Previous Article | Next Article ![]()
Journal of Virology, August 2002, p. 8003-8010, Vol. 76, No. 16
0022-538X/02/$04.00+0 DOI: 10.1128/JVI.76.16.8003-8010.2002
Copyright © 2002, American Society for Microbiology. All Rights Reserved.
Barak Maguen,1 Ling Jin,1,
Kevin R. Mott,1,
John Kurylo,1 Lbachir BenMohamed,1 Ada Yukht,1 Nelson Osorio,1,
Anthony B. Nesburn,1,2 Gail Henderson,3 Melissa Inman,3 Clinton Jones,3 and Steven L. Wechsler1,2*
Ophthalmology Research Laboratories, Cedars-Sinai Medical Center Burns & Allen Research Institute, Los Angeles, California 90048,1 Department of Ophthalmology, UCLA School of Medicine, Los Angeles, California 90024,2 Department of Veterinary and Biomedical Sciences, University of NebraskaLincoln, Lincoln, Nebraska 68583-09053
Received 11 March 2002/ Accepted 10 May 2002
|
|
|---|
|
|
|---|
Latency-associated transcript (LAT) is the only abundantly transcribed viral gene during neuronal latency (28, 35). The primary LAT transcript is 8.3 to 8.5 kb long (9, 43). A very stable intron, the 2-kb LAT, is spliced from the primary transcript (11). This 2-kb LAT intron is the major LAT RNA detected during latency (9, 32, 34, 38, 39, 42). The LAT gene is located in the viral long repeats and is thus present in two identical copies. As such, the LAT region accounts for over 16 kb of the 152-kb HSV-1 genome. Having one gene occupy over 10% of the genome of a virus that contains over 80 genes suggests that there is high selective pressure to maintain the sequence of the LAT region. Not surprisingly, there are also other genes in the LAT region. LAT overlaps the viral genes for ICP0 and ICP34.5 in an antisense direction (28, 35).
Studies with various LAT mutants have shown that LAT increases the induced and spontaneous reactivation phenotypes in the rabbit ocular model (14, 20) and the induced reactivation phenotype in mice (3, 8, 16, 23, 29, 33). This may be the result of LAT enhancing establishment of latency, LAT being directly involved in the reactivation stage, or a combination of both (19, 25). Support for a role for LAT in the establishment of latency comes from reports showing that in rabbits and mice more neurons become latently infected with LAT+ virus than with LAT- virus (25, 29, 36). In addition, LAT has antiapoptosis activity (1, 15, 19). This could protect acutely infected neurons from death and result in increased establishment of latency. It should be noted that these findings do not rule out an additional role for LAT in the reactivation stage. Interestingly, a LAT-null mutant containing the LAT promoter driving the first 1.5 kb of LAT from an ectopic location in the virus (22) and a mutant containing a polyadenylation signal inserted at LAT nucleotide 1499 (5) both have wild-type levels of reactivation. Thus, the major LAT function(s) involved in the latency-reactivation cycle appears to reside within the first 1.5 kb of the primary 8.3- to 8.5-kb transcript.
Although LAT-null mutants (i.e., mutants unable to transcribe any LAT) appear to have virulence properties similar to their wild-type parents (3, 4, 14, 20, 30), it has been proposed elsewhere that the 5' end of LAT or a gene in this region of the HSV-1 genome may affect viral virulence (21, 41). This is based, at least partially, on the altered viral virulence of some LAT mutants with deletions in the 5' end of the primary LAT transcript that do not block transcription of the remainder of LAT (21, 24). This suggests that there are two functions, one involved in reactivation and one involved in virulence, that map to the genomic location corresponding to the 5' end of LAT (i.e., the first 1.5 kb of LAT discussed above). One or both of these functions may reside in LAT or in a gene overlapping the 5' end of LAT. Several LAT sense transcripts that terminate just prior to the 5' end of LAT have been reported previously (31, 41), and one of these transcripts has been proposed previously to be involved in virulence (41). However, since this transcript does not overlap any of the LAT RNA it is unlikely to account for the altered virulence seen for mutants containing deletions only within the LAT RNA (21, 24). Consequently, we searched for other possible non-LAT transcripts near the start of LAT transcription.
We report here the discovery of a transcript antisense to LAT (AL-RNA) that overlaps the 5' end of LAT and the core LAT promoter. AL-RNA was detected by reverse transcription-PCR (RT-PCR) in PC-12 cells acutely infected with wild-type HSV-1 but not in uninfected PC-12 cells or PC-12 cells infected with the LAT-null mutant dLAT2903 (20) (which also has the entire AL region deleted). We mapped the 5' and 3' ends of the AL-RNA and found that the transcript is polyadenylated and extends from LAT nucleotide +158 to LAT nucleotide -198 (on the other DNA strand). The deletions within the three LAT mutants with altered viral virulence discussed above each remove the 5' end of AL and hence result in AL-null mutants. The findings reported here are therefore consistent with the notion that AL is involved in viral virulence. We also found that serum from rabbits infected with wild-type HSV-1 recognized recombinantly expressed AL protein, suggesting that an AL protein is made in vivo.
|
|
|---|
Rabbits. Eight- to 10-week-old New Zealand White male rabbits (Irish Farms) were used. Rabbits were treated in accordance with Association for Research in Vision and Ophthalmology, American Association for Laboratory Animal Care, and National Institutes of Health guidelines. Rabbits were bilaterally infected without scarification or anesthesia by placing 2 x 105 PFU of virus into the conjunctival cul-de-sac, closing the eye, and rubbing the lid gently against the eye for 30 s as we previously described (28).
RNA isolation. Subconfluent cell monolayers were infected at a multiplicity of infection (MOI) of 5 PFU/cell, total RNA was isolated at various times postinfection (p.i.), and Northern blot analysis was performed according to standard protocols with various 32P-labeled oligonucleotide probes specific for the detection of anti-LAT transcripts in the immediate vicinity of the LAT promoter and the start of LAT transcription.
RT-PCR. Cells were infected at an MOI of 5 PFU/cell or mock infected. Poly(A) mRNA was isolated from total cell extracts with the PolyATract mRNA Isolation System IV (catalogue no. Z5310; Promega, Madison, Wis.), according to the manufacturer's directions. RT was done with the ThermoScript RT-PCR system (catalogue no. 11146-024; Invitrogen Life Technologies, San Diego, Calif.) according to the manufacturer's directions. The primer used for cDNA synthesis was either oligo(dT) or an oligonucleotide corresponding to HSV-1 genomic nucleotides 118661 to 118690 (LAT nucleotides -140 to -111; see Fig. 1E, primer a). All PCRs were done with primer a (see previous sentence) and primer d (corresponding to HSV-1 nucleotides 118910 to 118881 or LAT nucleotides 109 to 80; see Fig. 1E, primer d). The PCR was initiated at 90°C for 5 min. This was followed by 30 cycles of 94°C for 40 s, 60°C for 30 s, and 71°C for 90 s. Extension was at 72°C for 10 min. Southern analysis was performed on the RT-PCR products with a probe corresponding to HSV-1 nucleotides 118811 to 118840 (LAT nucleotides 10 to 39; see Fig. 1E, probe e).
![]() View larger version (19K): [in a new window] |
FIG. 1. Relative location of the AL gene. (A) Schematic representation of the wild-type HSV-1 genome. TRL and IRL indicate the terminal and inverted long repeats, respectively. IRS and TRS indicate the inverted and terminal short repeats, respectively. UL and US indicate the unique long and unique short regions, respectively. The TRL and IRL are expanded, and the TRL is flipped left to right as indicated by the dashed lines, so that both repeats can be represented together in the subsequent panels. (B) Blowup of the long repeats. The primary LAT transcript is indicated by the large arrow. The solid rectangle represents the very stable 2-kb LAT intron. The LAT TATA box is indicated by TATA. The start of LAT transcription is indicated by the arrow at +1 (genomic nucleotide 118801). Several restriction enzyme sites and the relative locations of the ICP0 and ICP34.5 transcripts are shown for reference. (C) Blowup of the 5' LAT region. The LAT promoter (open rectangle), the start of LAT transcription (arrow, +1), a putative secondary LAT promoter (LAP2), and the beginning of the stable 2-kb LAT are indicated. Nucleotide positions relative to the start of LAT transcription are shown in parentheses. (D) AL gene. The DNA strand opposite that of LAT in panel C is shown. The numbers in parentheses indicate nucleotide positions relative to the start of LAT transcription on the other strand. The remaining numbers indicate nucleotide positions relative to the start of AL transcription. The ORF encoding a putative AL protein is shown as an open rectangle. The AL promoter is located somewhere between the large parentheses in the same region as LAP2. (E) Relative positions of oligonucleotide primers (a to d) and probes (e) used in this report. a, 118661 to 118690; b, 118681 to 118710; c, 118841 to 118870; d, 118881 to 118910; e, 118811 to 118840
|
DNA sequencing. Sequencing of cloned DNA was done by standard dideoxy sequencing.
Expression of AL protein in Escherichia coli. The AL open reading frame (AL ORF) from the ATG codon to the last amino acid before the stop codon was amplified from a cloned full-length RT-PCR product by PCR and fused to a His tag-containing plasmid (Xpress system protein expression pTrcHis; Invitrogen Life Technologies) and expressed in E. coli according to the manufacturer's protocol.
Anti-AL ORF peptide antibodies. Two peptides, one representing the putative 5' end of the AL protein (peptide 1, GRARKVFRTTRDRHGC) and one representing the putative 3' end of the AL protein (peptide 2, KAEMTQQKQATAPRGC), were synthesized, and antibody was raised commercially in rabbits (BioSource International, Camarillo, Calif.). The E. coli-expressed AL ORF was recognized by both peptide antibodies on Western blots (data not shown), suggesting that these antibodies would be capable of detecting an AL protein produced by HSV-1. In some experiments, numerous AL-related bands were detected in a laddering pattern (data not shown), suggesting that the E. coli-expressed AL protein binds tightly to itself or other proteins.
Western blots. The E. coli-expressed AL-His tag fusion protein was either used as a crude total cell extract or partially purified with a His-Bind Quick Column (Novagen; catalogue no. 70159-4) according to the manufacturer's directions. The total extract or partially purified protein was run on standard 15% Tricine gels with a wide sample well and transferred to a polyvinylidene difluoride (PVDF) membrane. The membrane was cut into strips, and individual strips were incubated with serum obtained from rabbits on day 57 p.i. The antibody bound to the blots was visualized by chemiluminescence with secondary antibody conjugated to horseradish peroxidase.
|
|
|---|
Sequence analysis revealed a potential small ORF downstream of ICP0 in HSV-1 strains 17syn+ and McKrae, the only strains for which this sequence is available. This ORF (Fig. 1D, AL ORF) overlaps part of the LAT promoter and part of the LAT region that is required for spontaneous reactivation (the first 1.5 kb of the primary LAT transcript). Interestingly, the transcript that would encode this novel ORF would be expressed antisense to LAT and would not be contained within the ICP0 mRNA. To look for a potential RNA near the 5' end of LAT that might encode this ORF, RS, CV-1, and PC-12 cells were infected with wild-type HSV-1 strain McKrae and Northern blot analyses were performed. Probes specific for the detection of anti-LAT transcripts in the immediate vicinity of the LAT promoter and the start of LAT transcription were used to detect this putative transcript. No such RNA was reproducibly detected on Northern blots in any of these cells (data not shown).
To look for a low-abundance transcript in this region, oligonucleotide primers corresponding to LAT nucleotides -140 to -110 and +109 to +80 (genomic nucleotides 118661 to 118690 and 118910 to 118881, respectively) (primers a and d in Fig. 1E) were used for RT-PCR. Primer a was also used for first-strand synthesis. Thus, the RT-PCR would produce a product only from RNA that was antisense to LAT. PC-12 cells were infected at an MOI of 5 with wild-type strain McKrae, and total RNA was isolated 6 h p.i. An RT-PCR product of the size predicted from the genomic sequence and that hybridized to an internal probe corresponding to LAT nucleotides 80 to 39 (genomic nucleotides 118811 to 118840; probe e in Fig. 1E), was readily detected (Fig. 2, lane 2). The expected size of the RT-PCR product is shown in lane 8, which shows the PCR product generated with wild-type McKrae genomic DNA with the same primers. As expected, no RT-PCR product was detected with RNA from uninfected PC-12 cells (lane 5) or when reverse transcriptase was left out of the reaction mixture (lanes 1, 6, and 7). In addition, no RT-PCR product was detected in PC-12 cells infected with the LAT- virus dLAT2903 (20) (lane 4). This was expected, since the deletion in this mutant (LAT nucleotides -161 to +1667) encompasses all of the RT-PCR primers. This acts as a further control to confirm that the RT-PCR product originated from the expected location. Also as expected, an RT-PCR product indistinguishable from that of wild-type-infected PC-12 cells was seen in PC-12 cells infected with marker-rescued dLAT2903R (lane 3). These results revealed the presence of an RNA antisense to and overlapping the 5' end of LAT (AL-RNA).
![]() View larger version (88K): [in a new window] |
FIG. 2. Detection of AL-RNA by RT-PCR. PC-12 cells were infected at an MOI of 5. Total RNA was isolated 6 h p.i., and RT-PCR was performed as described in Materials and Methods. Lanes: 1, no RT, wild-type McKrae virus-infected PC-12 cells; 2, wild-type McKrae virus; 3, dLAT2903R; 4, dLAT2903; 5, uninfected PC-12; 6, no RT, dLAT2903; 7, no RT, uninfected PC-12; 8, marker, PCR of wild-type McKrae viral DNA.
|
Mapping the 3' end of AL-RNA. The 3' end of the AL-RNA was mapped by poly(A) selection and RT-PCR. Briefly, RT-PCR was performed on poly(A)+ RNA isolated from PC-12 cells infected with wild-type McKrae virus as described above. Oligo(dT) was used as the first primer, and a synthetic oligonucleotide corresponding to LAT nucleotides -120 to -91 (genomic nucleotides 118710 to 118681; primer b, Fig. 1E) was used as the second primer. The resulting RT-PCR product was cloned and sequenced. Reading from the 5' toward the 3' end of the AL-RNA, the sequence was identical to that of the genomic sequence in this region up to genomic nucleotide 118603 (Fig. 3). This corresponds to the nucleotide complementary to LAT nucleotide -198. After this nucleotide, the sequence of the RT-PCR product diverged from the genomic sequence. The divergent sequence corresponded to a 22-nucleotide AU-rich region containing a consensus poly(A) attachment signal (AAUAAA) (40) followed by a 24-nucleotide-long poly(A) stretch.
![]() View larger version (17K): [in a new window] |
FIG. 3. Mapping the 3' end of the AL-RNA. RT-PCR was performed on poly(A)+ RNA isolated from PC-12 cells infected with wild-type McKrae virus, and the RT-PCR product was cloned and sequenced as described in Materials and Methods. The resulting sequence is labeled RNA. The "genome" sequence is from the same HSV-1 strain (McKrae). The asterisks in the "genome" sequence indicate identity to the RNA sequence. This identity ends after genomic nucleotide 118603. The positions of the 5' primer and the 3' oligo(dT) primer are shown.
|
Potential location of the AL promoter. There is no TATA box located near the 5' end of the AL-RNA. However, the region of LAT corresponding to the general region in which the AL promoter would be expected to be located (Fig. 1D) has been reported to act as a promoter for the stable 2-kb LAT intron (13) (Fig. 1C) which begins at LAT nucleotide 661. This putative secondary LAT promoter has been termed LAP2 (7, 13). The LAP2 promoter contains several bidirectional elements, such as Sp1 (12), and can enhance transcription in both orientations (2). Thus, it is possible that LAP2 acts as the AL gene promoter.
Kinetics of AL-RNA expression in tissue culture. PC-12 cells were infected as described above. RNA was extracted at different times p.i., and the presence or absence of AL-RNA was determined by RT-PCR as described in Materials and Methods with poly(dT) for first-strand synthesis and AL-specific primers (a and d; Fig. 1E) for the PCR. AL-RNA was detected as early as 4 h p.i., peaked at 6 h p.i., and then declined (Fig. 4). These kinetics were similar to those of the immediate-early gene ICP0 mRNA (data not shown). The kinetics of AL-RNA expression also appeared to be complementary to the kinetics of the expression of LAT RNA, which first became apparent at 6 to 8 h p.i. (data not shown). This is consistent with the possibility of some form of antisense regulation between these two genes.
![]() View larger version (59K): [in a new window] |
FIG. 4. Kinetics of AL-RNA expression. PC-12 cells were infected at an MOI of 5 with wild-type McKrae virus. Total RNA was isolated at various times p.i., and RT-PCR was performed as described in Materials and Methods. Lanes: 1, positive control (PCR of plasmid AL DNA); 2, uninfected; 3, 2 h p.i.; 4, 4 h p.i.; 5, 6 h p.i.; 6, 8 h p.i.; 7, 10 h p.i.; 8, 12 h p.i.; 9, 14 h p.i.; 10, no RT, 8 h p.i.
|
![]() View larger version (38K): [in a new window] |
FIG. 5. Complete sequence of AL-RNA. The sequence was determined as described in the text. AL-RNA indicates the DNA sequence corresponding to the AL-RNA sequence. The genomic sequence is from McKrae (the same HSV-1 strain as the RNA). The AL ORF begins with a methionine (M) at AL nucleotide 61 and ends with a TAG at AL nucleotide 229. The boxed C at genomic nucleotide location 118769 indicates the only nucleotide difference in AL between McKrae and HSV-1 strain 17syn+, which contains a G at this location. The boxed Q indicates that the nucleotide change from C to G changed the predicted amino acid from Q in McKrae to E in 17syn+.
|
Rabbits were infected in both eyes with 2 x 105 PFU of wild-type McKrae or dLAT2903 virus per eye, and serum was collected from individual rabbits 57 days p.i., a time at which latency had been well established. Total cell extract from E. coli expressing the AL-His tag fusion protein was run on a Tricine gel in a large loading well. The protein was transferred by blotting to PVDF membrane. The membrane was then cut into strips, and the different strips were reacted with different sera (Fig. 6A). Individual sera from three different wild-type-infected rabbits (lanes 1, 2, and 3) recognized a band corresponding in mobility to the AL band recognized by anti-AL ORF peptide antibody (lane M, arrow). In contrast, none of the dLAT2903-infected rabbit sera (lanes 4, 5, and 6) or uninfected rabbit sera (lane U) appeared to recognize this band. In preliminary experiments, we found that antibodies raised against two different synthetic AL peptides reacted with the same band recognized by the anti-His tag antibody and that this band had an apparent molecular weight consistent with that expected for the AL-His tag protein (data not shown).
![]() View larger version (60K): [in a new window] |
FIG. 6. Detection of E. coli-expressed AL protein by sera from infected rabbits. (A) Total extract from E. coli expressing the AL-His tag fusion protein was run on a 15% Tricine gel with a single large loading well and transferred to a PVDF membrane. The membrane was cut into strips, and each strip was separately reacted with the indicated antibody. The strips were then reacted with horseradish peroxidase-conjugated secondary antibody for chemifluorescence. The arrow indicates the location of the E. coli-expressed AL-His tag fusion protein. Lane M, anti-His tag antibody as marker. Lane U, serum from an uninfected rabbit. Lanes 1 to 3, sera from three different rabbits infected with wild-type virus. Lanes 4 to 6, sera from three different rabbits infected with dLAT2903 (a LAT and AL-null mutant). (B) The E. coli-expressed AL-His tag protein was partially purified as described in Materials and Methods and processed as for panel A. Lane M, anti-His tag antibody as marker. Lane U, serum from an uninfected rabbit different from the serum in panel A. Lanes 3 and 4, the same rabbit sera as in lanes 3 and 4 in panel A. wt, wild type.
|
The sera used in Fig. 6A, along with three additional sera from wild-type-infected rabbits and three additional sera from dLAT2903-infected rabbits, were further examined for the presence of AL antibodies by enzyme-linked immunosorbent assays (ELISAs). ELISA plates were coated with AL synthetic peptide 1 or 2. Sera from four of six wild-type-infected rabbits contained small but significant amounts of immunoglobulin G to either peptide 1 or 2 or both, based on their positive ELISA titers against these peptides. In contrast none of the sera from dLAT2903-infected rabbits produced an ELISA reading above background (P = 0.03, single-sided Fisher exact test).
Thus, the AL protein appeared to be expressed following ocular infection of rabbits. The minimal reactivity of the wild-type-infected rabbit serum with the AL-His tag fusion protein and the AL synthetic peptides suggests either that the AL protein is not very immunogenic and/or that only small amounts of AL protein were produced.
|
|
|---|
The amino acid sequence of the putative AL protein is highly conserved in HSV-1 strains McKrae and 17syn+ (>98% identical). This level of amino acid sequence conservation is similar to that seen with other HSV-1 proteins (10). However, it is in sharp contrast to the much lower amino acid sequence conservation of 10 to 80% in eight potential ORFs within the first 1.5 kb of LAT (located on the other DNA strand opposite AL) (10). Based on the low conservation, it was concluded that none of these potential LAT ORFs encode a protein involved in the HSV-1 reactivation phenotype. In contrast, the high conservation of the putative AL protein amino acid sequence suggests that there is strong selective pressure to maintain the amino acid sequence. This suggests that AL encodes a protein.
We found that rabbits infected with wild-type HSV-1 by the ocular route developed antibodies to the AL protein. This is strong evidence that the AL protein was produced during acute and/or latent infection in rabbits. Combined with the high sequence conservation, it is therefore likely that AL is expressed by the virus in certain cell types and may play a role in the pathogenic potential of HSV-1. The apparently low level of AL antibody and our inability to detect AL protein above the background in infected cells in tissue culture suggested that AL is a low-abundance protein and/or it is produced only in certain cell types or at certain times, such as during latency.
The relative locations of the LAT and AL genes as presented in this study suggest the possibility of antisense regulation of LAT by AL or vice versa. This may be important in the latency-reactivation cycle. We found an apparent discordance between transcription of AL and LAT in tissue culture. AL transcription is greatest prior to the time that LAT expression becomes readily apparent. Transcription of AL decreases after 6 h p.i., the time that LAT expression increases dramatically. This is consistent with the possibility of some form of antisense regulation between AL and LAT. It has been reported elsewhere that dexamethasone-induced reactivation of bovine herpesvirus 1 results in transient reduction of LAT expression (27). This decrease in LAT may be a key factor in reactivation. It is possible that a reactivation stimulus either directly or indirectly up regulates AL transcription and that this in turn down regulates LAT, which then leads to reactivation.
We (21, 24) and others (41) have suggested elsewhere that LAT, or the region of the HSV-1 genome near the start of the primary LAT transcript, may be involved in viral virulence. Zhu et al. (41) described a 0.7-kb RNA that overlaps the LAT promoter, is transcribed in the LAT direction, and terminates just prior to the start of LAT. They also showed that a mutant with a deletion in this 0.7-kb region has altered virulence in mice. Interestingly, this mutant also deletes a portion of the AL gene. The altered virulence of this mutant may therefore be due to partial deletion of the 0.7-kb LAT sense RNA or partial deletion of the AL gene. We have also reported alterations in viral virulence in three mutants with deletions within the 5' end of the primary 8.3-kb LAT transcript (21, 24). dLAT1.5 has a deletion of LAT nucleotides 76 to 1499 (corresponding to AL nucleotides 82 to -1342). dLAT371 has a deletion of LAT nucleotides 76 to 447 (corresponding to AL nucleotides 82 to -289). Both of these mutants express the remainder of LAT. In contrast, LAT2.9A, which contains the same deletion as dLAT371, also does not express LAT nucleotides 1500 to 8324 (the end of LAT). These deletions do not alter the structure of the genomic region encoding the 0.7-kb LAT sense RNA. However, they may impact expression of the 0.7-kb RNA, since expression of this RNA was abolished in an HSV-1 strain 17syn+ mutant similar to dLAT371 (41). Thus, it remains possible that the altered virulence phenotypes of all four of these mutants (one directly affecting the 0.7-kb transcript and AL and three directly affecting LAT and AL and probably indirectly affecting the 0.7-kb RNA) may be due to AL.
LAT-null mutants that also delete AL or significantly disrupt the AL ORF do not appear to have altered virulence (3, 4, 14, 20, 30). It is therefore possible that LAT and AL have opposing and balanced functions with respect to viral virulence. Blocking expression of both genes simultaneously would then have no obvious virulence phenotype. In contrast, three mutants with deletions that remove the 5' end of AL (AL-null mutants) but do not disrupt the LAT promoter (and hence allow the mutants to continue to express the region of LAT outside of the deletion) do have altered virulence (21, 24). It is possible that this is the result of altering the balance between an AL function and a LAT function. In addition, the virulence phenotypes of these three mutants differ from each other and also differ between rabbits and mice. Since each of these mutants expresses different LAT transcripts, it is also possible that there is an intricate balance between one or more AL functions and functions due to different regions of LAT.
Present address: Department of Ophthalmology, University of California Irvine, UCI Medical Center, Orange, CA 92868. ![]()
|
|
|---|
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»