Department of Microbiology and Immunology and
Markey Cancer Center, University of Kentucky, Lexington, Kentucky
40536-0096
As a consequence of being diploid, retroviruses have a high
recombination rate. Naturally occurring retroviruses contain two repeat sequences (R regions) flanking either end of their RNA genomes,
and recombination between these two R regions occurs at a high rate. We
deduced that recombination may occur between two sequences within the
same RNA molecule (intramolecular) as well as between sequences present
within two separate RNA molecules (intermolecular). Intramolecular
recombination would usually result in a deletion within the progeny
provirus. In this report, we demonstrate that intramolecular
recombination between two identical sequences occurred within a
chimeric RNA vector. In addition, high rates of recombination between
two identical sequences within the same RNA molecule resulted mostly
from intramolecular recombination.
 |
INTRODUCTION |
The presence of two identical
genomic RNA molecules in retroviral virions results in a high rate of
recombination (3). The homologous intermolecular
recombination rate is 4 × 10
5 per base per
replication cycle (6). When nonhomologous RNA is present
in these dimer RNA molecules, nonhomologous intermolecular recombination can occur, although the rate is very low, only 0.1% of
the rate of essentially homologous recombination (21). A similar high rate of recombination between short region of identities in the midst of otherwise nonidentical sequences also occurs. This
recombination depends on the length of sequence identity (23). Retroviruses contain two identical repeat sequences
called R regions, and recombination between these two R regions has
been observed (8, 10, 14).
Little is known about the physical relationship of the two RNA
molecules contained in the retroviral virion. During reverse transcription, if recombination does not occur, the minus- and plus-strand DNA primers are provided almost completely intramolecularly (7). This implies that the two RNA molecules are not
randomly intertwined in a virion (7). To study
recombination, two different proviruses were introduced into one
retroviral helper cell line. The recombinants were then judged by the
expression of markers from both parental proviruses. Evidence shows
that intermolecular recombinations occur frequently (6,
22). To further study recombinations between two identical
sequences within the same RNA molecules, a Moloney murine leukemia
virus (MLV)-based vector was constructed (19). This vector
contains a complete hygromycin resistance gene (hyg) and a
color marker reporter gene, gfp (green fluorescence
protein). Additionally, this vector carries an extra 3' hyg
gene segment 290 bp in length inserted into the 3' untranslated region
of gfp (Fig. 1A). After one
round of replication, deletion of gfp resulted from
recombination between the downstream 3' hyg sequence and the
upstream 3' hyg sequence. The ratio of the number of
colonies without the gfp gene phenotype to the total number of hygromycin-resistant (Hygr) colonies represented the
rate of recombination between the two identical sequences within the
same RNA molecule. The rate of deletion in the progeny proviruses was
about 62% (19). However, the high rate of recombination
can be either intra- or intermolecular.

View larger version (25K):
[in this window]
[in a new window]
|
FIG. 1.
Structures of retrovirus vectors used to determine the
recombination rate between two identical sequences in the same RNA
molecule. (A) Structure of a retrovirus containing two identical
sequences. JZ422 + 3' Hyg contains hyg, gfp, and an
additional 290 bp of the 3' hyg sequence downstream of
gfp. After one round of replication, the downstream 3'
hyg sequence will recombine with the identical upstream
hyg sequence and result in the deletion of gfp.
Recombinants, therefore, contain only hyg. Dotted lines
between the two recombinants indicate the identical 3' hyg
sequences, (B) Structures of a chimeric RNA, infectious virus vectors,
and resulting recombinants. JZ446 contains only the 5' MLV LTR, while
the infectious vector LN contains two MLV LTRs. JZ446 contains
hyg, gfp, and an additional 290 bp of the 3' hyg
sequence downstream of gfp. Recombinant proviruses, which
contain hyg, form only when recombination occurs between the
JZ446 and the infectious vector so that hyg is flanked by
two LTRs. Recombinations between JZ446 and LN are nonhomologous
(22). Most recombinations between JZ446 and LN destroy
gfp. (C) Structures of a chimeric RNA and infectious virus
vector containing 80-bp sequence identities and resulting recombinants.
LN80 is identical to LN except for containing an 80-bp sequence in
common with JZ446. After one round of replication, most recombinants
utilized the 80-bp sequence identities to recombine with JZ446 to form
a recombinant containing gfp flanked by two identical 290-bp
sequences within the same RNA molecule. Eighty-three percent of
recombinants with the two identical 290-bp sequences have undergone an
intramolecular recombination resulting in a deletion of gfp.
The dotted lines between JZ446 and LN80 indicate the identical 80-bp
sequences; the dotted lines between the two recombinants indicate the
identical 290-bp sequences.
|
|
It was deduced that recombination may occur between two sequences
within the same RNA molecule (intramolecular) as well as between
sequences present within two separate RNA molecules (intermolecular) (17). Intramolecular recombination would result in a
deletion within the progeny provirus. However, there is no direct
evidence indicating that intramolecular recombination occurs. This is
because, in the experiments described above, retroviral particles were able to package either two different RNA molecules or two identical RNA
molecules. The deletion can result from an intramolecular event or from
the recombination between a downstream sequence of one RNA molecule and
the upstream sequence of the other RNA molecule. Since current
technology cannot separate the viral particles containing two identical
molecules (homogeneous) from those containing two different molecules
(heterogeneous), deletion recombinants cannot be excluded as arising
from intermolecular recombination.
In this report, we demonstrate that intramolecular recombinations
actually occurred between two identical sequences within a chimeric RNA
vector. Furthermore, the rate of intramolecular recombination was
determined. The high rate of recombination between two identical
sequences within the same RNA molecules resulted from intramolecular
recombination. This observation was consistent with those regarding
intramolecular transfers of DNA primers during retroviral reverse
transcription (7).
 |
MATERIALS AND METHODS |
Nomenclature.
Plasmids are designated as, for example,
pJZ446; viruses made from these plasmids are designated as, for
example, JZ446. Some infectious MLV vectors contained a 290-bp sequence
derived from the 3' end of hyg (Fig.
2). When the 290-bp sequence was inserted at the 5' end of the neomycin resistance gene (neo) the
number of nucleotides inserted is on the left of the "N" (N stands
for neo) (for example, pL290N) (Fig. 2C). When the 290-bp
sequence was inserted at the 3' end of neo, the number of
nucleotides inserted is on the right of the "N" (for example,
pLN290) (Fig. 2B). An infectious MLV-based vector contained an 80-bp
sequence derived from the 5' end of the herpes simplex virus (HSV)
thymidine kinase (TK) poly (A) addition signal from the chimeric RNA
vector JZ446. Since the 80-bp sequence was inserted at the 3' end of
neo, the number of nucleotides inserted is on the right of
the "N," and the vector was designated pLN80.

View larger version (19K):
[in this window]
[in a new window]
|
FIG. 2.
Chimeric RNA vector JZ211, infectious virus vectors, and
resulting recombinants. JZ211 contains only the 5' MLV LTR, while the
infectious vectors LN, LN290, LN290N, and L290N290 contain two MLV
LTRs. The recombinant proviruses containing hyg form only
when recombination occurs between JZ211 and the infectious vector such
that hyg is flanked by two LTRs. Recombinations between
JZ211 and LN are nonhomologous (22). Most recombinations
between JZ211 and LN290, LN290N, and L290N290 occurred between the
identical 290-bp sequences. The dotted lines between the chimeric RNA
vector and the infectious vectors indicate the identical 290-bp 3'
hyg sequences.
|
|
Vector constructions.
All recombinant techniques were
carried out according to conventional procedures (15). All
vector sequences are available upon request.
(i) Construction of pJZ211 and pJZ446.
pJZ211 (Fig. 2), the
spleen necrosis virus (SNV) vector, and the truncated MLV vector
containing only hyg were previously described
(22). The chimeric RNA vector pJZ446 (Fig.
3), derived also from SNV, contained a
deletion in the U3 region of its 3' SNV long terminal repeat (LTR) and
an XhoI restriction site linker in the deletion site (Fig.
4). The vector also contains a truncated MLV vector between the two SNV LTRs, in the opposite transcriptional orientation to the SNV LTRs. This truncated MLV vector carried hyg, gfp, and an IRES (internal ribosome entry
segment) sequence between the two genes. An HSV TK poly (A) addition
signal replaced the deleted 3' MLV LTR. The IRES sequence of
encephalomyocarditis virus origin allows the ribosome to bind to the
internal AUG and initiate translation of gfp independently
of hyg (1, 2). pJZ442 + 3' Hyg (Fig. 1A)
(19) is an MLV vector containing hyg and
gfp separated by an IRES sequence plus an additional 3'
hyg segment at the 3' gfp gene. The
SgfI-HindIII (2.4-kb) fragment isolated from
pJZ442 + 3' Hyg, containing the 3' end of hyg, the IRES
sequence, gfp, and an additional copy of the 3'
hyg sequence, was inserted into the SgfI and
HindIII sites of pJZ211. The resulting plasmid was
designated pJZ446.

View larger version (26K):
[in this window]
[in a new window]
|
FIG. 3.
Chimeric RNA vector JZ446, infectious virus vectors, and
resulting recombinants. JZ446 is identical to JZ211 (Fig. 2) except
that it also includes gfp and an insertion of a second
identical sequence homologous to 290 bp of the 3' hyg
sequence into the 3' untranslated portion of gfp (downstream
of gfp or after the gfp stop codon). After one
round of replication, the JZ446 recombinants showed that the downstream
3' hyg sequence had recombined with the upstream
hyg sequence, resulting in a deletion of gfp.
These recombinants, therefore, are clear under a fluorescence
microscopy.
|
|

View larger version (38K):
[in this window]
[in a new window]
|
FIG. 4.
Outline of an experimental approach for determining the
rate of recombination during a single cycle of retroviral replication
between the chimeric RNA vector JZ446 (or JZ211) and infectious
vectors. Plasmid backbone sequences are not shown. Directions of
transcription in SNV and MLV are shown by long thin arrows.
Transfections are indicated by test tube shapes. Infections are
indicated by virion shapes. The different backgrounds represent the
indicated cell lines. SV, late polyadenylation signal of simian virus
40; and E, encapsidation sequences of MLV and SNV, respectively;
tk, thymidine kinase. The lines in the LTR separate the U3, R, and U5
regions.
|
|
(ii) Construction of pLN, pLN290, pL290N, pL290N290, and
pLN80.
All infectious vectors in this study are
neo-carrying MLV-based vectors. Vector pLN (13)
contains neo flanked by two LTRs (Fig. 1B). pLN290, pL290N,
and pL290N290 (Fig. 2B to D and 3B to D) contain one or two copies of
the 3' end of hyg and were described previously (19,
23). In pLN80, an 80-bp sequence derived from the 5' end of the
HSV TK poly(A) addition signal was inserted into the 3' untranslated
region of neo (Fig. 1C). Therefore, pLN80 was identical to
pLN except for containing the additional 80-bp sequence. To insert the
80-bp sequence of the TK segment into pLN, the TK sequence within JZ446
was amplified by PCR. The first primer hybridized with the MLV
packaging signal, and the second primer
(5'-TCTTATCGATTGCCGTCATAGCGC-3') hybridized with the 3' end
of the TK segment. The HindIII-ClaI fragment
of the amplified DNA was inserted between the HindIII
and ClaI sites of pLNCX (13). The inserted
sequence in the resulting vector was confirmed by DNA sequencing.
(iii) Construction of JZ525 and YM2.
The pJZ525 construct
(Fig. 5A), from 5' to 3', was assembled
as follows. The 2.5-kb BamHI-NotI fragment (from
positions 1630 to 4112) was isolated from pJZ442 + 3' Hyg (Fig.
1A) (19) and contained hyg, the IRES, and
gfp. The 0.6-kb NotI-NdeI fragment (positions 4113 to 4650) contained an identical IRES sequence. The
5.4-kp NdeI-BamHI fragment (positions 4651 to
1629) was isolated from pLN (13) and contained
neo and the two MLV LTRs. The NcoI site within
neo was digested, followed by repair with the Klenow fragment. NcoI digestion created two DNA ends that contained
a four-base overhang. Repair of the overhang by the Klenow fragment created two blunt ends. Ligation of these two blunt ends with T4 ligase
created a 4-bp insertion, which shifted the neo open reading
fragment by one (+1). The retroviral vector YM2 was identical to JZ425
except that YM2 contained a different frameshift mutation at the
SgfI site within the hyg open reading frame,
while neo was functional.

View larger version (16K):
[in this window]
[in a new window]
|
FIG. 5.
Structures of retrovirus vectors used to determine the
rate of intramolecular recombination. (A) Structure of a retroviral
vector encoding a frameshift mutation within neo. JZ525
contains, from 5' to 3', hyg, an IRES sequence,
gfp, an additional copy of the IRES sequence, and
neo. The neo gene encoded a frameshift mutation
at its NcoI site and thus was nonfunctional. (B) Structure
of a retroviral vector encoding a frameshift mutation in
hyg. YM2 was identical to JZ525 except that neo
was functional but hyg was nonfunctional because of a
frameshift mutation at the SgtI site within the
hyg open reading frame. (C) Recombinant of JZ525 and YM2.
This recombinant encodes functional hyg, neo, and
gfp. (D) Recombinant of JZ525 and YM2. This recombinant is
identical to C except that the gfp gene between the two IRES
sequences is deleted.
|
|
Introduction of a single chimeric RNA vector and a single
infectious vector into helper cell line PG13.
Retroviral vectors
and protocols used to measure rates of recombination have been
described previously (22). To study recombination between
chimeric and infectious MLV RNAs, a chimeric RNA vector DNA (pJZ446 or
pJZ211) was transfected into the SNV C3A2 helper cell line (containing
the SNV gag-pol and env genes) (18)
(Fig. 4). The cells were selected for Hygr, and the
resistant cells were pooled and designated STEP 1 cells (Fig. 4, STEP
1). Virus from step 1 cells was used to infect the MLV helper cell line
PG13. Infected cells were selected for Hygr, and individual
clones were isolated and designated STEP 2 cells (Fig. 4, STEP 2). The
structures of the proviruses formed from the SNV U3-minus vector in the
PG13 cells were monitored by Southern (DNA) analysis. The
XhoI linker in JZ446 is duplicated in the 5' LTR during
formation of the STEP 2 provirus (5). The STEP 2 clones
containing the expected XhoI fragment that hybridized to a
hyg probe were used for further analysis (Fig. 4, STEP 2). To test whether any virus capable of forming Hygr colonies
was produced by the STEP 2 cells, the supernatant medium (4 ml) from
each STEP 2 cell clone was used to infect D17 cells, and the infected
cells were selected for Hygr. No Hygr colonies
were detected. This was because the deletion of the U3 region (promoter
and enhancer) in the SNV 5' LTR prevented transcription from the SNV
vector (4). The infectious MLV vectors LN, LN290, L290N,
L290N290, and LN80 (Fig. 4) were each transfected into the amphotropic
helper cell line PA317 (11) (Fig. 4). Viruses from the
transfected PA317 cells were used to superinfect the STEP 2 cells
containing JZ446 or JZ211, and the infected cells were selected for
neomycin resistance (Neor). Individual Neor
clones were isolated and designated STEP 3 cells (Fig. 4, STEP 3). Each
STEP 3 cell clone contained a single JZ446 or JZ211 integration and a
single integration of LN, LN80, LN290, L290N, or L290N290.
Introduction of a single JZ525 and a single YM2 into helper cell
line PG13.
Plasmid DNA of pJZ525 (Fig. 5A) was transfected into
the MLV amphotropic helper cell line PA317 (11). Virus
from transfected PA317 cells was used to infect the MLV xenotropic
helper cell line PG13 (12). Infected cells were selected
for Hygr, and individual Hygr green clones were
isolated and designated STEP 2 cells. Plasmid DNA of pYM2 was used to
transfect fresh PA317 cells. Transfected PA317 cells were selected for
Neor. The green Neor cells were sorted by flow
cytometry. Viruses released from green cells were used to infect STEP 2 cells, which contained JZ525 proviruses. Infected STEP 2 cells were
selected for Neor, and individual Neor cells
were cloned. Due to the high rate of recombination between the two
IRESs within YM2, the gfp gene between the two IRESs was deleted in a large portion of Neor cells. To distinguish
between the Neor PG13 cells containing parental and
recombinant YM2, the cellular DNAs of each Neor colony were
digested with EcoRV and hybridized with a hyg
probe. Neor cells integrated with a parental YM2 generated
a 5.5-kb fragment, while cells with a recombinant YM2 generated a
4.1-kb fragment. The Neor cells with a parental YM2 were
designated STEP 3 cells and used for further study.
Cells, transfection, and infection.
The handling of D17
cells (a dog osteosarcoma cell line; ATCC CRL-8468), PA317 cells (ATCC
CRL-9078), and PG13 helper cells (ATCC CRL-10686), DNA transfections,
virus harvesting, and virus infections were as previously described
(22).
Fusion of D17 cells and PG13 cells.
The cells (5 × 105 cells of each cell line) were fused with polyethylene
glycol 1500 (PEG). The cells were mixed in 1 ml of Dulbecco modified
Eagle medium (DMEM) with 10% of PEG and 5% dimethyl sulfoxide for 3 min. PEG was diluted by adding 5 ml (DMEM), and cells were plated in a
60-mm dish and incubated at 37°C. The PEG medium was replaced with
fresh DMEM containing 10% fetal calf serum 1 h after fusion.
Viruses were collected 36 h after fusion.
Sequence analyses of junctions of recombinants between JZ446 and
LN80.
Cellular DNAs of STEP 4 D17 cells were isolated and
amplified by PCR using primer 5'-CTACTTCGAGCGGAGGCATCC-3',
which hybridized within the 5' end of hyg, and
5'-ATGCCTTGCAAAATGG-3', which hybridized to the U3 region of
the 3' LTR (Fig. 6C and D). Primers
5'-GCGCGGCCGTCTGGAC-3', 5'-CTGGAGTTCGTGACCG-3', and/or
5'-CTTAAGCTAGCTTGCC-3' were used for sequencing the
amplified fragments.

View larger version (17K):
[in this window]
[in a new window]
|
FIG. 6.
An intramolecular recombination follows an
intermolecular recombination. (A) The RNA chimeric virus JZ446 contains
only one LTR and hyg. Hygr colonies form only
when recombination between JZ446 and an infectious vector occurs so
that hyg is flanked by two LTRs. (B) Infectious MLV vector
LN80 containing an 80-bp sequence identical to the TK sequence in
JZ446. The 80-bp sequence was inserted into the 3' portion of the
neo sequence (downstream of the 3' neo gene or
after the neo stop codon). Most of the recombinants utilized
the 80-bp sequences identical between JZ446 and LN80. (C) Recombinants
between JZ446 and LN80. The 80-bp TK sequence is located at the 3' end
downstream of the second 3' hyg sequence. Recombinants
resulted from intermolecular recombination between the identical 80-bp
sequences within JZ446 and LN80 and contained two 290-bp 3'
hyg sequences with gfp between within the same
RNA molecule. The arrowheads represent locations of the PCR primers.
(D) Recombinants resulting from an intramolecular recombination
following an intermolecular recombination. Recombination between the
two identical 290-bp sequences within JZ446 formed a clear
Hygr colony by deletion of gfp.
|
|
Fluorescence microscopy.
A fluorescence inverted microscope
(Zeiss Axiovert 25) with a 100-W mercury arc lamp and a fluorescent
filter set (CZ909) consisting of a 470/40-nm exciter, a 515-nm emitter,
and a 500-nm beam splitter were used to detect GFP in living cells.
 |
RESULTS |
Direct evidence for intramolecular recombination.
Previous
study demonstrated that the rate of recombinations between two
identical sequences within the same RNA molecules was about 62% during
a single replication cycle (19) (Fig. 1A). However, the
high rate of recombination could be either intra- or intermolecular,
because it could result from an intramolecular event or from the
recombination of a downstream sequence of one RNA molecule with the
upstream sequence of the other RNA molecule.
To demonstrate an intramolecular event, chimeric RNA vector JZ446 (Fig.
1BC) was introduced into the MLV helper cell line PG13 as described in
the Materials and Methods JZ446, derived from SNV, contained a deletion
in the U3 region of the 3' SNV LTR. (Fig. 4, top). This vector also
contained a truncated MLV vector between the two SNV LTRs, in a
transcriptional orientation opposite that of the SNV LTRs. In this
truncated MLV vector, hyg was expressed from the 5' MLV LTR,
and an HSV TK poly (A) addition signal replaced the deleted 3' MLV LTR.
This truncated MLV vector contained hyg and gfp,
separated by an IRES sequence. hyg was expressed from the 5'
MLV LTR, while gfp was expressed from the IRES sequence. The
IRES was isolated from an encephalomyocarditis virus origin and allows
the ribosome to bind to the internal AUG that initiates translation of
the second gene independently from the upstream gene (1,
2). Cells that express the gfp gene are green under a
fluorescence microscope, while cells without the gfp gene
product are clear (16, 19). In addition, the MLV sequences
of JZ446 also included the insertion of a second sequence homologous to
290 bp of the 3' end of hyg inserted into the 3'
untranslated portion of gfp (Fig. 1B and C). As expected, after one round of replication, the JZ446 recombinants showed that the
downstream 3' hyg sequence had recombined with the upstream hyg sequence, resulting in a deletion of gfp
(Fig. 6D). Therefore, these recombinants are clear under a fluorescence microscope.
Next, the chimeric RNA vector JZ446 and infectious vectors LN and LN80
were separately introduced into the helper cell line PG13
(12) as described in Materials and Methods. Individual clones were designated STEP 3 cells (Fig. 4, STEP 3). Viruses from each
STEP 3 clone were used to infect D17 cells, and the infected cells were
selected separately for Hygr and for Neor. The
resulting cells were designated STEP 4 cells (Fig. 4, STEP 4). With
this approach, Hygr colonies form only when a recombination
between the JZ446 and LN or LN80 vectors has occurred so that
hyg is flanked by two LTRs. The target cells do not contain
viral gag-pol and env gene products for
retrovirus replication. Therefore, progeny virus cannot be released
from them (22). Consequently, these vector viruses have
undergone only one cycle of replication.
Since the STEP 4 resulted from selection for Hygr, the LN
3' LTR had to recombine with the sequence downstream of hyg
within JZ446. This recombination was nonhomologous because the 3' end of JZ446 did not contain any sequence identical to LN. Therefore, formation of a Hygr colony resulted from a nonhomologous
recombination between JZ446 and LN. The ratio of the Hygr
titer to the Neor titer represented the rate of
recombination between the JZ446 and LN, which was 7.8 × 10
5 ± 2.7 × 10
5 (Table
1). This result was consistent with
previous report that the rates of nonhomologous recombinations were
very low (21).
View this table:
[in this window]
[in a new window]
|
TABLE 1.
Assay of D17 cells infected with STEP 3 clones of PG13
cells containing JZ446 and an infectious LN- or LN80-based vector
|
|
The second infectious MLV vector, LN80 (Fig. 1C), was identical to LN
except that an 80-bp HSV TK sequence before the poly(A) addition site
(or AATAAA) was inserted into the 3' untranslated portion of the
neo sequence. This 80-bp TK sequence is identical to the TK
sequences within JZ446 (Fig. 1C). Since there was an 80-bp sequence
homology between JZ446 and LN80, most recombinations between the two
vectors utilized the 80-bp sequences. The rate of recombination for
vectors containing shared homologous sequences is higher than the rate
of nonhomologous recombination (22, 23). The rate of
recombination (48.6 × 10
5 ± 39.9 × 10
5) between JZ446 and LN80 was approximately six times
that between JZ446 and LN (7.8 × 10
5 ± 2.7 × 10
5) (Table 1); i.e., only 16% (7.8/48.6) of
recombinations between JZ446 and LN80 resulted from nonhomologous recombination.
Two crossovers may occur to form clear Hygr LN80 colonies
(Fig. 6). First, most recombinations between JZ446 and LN80 resulted from recombination between the two identical 80-bp TK sequences such
that hyg was flanked by two LTRs (Fig. 6A and B).
Recombinants resulting from the first intermolecular recombination
contained two 290-bp 3' hyg sequences flanking
gfp within the same RNA molecule (Fig. 6C). The second event
would be an intramolecular crossover between the two identical 290-bp
sequences within JZ446 to delete gfp and form a clear
Hygr colony (Fig. 6D). If this intramolecular recombination
did not follow the intermolecular recombination, a green
Hygr colony would be formed (Fig. 6C).
Recombination between JZ446 and LN80 occurred at the shared 80-bp
homologous TK sequence, as well as at nonhomologous sequences. As
mentioned above, there was approximately 84% (100%
16%)
homologous recombinations between JZ446 and LN80 which utilized the
80-bp sequences identical between the two vectors. Only 14% of the
Hygr colonies for JZ446 and LN80 examined by fluorescence
microscopy were green, while 86% of colonies were clear (Table 1).
Most green colonies were assumed to represent recombinants between LN80
and JZ446 within the 80-bp sequences without further recombinations (Fig. 1C). Clear colonies (86%) represented the sum of the numbers of
nonhomologous recombination between the LN80 and upstream sequences of
the gfp gene of JZ446 and the numbers of homologous
recombination between the two 80-bp sequences within the two vectors
plus a second intramolecular recombination between the two 290-bp
sequences within JZ446 (Fig. 6D), which was more than 70% (86%
16%). In other words, 70% of the total 84% recombinants utilizing
the 80-bp TK homologous sequences had further undergone an
intramolecular recombination (Fig. 6C and D). Therefore, in this case,
the rate of intramolecular recombination was approximately 83%
(70%/84%).
Since most green recombinants resulted from a recombination between the
two shared 80-bp sequences within JZ446 and LN80, the RNAs transcribed
from those green recombinant proviruses with two identical 290-bp
sequences should undergo recombination if the RNAs were reverse
transcribed. To determine the phenotypic nature of green recombinants
between JZ446 and LN80, the green STEP 4 D17 cells (Fig. 6C) and fresh
PG13 cells were fused as described in Materials and Methods. The MLV
proteins (Gag-Pol and Env) from PG13 helper cells packaged the viral
RNA transcribed from the green provirus (Fig. 6C) in the STEP 4 D17
cells. Viruses released from the fused cells were used to infect fresh
D17 cells. The infected D17 cells were selected for Hygr.
Five of six STEP 4 clones analyzed produced viruses after fusing with
PG13 cells. The titers of those viruses ranged from 300 to 3,000 CFU/ml. The Hygr colonies were examined under a
fluorescence microscope. The ratio of the number of the clear colonies
to the number of total colonies represented the rate of recombination
between the two 290-bp sequences. The rate was 59% ± 5%, which was
similar to the rate for JZ442 + 3' Hyg (62% ± 9%) (Fig. 1A)
(19).
To determine the genomic structure of recombinants of JZ446 and LN80,
proviral sequences of the STEP 4 cells were analyzed. Cellular DNAs
were isolated from individual Hygr STEP 4 green and clear
clones and were amplified by PCR using two primers hybridized within
the MLV proviruses as described in Materials and Methods. The amplified
proviral DNAs were sequenced. All four green colonies analyzed
coincided with the structure predicted in Fig. 6C, which encoded, from
5' to 3', the gfp sequence and the 290-bp sequence from
JZ446, the 80-bp shared TK sequence, and the U3 region from LN80. All
10 clear colonies analyzed were the structure predicted in Fig. 6D,
which encoded the hyg sequence from JZ446, the shared 80-bp
sequence, and the U3 region from LN80. The high rate of recombination
must occur during reverse transcription, because the deletion rate
after reverse transcription is very low (less than 10
5)
(9). Therefore, intramolecular recombination actually
occurred between the two 290-bp sequences within JZ446.
High rates of recombination between two identical sequences within
the same RNA molecule resulted mostly from intramolecular
recombination.
We have provided evidence that intramolecular
recombination actually occurred within the chimeric RNA vector JZ446.
The high rate of recombination, however, might result from a negative
interference since the intramolecular recombinations within JZ446
followed an intermolecular recombination between the two vectors. The
above observation demonstrated that intramolecular recombinations
occurred, but we could not determine whether the high rate of
recombination of two identical sequences (Fig. 1A) (19)
resulted mostly from intramolecular recombination.
Previous work indicated that the recombination (both inter- and
intramolecular) rate between two identical sequences within the same
retroviral RNA molecule is 62% per replication cycle (Fig. 1A)
(19). The intermolecular recombination rate between a 1-kb
homologous sequence is only 4% (6). If the high rate (62%) of recombination between the two identical sequences resulted from intermolecular recombinations, it would be expected that the
presence of four copies of identical 290-bp sequences within the two
RNA molecules would result in a significantly higher rate of
intermolecular recombination. To separate intermolecular from intramolecular recombination, chimeric RNA vectors JZ211 and JZ446 (Fig. 2 and 3) were introduced into the MLV helper cell line PG13 as
described above. JZ211 has been described previously (22). Briefly, JZ211 was identical to JZ446 (Fig. 2 and top of Fig. 4) except
that it carried only hyg and thus contained only one copy of
the 290-bp 3' hyg sequence. As described above, JZ446 contained two copies of the 290-bp sequences. Infectious vectors were
used to recombine with JZ211 and JZ446. In addition to LN described
above, three infectious vectors were constructed. LN290 (Fig. 2B and
3B) and L290N (Fig. 2C and 3C) contained a neo gene and one
copy of the 290-bp sequence from the 3' end of the hyg gene,
except that the 290-bp sequence of LN290 was inserted into the 3' end
of neo and served as the 3' untranslated sequence, whereas
the 290-bp sequence of L290 was inserted into the 5' end of
neo and served as the 5' untranslated sequence. L290N290
contained two copies of the 290-bp sequence (Fig. 2D and 3D).
Next, the chimeric RNA vectors JZ211 (and JZ446) and infectious vectors
LN, LN290, L290N, and L290N290 were separately introduced into the
helper cell line PG13 (12) as described in Materials and
Methods. Individual clones were designated STEP 3 cells (Fig. 4, STEP
3). Viruses from each STEP 3 clone were used to infect D17 cells, and
the infected cells were selected separately for Hygr and
for Neor. The resulting cells were designated STEP 4 cells
(Fig. 4, STEP 4). With this approach, Hygr colonies form
only when a recombination between JZ211 (or JZ446) and any one of
LN290, L290N, and L290N290 vectors has occurred at the shared 290-bp
homologous sequences so that the hyg gene is flanked by two LTRs.
If a nonhomologous recombination event occurs between JZ211 (or JZ446)
and LN, it results in a Hygr colony. The rate of
recombination is much lower than the rates for vectors containing the
shared 290-bp homologous sequences (Fig. 2A and 3A; Table
2) (22, 23); therefore, most
recombinations were between the 290-bp sequences. The ratios of
Hygr CFU to Neor CFU produced were 170 × 10
5 ± 50 × 10
5 and 80 × 10
5 ± 60 × 10
5 for LN290 and L290N,
respectively, as described previously (19) (Table 2; Fig.
2B and C). Those ratios represented the rate of recombination between
one copy of the 290-bp within the chimeric RNA vector (JZ211) and one
copy of the 290-bp sequence within the infectious vector (L290N or
LN290). L290N290 contained two copies of the 290-bp sequences identical
to that in JZ446. Specifically, the 290-bp 3' hyg sequence
was inserted both upstream and downstream of neo within this
vector (Fig. 2D and 3D). The actual titers of the infectious viruses
(L290N290) as measured by Neor selection should be higher
than that. This difference could be due to a recombination between the
two identical sequences on either side of the neo gene of
L290N290 that would result in a deletion of neo. To estimate
the rate of deletion within L290N290 during a single round of
replication, STEP 3 viruses of L290N290 were used to infect D17 cells.
Infected cells were pooled without selection, and the DNAs of pooled
D17 cells were digested with EcoRV and hybridized with a
probe containing the MLV packaging signal. EcoRV digested
within the LTRs of the vectors such that LN290 produced a 2.9-kb
fragment and L290N produced a 2.7-kb fragment (Fig.
7A). The parental provirus of L290N290
produced a 3.1-kb fragment, while the recombinant provirus with the
neo deletion produced a 1.5-kb fragment (Fig. 7A). The ratio
of the intensity of the 3.1-kb fragment and the 1.5-kb fragment was
about 5. After normalization of the L290N290 titer, the JZ446/L290N290
recombination rate was 430 × 10
5 (Fig. 3D) (Table
2). Intermolecular recombination between JZ446 and L290N290 (i.e., two
RNA molecules that together have four copies of the identical 290-kb
sequences) increased only three to six times above that of
intermolecular recombinations between JZ211 and LN290 or L290N (i.e.,
two RNA molecules that together have two copies of the 290-kb identical
sequences). The recombination rate between two identical sequences
within the same RNA molecules (i.e., two RNA molecules containing four
copies of the 290-kb identical sequences) was 62%, (19),
while the rate of the intermolecular recombination between two 290-bp
homologous sequences (i.e., two RNA molecules containing two copies of
the 290-kb identical sequences) should be 1.16% (290/1,000 × 4%) (6). The recombination rate between the two identical
sequences within the same RNA molecule would increase 50-fold.
Therefore, this observation suggested that the high rate of
recombinations between two identical sequences within the same RNA
molecule resulted mostly from intramolecular recombination.
View this table:
[in this window]
[in a new window]
|
TABLE 2.
Assay of D17 cells infected with STEP 3 clones of PG13
cells containing JZ446 and an infectious LN290- or 290LN-based
vector
|
|

View larger version (38K):
[in this window]
[in a new window]
|
FIG. 7.
Southern analyses of recombinants between JZ446 and
LN290, L290N, and L290N290. (A) Southern analysis of chromosomal DNA of
unselected STEP 4 cells. Cellular DNAs were isolated from
JZ446/L290N290 STEP 4 cells, digested with EcoRV, and
hybridized with a package signal probe. (B) Southern analysis of
chromosomal DNA of Hygr STEP 4 cells. Cellular DNAs were
isolated from STEP 4 cells (Fig. 4) that had been pooled from
Hygr colonies; the DNA was then digested with
EcoRV and hybridized with a hyg probe (Fig. 3).
The cells examined are listed above the lanes. Molecular sizes
(kilobase pairs) are shown on the right. The additional fragment (2.9 kb) in JZ446/LN290 proviruses was an EcoRV fragment of LN290
which contains a 290-bp 3' hyg sequence. The additional
2.7-kb EcoRV fragment of L290N and 3.1-kb EcoRV
fragment of L290N290 also resulted from the infectious viruses.
|
|
To determine whether recombinations had occurred at the 5' or the 3'
end of the 290-bp hyg sequence, the DNA from STEP 4 cells was digested with EcoRV and hybridized with a hyg
probe. The proviruses resulting from recombination between JZ446 and
L290N290 using the upstream and downstream hyg sequence gave
four different-sized EcoRV fragments that hybridized with a
hyg probe (Fig. 7B). The clear recombinants (92% of total
Hygr colonies) utilizing the upstream hyg
sequence of L290N290 gave a 4.0-kb EcoRV fragment, and those
utilizing the downstream hyg sequence gave a 2.6-kb
EcoRV fragment. The green recombinants (8% of total
Hygr colonies) utilizing the upstream hyg
sequence of L290N290 gave a 5.7-kb EcoRV fragment, while
those utilizing the downstream hyg sequence gave a 4.2-kb
EcoRV hyg fragment. The results shown in Fig. 7B
indicate that both identical sequences were used to form recombinants
between the chimeric RNA and the infectious vector.
Determination of the rate of intramolecular recombination.
Evidence for intramolecular recombination provided above was based on
assays using two heterologous RNA molecules. The packaging and
arrangement of the two different RNA molecules within a viral particle
may be very different from those of two homologous RNA molecules. To
determine the rate of intramolecular recombination, two homologous
MLV-based vectors were constructed. The first vector, JZ525, carries,
from 5' to 3', hyg, an IRES sequence, gfp, an additional copy of the IRES sequence, and neo (Fig. 5A). In
this vector, hyg was expressed from the 5'-end LTR, and
gfp and neo were expressed from two IRESs. The
gfp gene was flanked by the two identical IRES sequences.
After one round of replication, recombination between the two IRES
sequences would delete the gfp gene (Fig. 5D), so that cells
encoding the parental JZ525 were green under a fluorescence microscope
whereas cells with the recombinant JZ525 were clear (Fig. 5D). The
neo gene encoded a frameshift mutation at its
NcoI site so that the neo gene was nonfunctional
within this vector. Cells infected with JZ525 were Hygr and
Neos. The second vector, YM2, was identical to JZ525 except
that neo was functional but hyg was nonfunctional
because of a frameshift mutation on the SgtI site within the
hyg open reading frame (Fig. 5B). As a result, cells
containing YM2 were Hygs and Neor. JZ525 and
YM2 were introduced into the PG13 cell line as described in Materials
and Methods. Approximately half of the viruses released from these
cells would contain a JZ525 RNA molecule and a YM2 RNA molecule.
(Retroviruses package two RNA molecules in each virion.) Retroviral
recombinations occur during minus-strand DNA synthesis, which begins
from the 3' end of the viral RNA molecules (20). An
intermolecular recombination occurring between the downstream IRES
within YM2 and the upstream IRES within JZ525 would not only delete
gfp but also form a double-resistant (Hygr
Neor) provirus. Only half of the intermolecular
recombinations would result in double resistance, while the other half
would result in double-sensitive proviruses which could not be detected
after selection. Therefore, the intermolecular recombination rate
should be twice the ratio of the number of the double-resistant
colonies to the number of the single-resistant colonies. Most
intramolecular recombination between the two identical IRES sequences
would form a single-resistant provirus without gfp.
Viruses released from the helper cell line PG13 containing JZ525 and
YM2 were used to infect D17 cells. Infected cells were selected for
Hygr, Neor, and Hygr
Neor. The rate of recombinations between the two vectors
should be twofold above the ratio of the number of the double-resistant colonies to the sum of two single-resistant colonies. The recombination rate between the two frameshift mutations was 1.3% ± 0.6% (Table 3).
The rate of recombination between the two IRES sequence resulting in
gfp deletion would be the ratio of the number of clear colonies to the number of total single-resistant colonies, which was 46 to 51% or 48.5% on average (Table 3). The rate of intermolecular recombination between a downstream IRES within one vector and the
upstream IRES within the other vector was the product of the rate of
intermolecular recombination (1.3%) and the deletion rate of the
double-resistant colonies (36% [Table 3]), which was only 0.5%
(1.3% × 36%). Therefore, approximately 99% [(48.5%
0.5%)/48.5%] of recombinations between the two identical sequences
resulted from an intramolecular recombination.
To determine whether recombinations occurred between the two identical
IRES sequences within JZ525 and YM2, DNA from STEP 4 cells was digested
with EcoRV and hybridized with a hyg probe. The
green proviruses produced a 5.5-kb fragment, and the clear proviruses
produced a 4.1-kb fragment. The results shown in Fig. 8 indicate that the clear clones resulted
from recombinations between the two identical IRES sequences within
JZ446 and YM2.

View larger version (51K):
[in this window]
[in a new window]
|
FIG. 8.
Southern analysis of recombinants between JZ425 and YM2.
Cellular DNAs from green or clear Hygr, Neor,
and Hygr Neor STEP 4 clones were digested with
EcoRV and hybridized with a hyg probe. The clones
examined are listed above the lanes. Molecular sizes (kilobase pairs)
are shown on the right.
|
|
 |
DISCUSSION |
Retroviruses, as a consequence of having two RNA molecules in
their virions, recombine at a high rate. The rate of recombination between two identical sequences within the same RNA molecule is about
62% (19). It was deduced that recombination may occur between two sequences on one RNA molecule (intramolecular) as well as
between the two RNA molecules (intermolecular) (17). We
demonstrated intramolecular recombination occurred within the chimeric
RNA vector JZ446. Using two homologous vectors, we determined that
approximate 99% recombinations between the two identical sequences
within the same RNA molecules resulted from an intramolecular recombination.
The rate of intramolecular recombination (83%) within JZ446 (Fig. 6)
is higher than the rate of deletion (62%) between the same 290-bp
sequence identities within each of two essentially homologous RNA
molecules (JZ442 + 3' Hyg) (Fig. 1A). The deletion within the two
homologous RNA molecules requires only one step of recombination:
either an intramolecular or an intermolecular recombination between the
downstream sequence from one RNA molecule and the upstream sequence of
the same or the other RNA molecule. However, the deletion within the
same two 290-bp sequences within the chimeric RNA molecule (JZ446)
requires two steps. The first step is intermolecular, allowing
hyg to be located between two LTRs; the second step is
intramolecular, deleting gfp. The discrepancy may have
resulted from negative genetic interference.
We hypothesized that the rate of intramolecular would be higher if it
followed an intermolecular recombination. However, the rate of deletion
within JZ525/YM2 with an intermolecular recombination was a little bit
lower than that without an intermolecular recombination (Table 3).
Retroviral recombination occurs during minus-strand DNA synthesis
(20), and so a minus-strand DNA from downstream sequences
lands on an upstream sequence of the RNA template. The difference
between JZ446/LN and JZ525/YM2 is that in the former system, the two
identical sequences were available within JZ446 after an intermolecular
recombination (Fig. 6C). However, in the latter system, as soon as
reverse transcription passed the downstream IRES, the intramolecular
recombination between the two identical sequences (IRESs) could not
occur. In this system, intermolecular recombinants between the two
frameshift mutations (SgfI and NcoI) were
selected (Fig. 5A and B). In JZ525/YM2, the hyg sequence downstream of the SgfI site, the upstream IRES, and the
gfp gene were 2 kb in length, while the downstream IRES and
the NcoI site upstream of the neo sequence were
only 1 kb in length. The difference in the rates of deletion resulting
from a single selection and the double selection was probably due to
intermolecular recombinations within the 2-kb sequence, which were not
followed by an intramolecular recombination.
The high rates of deletion between two identical sequences within a
retroviral RNA molecule are mostly due to interamolecular recombinations. This observation is consistent with the fact that the
newly synthesized minus-strand DNA primers preferably anneal on the
intra-RNA molecule rather than on the inter-RNA molecule (7). This suggests that the two RNA molecules are not
randomly intertwined in a virion. Alternatively, the reverse
transcriptase, with partially synthesized minus-strand DNA, may
slide along with the template RNA molecule until the minus-strand DNA
anneals at a homologous sequence within the same RNA molecule. Once DNA
synthesis is completed, the resulting DNA would contain a deletion
between the two identical sequences.
We thank William Bargmann, Robert Jacob, Ting Li, and Alan
Simmons for helpful comments on the manuscript.
| 1.
|
Adam, M. A.,
N. Ramesh,
A. D. Miller, and W. R. Osborne.
1991.
Internal initiation of translation in retroviral vectors carrying picornavirus 5' nontranslated regions.
J. Virol.
65:4985-4990[Abstract/Free Full Text].
|
| 2.
|
Boris-Lawrie, K. A., and H. M. Temin.
1993.
Recent advances in retrovirus vector technology.
Curr. Opin. Genet. Dev.
3:102-109[CrossRef][Medline].
|
| 3.
|
Coffin, J. M.,
S. H. Hughes, and H. Varmus.
1997.
Retroviruses.
Cold Spring Harbor Laboratory Press, Plainview, N.Y.
|
| 4.
|
Dornburg, R., and H. M. Temin.
1988.
Retroviral vector system for the study of cDNA gene formation.
Mol. Cell. Biol.
8:2328-2334[Abstract/Free Full Text].
|
| 5.
|
Dougherty, J. P., and H. M. Temin.
1987.
A promoterless retroviral vector indicates that there are sequences in U3 required for 3' RNA processing.
Proc. Natl. Acad. Sci. USA
84:1197-1201[Abstract/Free Full Text].
|
| 6.
|
Hu, W. S., and H. M. Temin.
1990.
Genetic consequences of packaging two RNA genomes in one retroviral particle: pseudodiploidy and high rate of genetic recombination.
Proc. Natl. Acad. Sci. USA
87:1556-1560[Abstract/Free Full Text].
|
| 7.
|
Jones, J. S.,
R. W. Allan, and H. M. Temin.
1994.
One retroviral RNA is sufficient for synthesis of viral DNA.
J. Virol.
68:207-216[Abstract/Free Full Text].
|
| 8.
|
Klaver, B., and B. Berkhout.
1994.
Premature strand transfer by the HIV-1 reverse transcriptase during strong-stop DNA synthesis.
Nucleic Acids Res.
22:137-144[Abstract/Free Full Text].
|
| 9.
|
Li, T., and J. Zhang.
2000.
Determination of the frequency of retroviral recombination between two identical sequences within a provirus.
J. Virol.
74:7646-7650[Abstract/Free Full Text].
|
| 10.
|
Lobel, L. I., and S. P. Goff.
1985.
Reverse transcription of retroviral genomes: mutations in the terminal repeat sequences.
J. Virol.
53:447-455[Abstract/Free Full Text].
|
| 11.
|
Miller, A. D., and C. Buttimore.
1986.
Redesign of retrovirus packaging cell lines to avoid recombination leading to helper virus production.
Mol. Cell. Biol.
6:2895-2902[Abstract/Free Full Text].
|
| 12.
|
Miller, A. D.,
J. V. Garcia,
N. von Suhr,
C. M. Lynch,
C. Wilson, and M. V. Eiden.
1991.
Construction and properties of retrovirus packaging cells based on gibbon ape leukemia virus.
J. Virol.
65:2220-2224[Abstract/Free Full Text].
|
| 13.
|
Miller, A. D., and G. J. Rosman.
1989.
Improved retroviral vectors for gene transfer and expression.
BioTechniques
7:980-982[Medline], 984-986, 989-990.
|
| 14.
|
Ramsey, C. A., and A. T. Panganiban.
1993.
Replication of the retroviral terminal repeat sequence during in vivo reverse transcription.
J. Virol.
67:4114-4121[Abstract/Free Full Text].
|
| 15.
|
Sambrook, J.,
E. F. Fritsch, and T. Maniatis.
1989.
Molecular cloning: a laboratory manual, 2nd ed.
Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y.
|
| 16.
|
Sapp, C. M.,
T. Li, and J. Zhang.
1999.
Systematic comparison of a color reporter gene and drug resistance genes for the determination of retroviral titers.
J. Biomed. Sci.
6:342-348[CrossRef][Medline].
|
| 17.
|
Skalka, A. M., and S. Goff.
1993.
Reverse transcriptase.
Cold Spring Harbor Laboratory Press, Plainview, N.Y.
|
| 18.
|
Watanabe, S., and H. M. Temin.
1983.
Construction of a helper cell line for avian reticuloendotheliosis virus cloning vectors.
Mol. Cell. Biol.
3:2241-2249[Abstract/Free Full Text].
|
| 19.
|
Zhang, J., and C. M. Sapp.
1999.
Recombination between two identical sequences within the same retroviral RNA molecule.
J. Virol.
73:5912-5917[Abstract/Free Full Text].
|
| 20.
|
Zhang, J.,
L. Y. Tang,
T. Li,
Y. Ma, and C. M. Sapp.
2000.
Most retroviral recombinations occur during minus-strand DNA synthesis.
J. Virol.
74:2313-2322[Abstract/Free Full Text].
|
| 21.
|
Zhang, J., and H. M. Temin.
1993.
3' junctions of oncogene-virus sequences and the mechanisms for formation of highly oncogenic retroviruses.
J. Virol.
67:1747-1751[Free Full Text].
|
| 22.
|
Zhang, J., and H. M. Temin.
1993.
Rate and mechanism of nonhomologous recombination during a single cycle of retroviral replication.
Science
259:234-238[Abstract/Free Full Text].
|
| 23.
|
Zhang, J., and H. M. Temin.
1994.
Retrovirus recombination depends on the length of sequence identity and is not error prone.
J. Virol.
68:2409-2414[Abstract/Free Full Text].
|