicitor. This observation prompted the authors to suggest a link between plant defense responses and increased turnover of the activated methyl cycle. It was intriguing, therefore, to observe constitutive increases in transcript accumulation for homologues of SAMS (AEST-08-F-10) and SAHH (AEST-22-B-11) (Tables 2 and 3) of six- and fourfold, respectively, following CHV1-EP713 infection.
SAMS catalyzes the condensation of L-methionine and ATP to produce SAMe, which serves as the primary methyl donor in transmethylation reactions involving proteins, nucleic acids, fatty acids, and polysaccharides and as the precursor for polyamine synthesis. Transfer of the methyl group from SAMe results in the production of SAH. Since increased accumulation of SAH inhibits SAMe-dependent methylation, SAH is hydrolyzed by SAHH to L-homocysteine and adenosine. Given the central role of SAMe in cellular metabolism, the constitutive up-regulation of these key enzymes could be anticipated to have a number of possible metabolic or physiologic consequences. Altered transmethylation activity due to increased SAMe levels could influence physiologic processes ranging from protein synthesis to membrane integrity. Related alterations in polyamine biosynthesis could influence cell cycle progression and development in C. parasitica, as has been reported for the pathogenic fungus Sclerotinia sclerotiorum (39) and the fission yeast Schizosaccharomyces pombe (4). Associations between abnormal intracellular SAMe levels and altered rates of DNA mutations and genome stability due to changes in DNA methylation patterns have been suggested (28, 29, 30, 44), and SAHH has been reported to influence senescence and cell growth (33, 54) through its role in regulating homocysteine levels. In this context, persistent hypovirus infection of C. parasitica may provide a particulary useful model for examining the consequences of chronic RNA virus infection on the stability of host nuclear and organellar genomes.
Hypovirus infection alters a subset of C. parasitica transcriptional regulatory factors. There is increasing evidence to support the proposal that the pleiotrophic nature of hypovirus-mediated phenotypic changes is related to the perturbation of one or more key regulatory pathways (reviewed in reference 13). Thus, an understanding of the mechanisms underlying these changes is likely to require identification of key control points or elements, such as receptors, that are involved in initiating signaling pathways and transcription factors that convert signals into changes in gene expression. Therefore, it was of interest that only 3 of 26 genes classified under the molecular function category "transcription regulation/transcription factors" by Dawe et al. (12), AEST-27-F-10, AEST-30-C-09, and AEST-05-C-02, were found to be responsive to CHV1-EP713 infection; transcript accumulation was reduced by at least threefold for each gene (Table 2). Additionally, the products of two of the three genes are homologues of fungal transcription factors that have been reported to regulate processes that are altered by hypovirus infection. Transcription factor Pro1 (AEST-27-F-10) is involved in controlling fruiting body formation and sexual sporulation in several filamentous fungi (34). Mst12 (AEST-30-C-09) from the rice pathogen Magnaporthe grisea, a recently identified homologue of yeast Ste12, has been shown to be important in regulating infectious hypha growth (36). The third responsive regulatory factor homologue, HoxX (AEST-05-C-02), is part of a bacterial two-component system involved in the regulation of a hydrogenase that oxidizes hydrogen into constituent protons and electrons and passes the electrons to the electron transport chain (14). The potential significance of reduced expression of this putative regulatory factor to hypovirus infection is more difficult to envision. Significant on the basis of the absence of responsiveness were the C. parasitica cpc1 cross-pathway control transcription factor (50) and a number of putative transcription factors, including homologues of transcription factor PacC from N. crassa, involved in regulation of the pH response, N. crassa Cys-3, involved in regulation of sulfur metabolism, the hac1 transcription factor from Hypocrea jecorinal, and transcription factor CON7 from M. grisea.
In addition to revealing new hypotheses for testing, the C. parasitica microarray now provides the opportunity to address a number of long-standing questions about hypovirus infection. It will now be possible to examine whether hypoviruses that differ in the severity of symptom expression (6) elicit similar or quite different transcriptional responses by the host. Additionally, it should be possible to determine whether a symptom-inducing, hypovirus-encoded gene product, such as p29 (10, 47), alters the expression of a specific set of cellular genes. In combination with a collection of available C. parasitica signaling mutants, microarray analysis could be used to identify sets of cellular genes that are regulated through specific signaling pathways and simultaneously monitor the effects of hypovirus infection on their expression. It is anticipated that this information, in turn, will provide insight into the roles of specific cellular genes and signaling pathways in the elaboration of specific virus-mediated phenotypic changes, including attenuation of fungal virulence.
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»