CrFK Feline Kidney Cells Produce an RD114-Like Endogenous Virus That Can Package Murine Leukemia Virus-Based Vectors

Abstract

The feline kidney cell line CrFK is used extensively for viral infectivity assays and for study of the biology of various retroviruses and derived vectors. We demonstrate the production of an endogenous, RD114-like, infectious retrovirus from CrFK cells. This virus also is shown to efficiently package Moloney murine leukemia virus vectors.

CrFK feline kidney cells have been used extensively both to propagate and to investigate the biology of a number of enveloped viruses, including retroviruses such as feline immunodeficiency virus (FIV) (1, 2, 6, 19, 21, 27, 32), feline leukemia virus (FeLV) (23), mouse mammary tumor virus (MMTV) (13, 20, 24), herpesviruses (7, 35), coronaviruses (3, 12), and feline syncytial virus (18). Further, retrovirus vectors based on MMTV (4, 25) and FIV (22) have been constructed with CrFK cells.

We investigated the production of type C retroviruses from CrFK cells which might play a role in the mobilization of transfected viral genes, other viruses, or viral vectors. For these studies, CrFK cells were obtained on two independent occasions from the American Type Culture Collection (CCL-94). With a 1:500 dilution of an antibody (goat anti-RD114 CA) specific for the capsid protein of RD114, the most common feline endogenous retrovirus (15), and a 1:1,000 dilution of secondary anti-goat alkaline phosphatase-conjugated antibody (Vector Laboratories, Burlingame, Calif.), a protein of 28 kDa, the expected size of the RD114 capsid protein (Fig. 1A), was detected in membrane extracts from noninfected parental CrFK cells (Fig. 1B, lane 5). Using the same antiserum, we also detected FeLV capsid protein (p26) (Fig. 1A) in feline embryonic fibroblasts (FEA) (10) infected with FeLV (Fig. 1B, lanes 2 and 3). This antiserum also detected Moloney murine leukemia virus (MuLV) Gag processing precursors (Fig. 1A) in CrFK cells transfected with an MuLV Gag-Pol expression construct (14) (Fig. 1B, lane 4). The lack of detectable proteins of 26 kDa or of processing intermediates in the parental CrFK cells (Fig. 1B, lane 5) suggests that the major type C retrovirus Gag protein produced is related to RD114.

FIG. 1.

(A) Schematic representation of the Gag proteins of FeLV, RD114, and MuLV. The sizes of the precursor proteins as well as the processed end products are shown. (B) Expression of RD114-related Gag proteins in QN10S cells (lane 1), FeLV-infected FEA cells (lanes 2 and 3), MuLV Gag–Pol-transfected CrFK cells (lane 4), and CrFK cells (lane 5) in membrane extracts with antiserum directed against the RD114 CA protein, after application of protein extract from 10⁶ cells to a 7.5% sodium dodecyl sulfate-polyacrylamide gel, resolution at 40 V overnight, and transfer (2.5 mA/cm², 1 h) to a nitrocellulose (Schleicher and Schüll) membrane (8). The size of the marker protein (30 kDa) is indicated, as is the RD114 Gag protein p28 and the FeLV Gag protein p26.

To investigate whether CrFK cells produce infective, RD114-related virions (CrLE virus), supernatant was taken from these cells and used to infect feline QN10S fibroblasts (9), previously shown not to express any RD114-related Gag proteins (Fig. 1B, lane 1). After passage of these initially infected QN10S cells for 4 days, viral proteins were extracted from 0.45-μm-pore-size-filtered 24-h supernatant from 5 × 10⁶ cells and analyzed by Western blotting with the same antiserum directed against RD114 p28^Gag. In contrast to the supernatant from noninfected QN10S cells (Fig. 2, lane 6), supernatant from QN10S cells infected with putative CrLE virus from CrFK cells (Fig. 2, lane 4) contained a 28-kDa protein with a mobility similar to that found in supernatant from CrFK cells, RD114-producing FER cells (Fig. 2, lane 1), and QN10S cells infected with RD114-containing supernatant (Fig. 2, lane 3). These results demonstrate the presence of infective, biologically active CrLE virus produced from CrFK cells.

FIG. 2.

Western blot analysis of viral proteins in supernatants from RD114-producing FER cells (lane 1), QN10S cells infected with RD114 (lane 3), QN10S cells infected with CrLE virus (lane 4), CrFK cells (lane 5), and QN10S cells (lane 6) with antiserum directed against the RD114 CA protein. The size of p28 Gag is indicated. Lane 2 is empty.

Infection of CrFK cells with an MuLV-based retrovirus vector, LXSN-EGFP, carrying neomycin resistance and enhanced green fluorescent protein genes (11) packaged in an amphotrophic packaging cell line, PA317 (17), allowed the establishment of neomycin (G418)-resistant, transduced CrFK clones. These were analyzed for the ability of the CrLE virus to provide helper functions in trans and to give rise to recombinant virus in which the LXSN-EGFP genomic RNA was packaged into CrLE Gag, Pol, and Env proteins. Two milliliters of 0.45-μm-pore-size-filtered 24-h supernatant from LXSN–EGFP-infected CrFK clones was used to infect 10⁶ target mink lung cells (Mv1Lu) (5), known to be infectable with RD114 (28). After selection in 800 μg of G418 (Life Technologies) per ml, titers of around 10⁵ CFU/ml were obtained (Table 1), similar to the titers obtained when the same cells were infected with LXSN-EGFP vector generated by a three-way transfection (29) of COS-7 cells (16) with the vector, an MuLV Gag-Pol expression construct (pGagpol-gpt) (14), and either an RD114 Env expression construct (pRDLF) (Table 1) or a vesicular stomatitis virus G (VSV-G) protein expression construct (pHCMV.G) (34) (Table 1). Thus, the CrLE infectious virus is able to package MuLV genomic RNA, as has been previously reported for RD114 (31).

TABLE 1.

Titers of various viruses on mink lung cells

Envelopea	Titer (CFU/ml)b
	Mv1 Lu	MinkRD
CrLE	8.6 × 104 ± 2.1 × 104	0
RD114	2.2 × 104 ± 1.2 × 103	0
VSV-G	1.8 × 105 ± 5.1 × 104	1.5 × 105 ± 5.9 × 104

^aCrLE, recombinant LXSN-EGFP/CrLE virus; RD114, recombinant LXSN-EGFP/MuLV Gag-Pol and RD114 Env virus; VSV-G, recombinant LXSN-EGFP/MuLV Gag-Pol and VSV-G surface protein virus. The RD114 and VSV-G recombinant viruses were generated by three-way transfection (29) of LXSN-EGFP (11), MuLV Gag-Pol (14), and either an RD114 Env (provided by Yasuhiro Takeuchi) or a VSV-G (34) expression construct (3 μg/3 μg/3 μg) (CellPhect; Pharmacia). 

^bMean titers and standard errors of the mean of five (CrLE and VSV-G) or three (RD114) independent experiments on mink lung cells (Mv1 Lu) or mink lung cells infected with RD114 (MinkRD). 

The receptor usage of the CrLE virus was determined by a receptor-interference assay. The same virus-containing supernatants described above were used to superinfect mink lung cells already productively infected with RD114 (MinkRD [provided by Yasuhiro Takeuchi]) and assayed either by neomycin resistance after G418 selection or by fluorescence-activated cell sorter (FACS) analysis for EGFP-positive cells (11). Only the VSV-G pseudotyped virus was able to infect these cells (Table 1), demonstrating that the CrLE virus utilizes the same receptor as RD114 (26).

Finally, the infection spectrum of the CrLE virus was investigated. As has already been reported for the RD114 virus (26), NIH 3T3 cells could not be infected with CrLE virus (Table 2), thus demonstrating that the virus is not contaminated by replication-competent, amphotropic MuLV. In contrast, rat XC cells (30), mink lung cells, and simian COS-7 cells could be infected by CrLE (Table 2). Interestingly, a titer was also sporadically recorded on CrFK cells, in two of five experiments (Table 2), suggesting either that not all CrFK cells produce CrLE virus or that not all receptor molecules are blocked. Early work by Lasfargues and colleagues (13) with immunofluoresence assays suggested that up to 5% of CrFK-F2 cells express RD114-related proteins. However, if these cells were producing infectious virus able to infect CrFK cells (ecotropic or amphotropic), we would expect the cells to rapidly become 100% infected. RD114 is known as a xenotropic virus, yet the closely related, if not identical, CrLE virus can infect the cell line from which it is produced, albeit inefficiently. RD114 has also been observed to occasionally infect feline cells (19a).

TABLE 2.

Survey of infectivity of CrLE virus^a

Target cells	Titer
NIH 3T3	0
CrFK	9.3 × 101 ± 8.5 × 101
XC	5.1 × 102 ± 6.1 × 101
Mv1Lu	7.0 × 104 ± 1.8 × 104
COS-7	5.2 × 103 ± 3.3 × 103

^aNIH 3T3 mouse fibroblasts, CrFK feline kidney cells, XC rat sarcoma cells, Mv1Lu mink lung cells, and COS-7 simian kidney cells were infected with recombinant LXSN-EGFP/CrLE virus, and titers were determined by the generation of G418-resistant colonies (for NIH 3T3, Mv1Lu, and COS-7 cells [CFU/ml]) or by FACS analysis for EGFP activity (for CrFK and XC cells [number of infected cells/ml]). Averages and standard errors of the mean of five (NIH 3T3, CrFK, XC, and Mv1Lu) or three (COS-7) independent experiments are shown.  

Our data demonstrate that CrFK cells produce an infectious virus which shares all tested properties with the RD114 virus. This virus is able to mobilize murine type C retroviruses by acting as a helper and, since it is an enveloped virus, is expected to form pseudotypes with other enveloped viruses (33). CrFK cells are commonly used both in classic diagnostic virology and in molecular virology. Thus, the possibility of mobilization of heterologous retroviral sequences, and potentially more importantly the possibility of generating pseudotyped virus with diagnostically relevant nonhuman pathogenic viruses, should be considered. Since we have shown that the CrLE virus can infect COS-7 cells and RD114 has been shown to infect other human cells (28), this could allow a normally host-restricted virus entry into human cells. Also, much of the biology of FIV, and to some extent MMTV, has been elucidated with CrFK cells. Indeed, CrFK is one of the few cell lines known to be permissive for production of MMTV (13, 20, 24). Although it is not clear whether the CrLE virus can act as a helper for FIV or type B viruses, the formation of pseudotypes may also contribute to infection (33). We have shown that MMTV virions produced in CrFK cells carry the MMTV envelope and core proteins (24) and that MMTV vectors give infectious titers on NIH 3T3 cells (4, 25), suggesting that the RD114 envelope was not involved in specifying the infection here. Even so, caution should be observed in the choice of cell lines as starting points for the generation of retrovirus packaging cell lines.