Expression of an adenovirus encoded reporter gene and its reactivation following UVC and oxidative damage in cultured fish cells

Purpose: Recombinant human adenovirus, AdCA35lacZ, was used to examine expression of a reporter gene and its reactivation following UVC (200–280 nm) and oxidative damage in fish cells. Materials and methods: AdCA35lacZ is a recombinant nonreplicating human adenovirus, which expresses the β-galactosidase (β-gal) reporter gene. UVC light produces DNA damage repaired by nucleotide excision repair (NER). In contrast, methylene blue plus visible light (MB+VL) produces oxidative DNA damage, mainly 8-oxoguanine, that is repaired by base excision repair (BER). We examined expression of the reporter gene and host cell reactivation (HCR) of the UVC-treated and MB+VL-treated reporter gene in fish cells. Results: AdCA35lacZ infection of Chinook salmon cells (CHSE-214), eel cells (PBLE) and four rainbow trout cell lines (RTG-2, RT-Gill, RTS-34st and RTS-pBk), but not zebrafish (ZEB) or carp (EPC) cells resulted in expression of β-gal. HCR of UVC-treated AdCA35lacZ in fish cells varied from that obtained in NER-deficient xeroderma pigmentosum group A fibroblasts to greater than that for NER-proficient normal human fibroblasts. HCR of UVC-treated AdCA35lacZ correlated with β-gal expression levels for untreated AdCA35lacZ. Exposure of cells to fluorescent light (400–700 nm) increased expression of the undamaged reporter gene in normal human fibroblasts and in all fish cells except PBLE and increased HCR of the UVC-damaged reporter gene in fish cells but not in human fibroblasts. HCR of the MB + VL-treated reporter gene was similar to that in human cells for PBLE, CHSE-214, RTG-2 and RTS-pBk, but was reduced in RT-Gill and RTS-34st cells. Conclusions: These results indicate the detection of functional photoreactivation (PR) of UVC-induced DNA damage in fish cells but not in normal human fibroblasts and a link between NER and transcription of the reporter gene in the fish cells in the absence of PR. We show also efficient BER of the reporter gene in several fish cell lines.

Keywords: UVC; fish cells; nucleotide excision repair; base excision repair; photoreactivation; recombinant adenovirus; host cell reactivation; oxidative damage; methylene blue plus visible light

The integrity of the vertebrate genome is constantly being compromised by alterations induced by a wide variety of exogenous physical and chemical agents, including oxidative damage due to aerobic respiration and ultraviolet (UV) light from the sun. Aerobic respiration generates reactive oxygen species (ROS) in our cells and it has been estimated that 105 damaging events occurs in every cell, every day as a result of ROS (Fraga et al. [11]). This damage can be directed towards many components of the cell, including the lipid membrane, proteins, and most importantly the DNA (as reviewed by Kawanishi et al. [23]). The main damage in DNA resulting from ROS is the oxidation of the purine and pyrimidine bases and oxidation of guanine residues is the most common giving rise to 8-hydroxyguanine (8-OxoG) (Muller et al. [35], Wang et al. [56]). UV radiation from sunlight can be divided into three different wavelength regions: Short-wavelength UVC (200–280 nm), medium-wavelength UVB (280–320 nm) and long-wavelength UVA (320–400 nm). However, due to the absorption by the ozone layer in the earth's stratosphere, UVC does not actually reach the earth's surface and thus is generally regarded to not have any significance on the environment or on human health. Therefore, the major relevant components of sunlight are UVA and UVB (as reviewed in Latonen & Laiho [26]). UVB, like UVC is directly absorbed by DNA generating cyclobutane pyrimidine dimers (CPD) and (6-4) photoproducts (6-4PP) (Friedberg et al 1995). In contrast, UVA does not directly excite DNA but rather indirectly damages DNA by a photosensitized reaction through the formation of ROS and results in the formation of non-bulky oxidative DNA base modifications including 8-OxoG (Douki et al. [8], Wondrak et al. [59]).

Direct repair of UV-induced CPDs occurs when photolyase in the presence of activating light reverses rather than excises DNA damage by the phenomenon of photoreactivation (PR). Single oxidative DNA base changes produced by UVA and through aerobic respiration, such as 8-OxoG, are removed by base excision repair (BER). Abasic sites are generated by glycosylases, processed by apurinic/apyrimidinic endonucleases and the oxidatively damaged base is replaced with an undamaged base. In contrast, nucleotide excision repair (NER) or dark repair is responsible for the repair of bulky adducts including UV-induced CPD and 6-4PP (Friedberg et al. [15], Friedberg [16]). NER can be divided into two interrelated sub-pathways: (i) Transcription coupled repair (TCR) which preferentially removes DNA damage at a faster rate from the transcribed strand of actively transcribed genes, and (ii) global genomic repair (GGR) which removes damage more slowly from throughout the entire genome and from the non-transcribed strand as well as the transcribed strand of active genes (as reviewed by Balajee & Bohr [2], Hanawalt [19]). In vertebrates, the processes of PR, BER and NER all contribute by varying degrees, to the the removal of DNA damage induced by aerobic respiration and sunlight. UVC induces DNA damage that is repaired primary by NER and PR, whereas oxidative-damaged DNA is repaired by BER.

Although the investigation of DNA repair in fish cells began more than two decades ago, there are still gaps in our understanding of the relative contribution of PR, TCR of NER, GGR of NER and BER in fish cells. Several fish cell lines have been reported to be relatively efficient in correcting UV-induced CPD by PR (Meador et al. [30], Wendi et al. [57]). In contrast, although there is evidence for the TCR sub-pathway of NER in fish cells (Komura et al. [25], [24]), they are characterized by a poor ability to perform the GGR sub-pathway of NER compared to human cells (Woodhead et al. [60], Shima et al. [47], Willett et al. [58]) and evidence for BER in fish cells is more limited (Ploch et al. [42], Walter et al. [55], Wendi et al. [57]). In human cells NER involves at least 25 polypeptides that function in recognition, excision, resynthesis of the repair patch and ligation (Sancar [46]). Individuals with the genetic diseases xeroderma pigmentosum (XP) and Cockayne syndrome (CS) have some deficiency in NER and cells from XP patients in complementation group A (XP-A) are deficient in both GGR and TCR (Friedberg et al. [15]).

Several previous reports have used host cell reactivation (HCR) of UV-irradiated plasmids to examine NER and PR of UVC-induced DNA damage in cultured fish cells (Mitani et al. [31], [32]). We have used recombinant non-replicating adenovirus constructs, which expresses the β-galactosidase (β-gal) reporter gene, to examine both constitutive and inducible BER of oxidative-damaged DNA and NER of UVC-damaged DNA in both human and rodent cells (Rainbow et al. 2005, Liu & Rainbow [29], Pitsikas et al. [41], Kassam & Rainbow [22]). UV-induced lesions in the template strand of active genes inhibit progression of RNA polymerase II (Donahue et al. [7]) and a single UV-induced cyclobutane pyrimidine dimer (CPD) is thought to be sufficient to inhibit reporter gene expression (Protic-Sabljic & Kraemer [43], Francis & Rainbow [12]). Using a recombinant adenovirus based HCR assay, we have previously shown that the extent of expression of (β-gal) from the UVC-damaged reporter gene in human cells is correlated with the degree of removal of CPD lesions, as measured by the ligation mediated-PCR method (Boszko & Rainbow [4]). Thus HCR of reporter gene activity is thought to require the repair of transcription blocking DNA lesions and reflect repair of DNA lesions in the transcribed strand.

The level of expression in cells infected by adenovirus vectors is greatly influenced by the promoter controlling expression of the transgene. The cytomegalovirus immediate early (CMV-IE) promoter is one of the most commonly used promoters in eukaryotic expression vectors, due primarily to its ability to yield high expression levels in many different mammalian cell types. In addition, Addison and colleagues (Addison et al. [1]) showed that the murine cytomegalovirus immediate early (MCMV-IE) rivals the human CMV-IE promoter for in vitro expression in both murine and human cells. AdCA35lacZ is a recombinant, non-replicating, human adenovirus that expresses the β-gal reporter gene under the control of the MCMV-IE promoter (Addison et al. [1]). In the present work, we were able to infect several cultured fish cell lines with AdCA35lacZ and express the β-gal reporter gene. This allowed us to examine HCR of β-gal expression for UVC-damaged and oxidative-damaged AdCA35lacZ in those fish cell lines capable of expressing the reporter gene. For an assessment of NER, cells were incubated in the dark and subsequently infected with either untreated or UVC-treated virus. In order to examine the extent of PR of UVC-induced DNA lesions, cells were also incubated in the presence of fluorescent light (400–700 nm) (FL) following infection with AdCA35lacZ. For an assessment of base excision repair (BER), cells were incubated in the dark and subsequently infected with either untreated virus or virus treated with photoactivated methylene blue. Methylene blue plus visible light (MB + VL) is known to create predominantly 8-oxoG in the DNA of cell free extracts (Floyd et al. [10]) as well as in whole cells and viruses (Epe et al. [9], Tuite & Kelly [52]) and the major repair mechanism of this oxidative DNA damage is BER (reviewed in Dianov et al. [6], Slupphaug et al. [49]).

HCR of the UV-treated reporter gene in the various fish cells varied from a level similar to that of the NER-deficient XP-A cells to values greater than that for the NER-proficient normal human fibroblasts. In addition the level of HCR correlated with β-gal expression levels for untreated AdCA35lacZ, suggesting a link between NER and transcription of the reporter gene in the fish cells. Exposure of cells to FL resulted in a substantial increase in HCR of the UVC-damaged reporter gene in all the fish cells examined, but not in the normal human fibroblasts, consistent with the detection of functional PR in the fish cells but not in the normal human fibroblasts. HCR of the MB + VL-treated reporter gene was similar to that in the normal human fibroblasts for the fish cell lines PBLE, CHSE-214, RTG-2 and RTS-pBk, but was significantly reduced in RT-Gill and RTS-34st cells.

The fish cell lines were Chinook salmon embryo (CHSE-214), carp epithelioma papulosum (EPC), zebrafish embryo (ZEB 2J), American eel leukocytes (PBLE), rainbow trout gonad (RTG-2), rainbow trout gill (RT-GillW1), rainbow trout spleen (RTS-34st) as well as a rainbow trout spleen cell line transfected with a neo-expression cassette (RTS-pBk+). All fish cell lines were obtained through Dr Carmel Mothersill (McMaster University, Ontario, Canada) and were a gift from Dr Neils Bols (University of Waterloo, Ontario, Canada). The cell lines CHSE-214, RTG-2, RT-GillW1, RTS-34st, RTS-pBk+ and EPC were originally obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA), whereas ZEB 2J, developed by Jerry Xing in Dr Bols' laboratory, is an original cell line derived from the zebrafish blastula stage embryonic stage stem cell line (ZEB2) that was originally obtained from ATCC. PBLE is a peripheral blood leukocyte cell line isolated in Dr Bols' laboratory from the American eel, Anguilla rostrata. The RTS34st line was isolated from the RTS34 line in Dr Bols' laboratory (Ganassin & Bols [17]). The EPC line was isolated from a herpes virus-induced hyperblastic lesion on the common carp, Cyprinus carpio (O'Dowd et al. [36]). Fish cell lines were grown in Leibovitz Media (L-15) supplemented with 2 mM L-Glutamine, 25 mM Hepes Buffer, antimycotic/antibiotic (100 μg/ml penicillin, 100 μg/ml streptomycin and 250 ng/ml amphotericin B, Gibco BRL, Grand Island, NY, USA) and 10% (RTG-2, CHSE-214, RT-Gill and EPC) or 15% (PBLE, RTS-34st, RTS-pBk and ZEB) fetal bovine serum (Gibco BRL, Grand Island, NY, USA). Fish cell cultures were trypsinized and seeded under minimal light conditions and maintained in seal cap tissue culture flasks in a light-tight box at room temperature (21–23°C).

The normal human fibroblast strain GM9503 and the XPA strain (XP12BE or GM5509) were obtained from the National Institute of General Medical Sciences (NIGMS) repository (Camden, NJ, USA). Except during the HCR experiments, the human cells were maintained in a humidified incubator kept at 5% CO2 and 37°C, and cultured in Eagle's α-MEM supplemented with 10% fetal bovine serum and antimycotic/antibiotic (100 μg/ml penicillin, 100 μg/ml streptomycin and 250 ng/ml amphotericin B, Gibco BRL, Grand Island, NY, USA).

AdCA35lacZ is a non-replicating, recombinant adenovirus containing the lacZ gene under the control of the MCMV-IE promoter inserted in the deleted E1 virus gene and was originally obtained from Dr F. L. Graham, McMaster University, Hamilton, Ontario, Canada (Addison et al. [1]). Virus stocks were prepared and titered in plaque forming units (PFU) per ml as previously described (Graham & Prevec [18]).

UVC-irradiation of virus has been described previously (Bennett & Rainbow [3]). Virus was suspended in 1.8 ml of cold PBS and irradiated in 35 mm dishes on ice with continuous stirring using a General Electric germicidal lamp (model G8T5) that emits predominantly at a wavelength of 254 nm with an incident fluence rate of 2 J/m2/s (as measured by a J-255 short-wave UV meter, Ultraviolet Products, San Gabriel, CA, USA). Aliquots of 200 μl were removed for each exposure to the virus and diluted appropriately with unsupplemented media. Under these conditions the induction of CPD in the transcribed strand of the adenovirus encoded lacZ gene was approximately 3.3×10−6 CPD/nucleotide/J/m2 as reported previously (Boszko & Rainbow [4]).

Preparation of MB and treatment of the virus with MB + VL has been described previously (Kassam & Rainbow [22], Pitsikas et al. [41]). 80 to 160 μl of stock AdCA35lacZ virus suspension was added to 3.6 ml of Phosphate Buffered Saline (PBS) containing 20 μg/ml MB in 35 mm Petri dishes on ice. With continuous stirring, the virus suspensions were irradiated (or mock irradiated) with visible light (VL). VL irradiation of virus employed a General Electric 1000-watt halogen lamp (GE R1000) at a distance of 70 cm from the bulb. After each time point, 400 μl of irradiated virus was removed, diluted appropriately in unsupplemented α-MEM and used to infect the cell monolayers.

Fish cells were seeded at 3×104 cells/well and primary human fibroblasts were seeded at 1.5×104 cells/well into 96-well plates (Falcon, Lincoln Park, NJ, USA). Some 18–24 h after seeding, growth media was aspirated and cells were infected with 40 μl of untreated, UVC-treated or MB + VL-treated AdCA35lacZ at an input multiplicity (MOI) of 100–400 plaque forming units (PFU) per cell for 90 min at room temperature (21–23°C). Infected cells were then overlaid with 160 μl of growth media and incubated for a further 24 or 40 h at 21–23°C and then assayed for β-galactosidase expression. In some experiments the primary human fibroblasts were incubated at 37°C for 40 h after infection and then assayed for β-galactosidase expression.

When infected cells were to be incubated in the presence of FL, the 96-well plates were placed above a light box employing a Westinghouse 32 watt circular fluorescent FC12T10/D 'Daylight' bulb emitting in the 400–700 nm wavelength range. Cells were exposed to light through the bottom of the 96-well plates at an incident exposure of 560 μW.

Cells were harvested at 24 or 40 h following the infection with AdCA35lacZ. The infected cell monolayer was incubated with 60 μl per well of 1 mM chlorophenol red β-d-galactopyranoside (CPRG; Roche, Indianapolis, IN, USA) in 0.01% Triton X-100, 1 mM MgCl2, and 100 mM phosphate buffer at pH 8.3. Light absorbance at 570 nm (A570) was determined several times following the addition of β-gal substrate using a 96-well plate reader (Bio-Tek Instruments EL340 Bio Kinetics Reader). Background readings of β-gal activity for uninfected cells were determined in all experiments and subtracted from readings of infected cells. Only readings for AdCA35lacZ infected cells that were twice background were included in the determination of β-gal expression and the β-gal survival curves.

Infection of Chinook salmon cells (CHSE-214), eel cells (PBLE) and four rainbow trout cell lines (RTG-2, RT-Gill, RTS-34st and RTS-pBk), but not zebrafish (ZEB) or carp (EPC) cells resulted in expression of β-gal. β-gal expression levels varied by more than 80-fold between the different fish cell lines when scored at 40 h after infection with undamaged AdCA35lacZ (Table I). Under the same conditions of incubation at 21–23°C, β-gal expression levels for undamaged AdCA35lacZ in the normal and XP-A fibroblasts were within the range detected for the fish cell lines.

Table I.  HCR of UVC-treated AdCA35lacZ in human primary fibroblasts and fish cells. Cells were infected at an MOI of 100–200 PFU/cell on the bench top at 21–23°C in the dark and β-galactosidase activity was measured 40 hours post infection. The average relative β-galactosidase expression levels of cells infected with untreated virus are listed in column A and the average absolute D37 values for HCR of UVC-treated AdCA35lacZ are listed in column B. The GM9503 normal human fibroblast was used in each experiment. The D37 value for each cell line relative to that for the GM9503 cells was determined for each individual experiment and the average relative D37 value for a number of independent experiments is shown in column C. Columns A, B and C show mean values and the standard error of the mean for a number of independent experiments as indicated in column D. (a) Significantly greater than 1 in a one-sample, two-tailed independent t-test (p < 0.05). (b) Significantly greater than GM5509 XP-A in a two sample independent t-test (p < 0.05). (c) Significantly different from 1 in a one sample two-tailed independent t-test (p < 0.05)

Since the fish cell lines grow at 21–23°C rather than 37°C, we first determined the effects of reduced temperature on HCR of the UVC-damaged reporter gene in NER–proficient normal human fibroblasts and NER-deficient XP-A fibroblasts. Relative β-gal expression of the UVC-damaged reporter gene in normal and XP-A fibroblasts incubated at room temperature (21–23°C) compared to incubation at 37°C are shown in Figure 1. HCR of the UVC-damaged reporter gene was reduced in NER–proficient normal human fibroblasts but not in NER-deficient XP-A fibroblasts incubated at room temperature (21–23°C) compared to incubation at 37°C. This is consistent with a reduced NER in human cells at lower temperatures due to temperature-dependent enzyme kinetics.

Graph: Figure 1. HCR of UVC-treated AdCA35lacZ in GM 9503 normal (▪, □) and GM 5509 XP-A (•, ˆ) primary human fibroblasts kept in the 5% CO2 incubator at 37°C (solid symbols) or at 21–23°C (open symbols). Shows representative results for cells infected with AdCA35lacZ virus at MOI 200 PFU/cell and β-gal activity was measured 40 hours post-infection. Error bars indicate the standard error of the mean for triplicate determinations.

Relative β-gal expression of the UVC-damaged reporter gene in the fish cells and the primary human fibroblasts is shown in Figure 2. The UVC fluence required to reduce β-gal activity to 37% of that for non-irradiated virus (D37) was used as a measure of HCR and compared to that obtained following infection of the NER proficient normal human fibroblast strain and the NER deficient XP-A fibroblast strain. Absolute D37 values and D37 relative to the NER-proficient normal fibroblast strain are shown in Table I. NER of the UVC-damaged reporter gene was significantly greater than that in the NER-deficient XP-A cells in eel cells (PBLE) and the rainbow trout cell lines RTS-34st and RTS-pBk but not in the Chinook salmon cells (CHSE-214) and the rainbow trout cell lines RT-Gill and RTG-2. HCR was significantly greater in the PBLE cells compared to the normal human fibroblasts.

Graph: Figure 2. HCR of UVC-treated AdCA35lacZ in primary human fibroblasts (open symbols) and fish cells (closed symbols). Shows representative results for cells infected with AdCA35lacZ at MOI 100–200 PFU/cell, incubated for 40 h at 21–23°C in the dark and then assayed for β-gal activity. Error bars indicate the standard error of the mean for triplicate determinations.

It has been suggested that a stalled transcription complex due to transcription blocking UV-induced photolesions is a signal for the recruitment of the DNA incision complex and subsequent removal and repair of the lesion by TCR in mammalian cells (Friedberg [16]). In addition, the induction of transcription results in increased repair of UV-induced lesions in the metallothionein gene of Chinese hamster ovary cells compared to the same gene in the uninduced state (Okumoto & Bohr [37]). This suggests that the level of gene expression influences the level of repair of lesions from transcribing genes by the TCR sub-pathway of NER in mammalian cells. It was therefore considered of interest to examine the relationship between expression of the undamaged reporter gene and HCR of UVC-induced lesions in the transcribed strand of reporter gene in the fish cells. The correlation of HCR for the UVC-damaged reporter gene and expression of the undamaged reporter gene (r = 0.85, p < 0.0001) shown in Figure 3 suggests a link between TCR of UVC-induced lesions and transcription of the reporter gene in the fish cells.

Graph: Figure 3. D37 value for HCR of UVC-damaged AdCA35lacZ in the fish cell lines versus relative β-gal expression levels of cells infected with undamaged virus. Each data point represents the D37 value obtained in a given experiment and the relative β-gal expression level obtained in that same experiment. The number of data points for each cell line is indicated in Table I.

We have reported previously that the expression of a replication deficient recombinant adenovirus reporter gene construct driven by a CMV-IE promoter is increased following exposure of human cells to various DNA damaging agents including oxidative DNA damage (Zacal et al. [62], Pitsikas et al. [41]). Members of the retinoblastoma protein family, but not p53 play an essential role in mediating DNA damage inducible expression from the CMV-IE promoter (Francis & Rainbow [14]). Since FL exposure has been shown to generate ROS within cells and produce oxidative DNA lesions (Lipinski et al. [28]) we were interested in the effects of FL exposure on the expression of the undamaged reporter gene. Results for expression of the undamaged reporter gene at 24 h after infection with AdCA35lacZ following exposure to FL in fish cells and normal human fibroblasts are shown in Table II. It can be seen that FL exposure resulted in varying degrees of increase in expression of the undamaged reporter gene in the normal human fibroblasts and the fish cells (Table II, column B).

Table II.  Effects of fluorescent light exposure on β-gal expression of the undamaged reporter gene. Normal human primary fibroblasts and fish cells were infected with AdCA35lacZ at an MOI of 200–400 PFU/cell and then kept on the bench top at 21–23°C in the dark or in the presence of fluorescent light for 24 hours, after which time β-gal activity was measured. The β-galactosidase expression level in each cell line relative to that in the CHSE line was determined in each of 2–3 independent experiments and the average relative β-galactosidase expression levels of cells infected with untreated virus are listed in column A. The relative β-galactosidase expression level in cells exposed to light relative to that in cells not exposed to light were determined for each cell line in each of 2–3 experiments and the average relative β-galactosidase expression levels are shown in column B. Columns A and B show mean values and the standard error of the mean for a number of independent experiments as indicated in column C. *Significantly different from 1 in a one-sample, two-tailed independent t-test (p < 0.05)

In order to examine the extent of direct PR of UV-induced DNA lesions in fish cells, we incubated cells in the presence or absence of FL following infection with untreated and UVC-treated AdCA35lacZ. Relative β-gal expression of the UVC-damaged reporter gene in the fish cells and normal human fibroblasts incubated in the presence or absence of FL is shown in Figure 4. Absolute D37 values and D37 relative to the NER-proficient normal fibroblast strain are shown in Table III. Exposure to FL resulted in a substantial increase of 3- to 6-fold in HCR of the UVC-damaged reporter gene in all the fish cell lines examined, but not in the normal human fibroblasts (Figure 4, Table III), consistent with the detection of functional PR in the fish cells but not the normal human fibroblasts.

Graph: Figure 4. HCR of UVC-treated AdCA35lacZ in cells exposed to FL (open symbols) compared to cells kept in the dark (closed symbols). Shows representative results for GM 9503 normal human fibroblasts and fish cells infected with AdCA35lacZ at an MOI of 200–400 PFU/cell, incubated for 24 h at 21–23°C and then assayed for β-galactosidase activity. Error bars indicate the standard error of the mean for triplicate determinations.

Table III.  Effects of fluorescent light exposure on HCR of UVC-treated AdCA35lacZ in normal human primary fibroblasts and fish cells. Cells were infected with AdCA35lacZ at an MOI of 200–400 PFU/cell and then kept on the bench top at 21–23°C in the dark or in the presence of fluorescent light for 24 hours, after which time β-gal activity was measured. The average D37 values for HCR of UVC-treated AdCA35lacZ in the absence of fluorescent light exposure are listed in column A and the average D37 values in the presence of fluorescent light are listed in column B. For each cell line, the D37 value obtained in the presence of fluorescent light relative to that obtained in the absence of fluorescent light was determined in individual experiments and the average relative D37 value is listed in Column C. Columns A, B and C show mean values and the standard error of the mean for a number of independent experiments as indicated in column D. *Significantly greater than 1 in a one-sample, two-tailed independent t-test (p < 0.05)

In order to examine the level of BER in the fish cell lines we examined HCR of β-gal expression for MB+VL-treated AdCA35lacZ following incubation of infected cells at 21–23°C in the dark. Relative β-gal expression of the MB+VL-treated reporter gene is shown in Figure 5. The VL exposure in seconds required to reduce β-gal activity to 37% of that for non-treated virus (D37) was used as a measure of HCR and compared to that obtained following infection of a BER proficient normal human fibroblast strain. Absolute D37 values and D37 relative to the BER-proficient normal fibroblast strain are shown in Table IV. HCR of the MB+VL-treated reporter gene was similar to that in the normal human fibroblasts for the fish cell lines PBLE, CHSE-214, RTG-2 and RTS-pBk, but was reduced in RT-Gill and RTS-34st cells. HCR of the MB+VL-treated reporter gene was greatest in the PBLE cells compared to the other fish cell lines.

Graph: Figure 5. HCR of MB + VL-treated virus in normal human fibroblasts (open symbols) and fish cells (closed symbols). Shows pooled results of three to four experiments each done with triplicate determinations for GM 9503 normal fibroblasts and fish cells infected with AdCA35lacZ at an MOI of 200–400 PFU/cell, incubated for 24 h at 21–23°C in the dark and then assayed for β-galactosidase activity. Error bars show the standard error of the mean for three to four independent experiments as indicated in Table IV.

Table IV.  Host cell reactivation of MB + VL-treated AdCA35lacZ in normal human fibroblasts and fish cells. Cells were infected with AdCA35lacZ at an MOI of 200–400 PFU/cell, incubated in the dark for 24 h at 21–23°C and then assayed for β-gal activity. Table shows the mean absolute D37 value (column A) and the D37 value for each cell line relative to that for the GM9503 normal human fibroblasts (column B). Columns A and B show mean values and the standard error of the mean for a number of independent experiments as indicated in column C. *Significantly different from 1 in a one-sample, two-tailed independent t-test (p < 0.05)

Infection of several of the fish cell lines with AdCA35lacZ resulted in expression of the MCMV-IE promoted β-gal reporter gene. β-gal expression levels varied by more than 80-fold between the different fish cell lines and were of the same order of magnitude as expression levels in the normal and XP human cells when scored at 40 h after infection and under the same conditions of incubation at 21–23°C.

Adenovirus entry into cells generally involves attachment to a primary receptor followed by interaction with a secondary receptor responsible for internalization. The primary receptor for most human adenoviruses is a 46-kDa transmembrane protein, the coxsackievirus and adenovirus receptor (hCAR) to which the knob domain of the Ad fibre binds (Zhang & Bergelson [61]). AdCA35lacZ used in the current study has knob fibre with standard tropism for hCAR. The ability of AdCA35lacZ to infect and express detectable levels of the β-gal reporter gene in CHSE-214, PBLE, RTG-2, RT-Gill, RTS-34st and RTS-pBk fish cells suggests that these fish cells express a functional homologue of hCAR. Our inability to infect and express the reporter gene in ZEB or EPC cells suggests these fish cells are CAR-negative. This is consistent with the results of Overturf and colleagues who have used competition with fiber knob blocking to show that CHSE-214 cells express a functional CAR homologue, whereas EPC cells do not (Overturf et al. [39]). Our inability to infect and detect expression of the reporter gene in ZEB cells is surprising since Petrella and colleagues have reported a zebrafish homologue of CAR (zCAR) that is sufficiently similar to hCAR to permit attachment and infection by Ad5 when expressed in Chinese hamster Ovary (CHO) cells (Petrella et al. [40]). Since expression of our β-gal reporter gene is MCMV-IE promoter based, the activity of the MCMV-IE promoter could be an issue. However, the ability to express a CMV promoted reporter gene in several other fish cell lines reported here and by others (Lee et al. [27], Overturf et al. [39]) would suggest this promoter is active in fish cells. Since we infected the different fish cell lines at the same viral input multiplicity we expect that the same number of copies of the reporter gene enter each cell. We therefore consider it most likely that the differences in expression of the reporter gene in the various fish cell lines reflect differences in the activity of the MCMV-IE promoter in the various fish cell lines, rather than differences in the number of copies of the reporter gene entering each cell.

Fish cells have poor NER ability in the genome overall, although they have proficient PR (Regan & Cook [45], Shima et al. [47], Shima & Setlow [48]). Willet and colleagues used alkaline gel electrophoresis to look at the repair of endonuclease sensitive sites (ESS) in two catfish species. No detectable repair was seen in total DNA extracted from primary hepatocytes following UV light exposure of (10–40 J/m2) when assayed up to 72 h after exposure (Willett et al. [58]). Although repair of the genome overall by GGR is low, there is evidence for TCR in fish cells. Komura and colleagues reported that NER in the RBCF-1 goldfish cell line is preferential for transcriptionally active genes, although PR is not (Komura et al. [24]).

In the present work we have used a HCR technique to measure repair of UVC-induced DNA damage in the transcribed strand of an actively transcribing adenovirus encoded reporter gene. In the absence of FL exposure, HCR of the UVC-damaged reporter gene for the fish cells varied from values similar to that of the NER-deficient XP-A cells to values greater than that for the NER-proficient normal human fibroblasts (Figure 2 and Table I). This indicated significant repair in the transcribed strand of the reporter gene in several of the fish cell lines. In particular, the PBLE line showed both the greatest expression and the greatest repair of the reporter compared to the other fish lines as well as the NER proficient normal human fibroblasts. It is possible that the variation in HCR among the fish cell lines and the particularly high HCR value in the PBLE cells might reflect other repair pathways in addition to NER, such as repair by intragenic homologous recombination, since multiple copies of the reporter gene enter the cell. However, UVC-induced lesions (CPD and 6-4PP) in the reporter gene at the UVC exposures employed here are predominantly intrastrand lesions, the vast majority of which are thought to be repaired by NER (Friedberg [16]). We therefore consider it most likely that variations in HCR of UVC-damaged reporter genes reflect, to a major degree, variations in the rate of repair of the reporter gene by NER. In addition, we show that HCR values correlated with β-gal expression for untreated AdCA35lacZ in the fish cells, suggesting a link between NER in the transcribed strand and transcription of the reporter gene (Figure 3). A similar dependence of NER in the transcribed strand of a plasmid encoded RPB2 reporter gene on transcription by RNA polymerase II has been reported in yeast cells (Sweder & Hanawalt [51]).

HCR of the UVC-damaged reporter gene was reduced in NER–proficient normal human fibroblasts but not in NER-deficient XP-A fibroblasts incubated at room temperature (21–23°C) compared to incubation at 37°C. This is consistent with a reduced NER in human cells at lower temperatures due to temperature-dependent enzyme kinetics. A reduced incision rate of UVC-induced DNA damage at 20 and 25°C compared to 37°C has been reported for the TO-2 fish cell line derived from the ovary of an adult Tilapia (Chen & Yew [5]).

Exposure to FL resulted in a substantial increase of 3- to 6-fold in HCR of the UVC-damaged reporter gene in all the fish cells examined, but not in the normal human fibroblasts, consistent with the detection of functional PR in the fish cells but not normal human fibroblasts (Figure 5 and Table III). Although PR is present in human cells (Sutherland & Bennett [50]) the ability of HCR assays to detect PR of UVC-damaged DNA in human cells has been successful in excision deficient XP cells but not in normal excision proficient human cells (Wagner et al. [54], Henderson [21]) consistent with our results for normal human fibroblasts. The inability to detect PR in normal human cells using HCR assays may result from competing dimer repair by NER or that PR is repressed in NER proficient cells and derepressed in NER deficient cells. We were only able to measure PR in the presence of NER. Since both mechanisms can repair the same types of photodamage (Mitchell & Nairn [34]) we cannot evaluate the relative contribution of each repair mechanism to the overall repair of the reporter gene when both mechanisms are operating. Olson and Mitchell ([38]) failed to detect PR in northern pike, brook trout and rainbow trout larvae and have suggested that one possible reason for this loss of PR may reflect evolutionary changes in species that spawn in low water temperatures. Our ability to detect substantial PR in several rainbow trout cell lines derived from gonad, gill and spleen suggests a regulation of PR during development as has been demonstrated in other species (Mitchell & Hartman [33]).

HCR of the MB + VL-treated reporter gene was similar to that in the normal human fibroblasts for the fish cell lines PBLE, CHSE-214, RTG-2 and RTS-pBk (Figure 5 and Table IV) suggesting efficient BER in several fish cell lines. In contrast, a significantly reduced HCR of the MB+VL-treated reporter gene was detected in rainbow trout gill (RTGill) and spleen (RTS-34st) compared to rainbow trout gonad (RTG-2), suggesting a reduced BER of oxidative damage in gill compared to gonad tissue of the rainbow trout. A reduced uracil initiated BER of a G:U mismatch has been reported for gill and liver tissue compared to brain tissue in hybrid fish of the genus Xiphorphorous (Wendi et al. [57]). Taken together, this suggests there are tissue specific differences in BER for fish.

We have reported previously that treatment of human cells with DNA damaging agents (including oxidizing agents) results in enhanced expression of the CMV-IE promoted reporter gene. The enhanced expression is induced at lower levels of DNA damage and at higher levels in TCR-deficient compared to TCR-proficient cells, indicating that persistent unrepaired DNA damage in active genes triggers increased activity from the CMV-IE driven reporter construct (Francis & Rainbow [13]; Zacal et al. [62], Pitsikas et al. [41]). We report here that exposure to FL resulted in varying degrees of increased expression of the undamaged reporter gene in the normal human fibroblasts and the fish cells (Table II). These results indicate that unrepaired oxidative damage induced by FL can also act as a trigger for increased activity from the CMV-IE driven reporter. The low level of increased expression of the reporter gene in PBLE cells suggests a more efficient removal of oxidative DNA damage by a TCR pathway of BER in PBLE cells compared to the other fish cells. This is consistent with the greater HCR for MB + VL-induced oxidative damage in PBLE cells (Table IV).

In summary we show that the recombinant adenovirus encoded reporter gene can be used to examine BER, NER and PR in several fish cell lines. These repair pathways in fish are critical for the removal of DNA damage induced by a variety of physical and chemical agents. Heavy metal exposure has been reported to increase ROS production and compromise the integrity of NER and BER in mammals (Hartwig [20], Valko & Morris [53]). Use of the recombinant adenovirus based HCR assay could be of value in examining the integrity of BER, NER and PR in fish cells following their treatment with physical and chemical environmental pollutants in order to more fully understand the impact of the environmental changes of lake and sea pollution and thinning of the ozone layer on fish.

This work was supported in part by the National Cancer Institute of Canada with funds from the Canadian Cancer Society.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.