1
Determination of sequences required for HERV-K transduction and its recognition by
2
foreign retroviral virions.
3
4
Otto Erlwein+, Nathan P. Sweeney+, Raffaele de Leon, Gillian Wills, Mark J. Robinson and
5
Myra O. McClure*
6
7
Jefferiss Research Trust Laboratories, Section of Infectious Diseases, Wright-Fleming
8
Institute, Faculty of Medicine, Imperial College London, St. Mary’s Campus, Norfolk
9
Place, London W2 1PG, United Kingdom
10
+Joint
11
*Corresponding author
first authors
12
13
Email addresses:
14
Otto Erlwein: o.erlwein@imperial.ac.uk
15
Nathan P. Sweeney: nathan.sweeney10@imperial.ac.uk
16
Raffaele de Leon: raffdeleon@gmail.com
17
Gillian Wills: g.wills@imperial.ac.uk
18
Mark J. Robinson: mark1of2@gmail.com
19
Myra O. McClure: mmcclure@imperial.ac.uk
20
1
21
ABSTRACT
22
23
Sequences necessary for transduction of HERV-Kcon, a consensus of the HERV-K(HML-2)
24
family were analysed and found to reside in the leader/gag region. They act in an
25
orientation-dependent way and consist of at least two sites working together. Having
26
defined these sequences, we exploited this information to produce a simple system to
27
investigate to what extent virions of HERV-Kcon, MLV and HIV-1 have the ability to
28
transduce each other’s genomes, leading to potential contamination of gene therapy
29
vectors. [77 words]
30
31
32
TEXT
33
34
Human endogenous retroviruses (HERVs) are groups of retroviral elements that infected
35
germ cells in the past and are now part of the human genome. In several instances HERV
36
expression has been linked to diseases and viral particles detected in tissues (reviewed
37
in 1). Of the different HERV families, HERV-K(HML-2) is believed to be the most recently
38
acquired (1). While many HERVs have acquired deletions and mutations rendering them
39
defective, HERV-K113 and HERV-K115 have retained functional open reading frames
40
(ORFs) enabling them to encode retroviral-like particles. HIV-1 infection can activate
41
HERV expression (reviewed in 2), but little is known about the packaging sequences of
42
HERVs and their relationship to other retroviruses. This oversight may be important as
43
cross-packaging of other retroviral genomes into HERV-K virions upon infection has the
44
potential to broaden their tropism. We mapped the sequences necessary for the
45
transduction of HERV-Kcon as the prototypic member of the HERV-K(HML-2) family and
46
looked for functional interactions with other retroviruses.
47
48
The generation of HERV-Kcon from the consensus of several HERV-K sequences and its
49
ability to encode functional open reading frames (ORFs) has been published (3). In our
2
50
experiments we used the described system of VSV-G pseudotyped HERV-Kcon particles,
51
able to transduce CHKCG, a full-length HERV-Kcon vector carrying a CMV/eGFP cassette
52
in its env ORF (Fig.1A). The HERV-Kcon proteins Gag, PR and Pol were made from
53
expression plasmid pCVI/Gag-PR-Pol. We also included expression plasmid pCR3.1/K-
54
Rev in the transfection cocktail as it encodes K-Rev (or Rec) a protein involved in HERV-
55
Kcon RNA export from the nucleus to the cytoplasm similar to HIV-1 Rev protein (3). VSV-
56
G glycoprotein was encoded by plasmid pMD2.G. All HERV-Kcon constructs were
57
generated using standard cloning techniques and confirmed by sequencing. The VSV-G
58
pseudotyped HERV-Kcon virions were prepared from transfected 293T cells, using
59
polyethylenimine (4). Experiments were performed in triplicate and flow cytometry
60
determined transfection efficiencies to be above 85% for all constructs. In all
61
experiments, GFP-expressing colonies were scored by fluorescence microscopy on
62
human melanoma A375 cells transduced by supernatant harvested two days post-
63
transfection of 293T cells with the vectors described.
64
65
We introduced deletions in CHKCG and investigated whether they could transduce
66
recipient cells (Fig.1A). We also generated mutant pSTOP carrying a stop codon in gag
67
after amino acid 228. In contrast to CHKCG but like the deletion mutants, pSTOP needed
68
plasmid pCVI/Gag-PR-Pol to provide HERV-Kcon Gag, PR and Pol for transduction (data
69
not shown), allowing unbiased comparison of titres.
70
71
Compared with plasmid pSTOP, constructs p1280/1799 and p1083/1799 (numbers
72
referring to positions of the proviral HERV-Kcon DNA) carrying deletions within the
73
gag/leader region showed a marked decrease in their transduction efficiency (Fig. 1B).
74
In contrast, deletion mutant 1800/6639 transduced better than the full-length mutant
75
pSTOP. These results indicate that important sequences for transduction are located
76
before position 1800 of HERV-Kcon. This is in addition to the sequence between 6852 and
77
9472 that contains the previously identified Rec-responsive element (5) which was
78
retained in all constructs.
3
79
80
Foamy viruses (FV) are a distinct subfamily of exogenous retroviruses (6). They are not
81
known to express a Rev-like protein. A FV packaging system based on the FV structural
82
proteins Gag, Pol and Env together with the vector, pΔΦ, has been described (7). We
83
generated vector pFV-eGFP by PCR amplifying the SFFV/eGFP cassette (eGFP under the
84
control of the spleen focus-forming virus promoter) of plasmid pMD9 (8) and inserting it
85
as an RsrII/NotI fragment into pΔΦ (Fig.2A). To see whether the identified HERV-Kcon
86
leader/gag ORF sequences are sufficient for transduction, several fragments covering
87
this area were PCR amplified and inserted as ClaI/RsrII fragments into pFV-eGFP. The
88
transduction efficiencies of the resulting pFV-eGFP/HERV-Kcon hybrid vectors were then
89
compared with that of the original pFV-eGFP.
90
91
While the FV-eGFP construct containing HERV-Kcon nucleotides (nt) 901 to 1740 resulted
92
in the highest transduction efficiency, constructs with shorter HERV-Kcon fragments were
93
weaker, but still active (Fig.2B), indicating that several sites act co-operatively. When
94
present in the inverse orientation, sequence 1740/901 transduced only poorly,
95
indicating that RNA secondary structure is important to transduction. Constructs pFV-
96
eGFP901/1740 and pFV-eGFP1740/901, which carry the cis-acting sequences CASI and
97
CASII necessary for delivery by FV virions (9), were produced to similar titres (1.7X105
98
and 1.5X105, respectively) when tested by the FV system (7). That inversion of the
99
901/1740 sequence diminished titre in HERV-Kcon virions but not in FV virions (which are
100
not dependent on a HERV-K packaging sequence) rules out the possibility of artefacts
101
from introduction of the inverted sequence and, thus, demonstrates that HERV-Kcon
102
positions 901 -1740 are necessary and sufficient to confer transduction by HERV-Kcon
103
virions. The experiments also indicate that the foreign FV LTRs are recognised and
104
integrated into the genome by HERV-Kcon integrase.
105
106
While generally retroviral packaging sequences are just 300 to 400 nt long (reviewed in
107
10), at about 840 nt the HERV-Kcon sequences necessary for transduction are even
4
108
longer than the 624 nt previously described for Mason Pfizer Monkey virus (11). Similar
109
to the situation in BLV, where a bi-partite packaging signal has been described (12), at
110
least two regions in HERV-Kcon work together cooperatively for full activity.
111
112
The activation of HERV gene expression by HIV-1 has been reported (2). We wanted to
113
see whether HERV-Kcon virions have the ability to transduce vectors carrying packaging
114
signals of HIV-1 and MLV. To this end, and to avoid problems that might arise from the
115
use of different vector backbones vector pCPX, a HERV-Kcon vector deleted between
116
positions 989 to 6640 was generated (Fig.3A). Packaging sequences from HIV-1 and MLV
117
of 310 and 304 nt, respectively, were PCR amplified and cloned as PacI/XbaI fragments
118
into pCPX, resulting in constructs pCPX/HIV-1 and pCPX/MLV. Their transduction titres
119
were compared, with pCPX and deletion mutant 1800/6639 serving as a negative and a
120
positive control, respectively. The results are shown in Fig.3B. HERV-Kcon virions
121
transduced construct pCPX/HIV-1 to about 8% of the HERV-Kcon positive control
122
p1800/6639 (Fig.3B, open bars) whereas pCPX/MLV sequences transduced similarly to
123
empty vector pCPX (1.6% and 1.3%, respectively). These results indicate that, provided
124
functional HERV-K proteins are produced, some HIV-1 packaging can occur. In this light
125
it is interesting that a newly detected HERV-K (HML2) provirus, termed K111, is
126
specifically active in HIV-1 infected individuals and the presence of virus-like particles
127
transmitting sequences to other cells have been described (13, 14). Moreover,
128
approaches aiming to clear HIV-1 infection should consider that some HERVs have the
129
potential to contribute to HIV transmission, yet may not be sensitive to anti-retroviral
130
therapies.
131
132
Many vectors commonly used in gene therapy experiments are based on MLV and HIV-
133
1. We investigated the above-mentioned pCPX hybrids for transduction by MLV capsids,
134
produced from expression plasmid pCG-gagpol (15) pseudotyped with VSV-G. We found
135
that HERV-Kcon vector p1800/6639 was recognised at 43% of pCPX/MLV, while
136
pCPX/HIV-1 was transduced at only 23% (Fig.3B, grey bars). When VSV-G pseudotyped
5
137
HIV-1 virions (generated from expression plasmid psPAX2) were used, transduction
138
efficiency was 16% and 15% of pCPX/HIV-1 for p1800/6639 and pCPX/MLV,
139
respectively (Fig3B, chequered bars).
140
141
Since all human packaging cells lines harbour HERV-K genomes, upon activation, they
142
might be packaged into MLV or HIV-1 virions. This could introduce new copies of
143
endogenous retroviruses into active chromatin, the preferred integration site of
144
lentiviral vectors (16). Furthermore, a contaminating HERV genome would retain any
145
LTR enhancer activity, a major contributor to genotoxicity that is exacerbated by MLVs
146
propensity to integrate within regulatory regions and near to oncogenes (16, 17). Since
147
there are increasing associations of HERV activation with human diseases, researchers
148
should be alert to the risk of introducing de novo endogenous virus insertions from
149
vectors used in human trials. [1255 words]
150
151
152
ACKNOWLEDGEMENTS
153
154
We are grateful to P. Bieniasz for providing the HERV-Kcon plasmids CHKCG, pCVI/Gag-
155
PR-Pol and pCR3.1/K-Rev. Plasmids pΔΦ and pMD9 were kindly donated by Professor D.
156
Russell and Professor D. Lindemann, respectively. Plasmids pMD.2 and psPAX2 were
157
generous gifts from Professor J. Luban.
158
159
160
REFERENCES
161
162
1. Hohn O, Hanke K, Bannert N. 2013. HERV-K(HML-2), the Best Preserved Family of
163
HERVs: Endogenization, Expression, and Implications in Health and Disease. Front
164
Oncol. 3: 1-12.
165
6
166
167
2. van der Kuyl AC. 2012. HIV infection and HERV expression: a review. Retrovirology 9:
6.
168
169
170
3. Lee YN, Bieniasz PD. 2007. Reconstitution of an infectious human endogenous
retrovirus. PLoS Pathog. 3(1):e10.
171
172
4. Durocher Y, Perret S, Kamen A. 2002. High-level and high-throughput recombinant
173
protein production by transient transfection of suspension-growing human 293-
174
EBNA1 cells. Nucleic Acids Res. 30(2):E9.
175
176
177
5. Yang J, Bogerd H, Le SY, Cullen BR. 2000. The human endogenous retrovirus K Rev
response element coincides with a predicted RNA folding region. RNA. 6: 1551-1564.
178
179
180
6. Rethwilm A, Bodem J. 2013. Evolution of foamy viruses: the most ancient of all
retroviruses. Viruses 5: 2349-2374. doi: 10.3390/v5102349.
181
182
183
7. Trobridge G, Josephson N, Vassilopoulos G, Mac J, Russell DW. 2002. Improved
foamy virus vectors with minimal viral sequences. Mol Ther. 3: 321-328
184
185
8. Heinkelein M, Dressler M, Jármy G, Rammling M, Imrich H, Thurow J, Lindemann
186
D, Rethwilm A. 2002. Improved primate foamy virus vectors and packaging
187
constructs. J Virol. 76: 3774-3783.
188
189
190
9. Erlwein O, Bieniasz PD, McClure MO. 1998. Sequences in pol are required for
transfer of human foamy virus-based vectors. J Virol. 72: 5510-5516.
191
192
193
10. D'Souza V, Summers MF. 2005. How retroviruses select their genomes. Nat Rev
Microbiol. 3: 643-655.
194
7
195
11. Vile RG, Ali M, Hunter E, McClure MO. 1992. Identification of a generalised
196
packaging sequence for D-type retroviruses and generation of a D-type retroviral
197
vector. Virology. 189: 786-791.
198
199
12. Mansky LM, Krueger AE, Temin HM. 1995. The bovine leukemia virus encapsidation
200
signal is discontinuous and extends into the 5’ end of the gag gene. J Virol. 69:3282-
201
9.
202
203
13. Contreras-Galindo R, Kaplan MH, Contreras-Galindo AC, Gonzalez-Hernandez MJ,
204
Ferlenghi I, Giusti F, Lorenzo E, Gitlin SD, Dosik MH, Yamamura Y, Markovitz DM.
205
2012. Characterization of human endogenous retroviral elements in the blood of
206
HIV-1-infected individuals. J.Virol. 86: 262-276.
207
208
14. Contreras-Galindo R, Kaplan MH, Dube D, Gonzalez-Hernandez MJ, Chan S, Meng
209
F, Dai M, Omenn GS, Gitlin SD, Markovitz DM. 2015. Human Endogenous Retrovirus
210
Type K (HERV-K) Particles Package and Transmit HERV-K-Related Sequences. J. Virol.
211
14: 7187-7201.
212
213
15. Ulm JW, Perron M, Sodroski JC, Mulligan RC. 2007. Complex determinants within
214
the Moloney murine leukemia virus capsid modulate susceptibility of the virus to Fv1
215
and Ref1-mediated restriction. Virology 363: 245-255.
216
217
218
16. Nowrouzi A, Glimm H, von Kalle C, Schmidt M. 2011. Retroviral vectors: post entry
events and genomic alterations. Viruses 3: 429-455.
219
220
17. Maruggi G, Porcellini S, Facchini G, Perna SK, Cattoglio C, Sartori D, AmbrosiA,
221
Schambach A, Baum C, Bonini C, Bovolenta C, Mavilio F, Recchia A. 2009.
222
Tanscriptional enhancers induce insertional gene deregulation independently from
223
the vector type and design. Mol Ther, 17: 851-856.
224
8
225
LEGENDS
226
227
Fig.1: Mapping of the sequences necessary to transduce HERV-Kcon vectors.
228
(A) Mutants generated from CHKCG, with the CMV/eGFP cassette in its env ORF
229
depicted in green. The nt positions refer to the provirus. Some important landmarks are
230
shown: 5’LTR, 1 - 968; gag ORF start codon, 1112; XbaI site, 6639; 3’LTR, 8505 - 9472;
231
RecRE indicates the sequence responsive to HERV-Kcon Rec protein for the nuclear
232
export of RNA, similar to the HIV-1 Rev/RRE system (5). Mutant STOP with a stop codon
233
after amino acid 228 was generated to allow an unbiased comparison of the titres. (B)
234
293T cells were transfected with the aforementioned mutant vectors from (A),
235
supernatant harvested and incubated with A375 cells. The number of eGFP-expressing
236
colonies were determined by fluorescence microscopy and the number per ml of
237
supernatant is shown. Experiments were carried out in triplicate. Error bars show the
238
standard error of the mean. Asterisks indicate values significantly different (P-value less
239
than 0.05) from pSTOP as determined by one-way analysis of variance with Dunnett’s
240
post-test using Graphpad Prism 5.0.
241
242
Fig.2: Generation of hybrid vectors to show that the leader/gag region of HERV-Kcon is
243
important for transduction by HERV-Kcon virions.
244
(A) Hybrid vectors between FV-eGFP and fragments of HERV-Kcon were generated. The
245
fragments were cloned between the ClaI and RsrII sites upstream of an SFFV/eGFP
246
reporter cassette, depicted in green. CASI/CASII represents cis-acting sequences,
247
necessary for transduction by FV virions (9) used in the control experiments to confirm
248
the integrity of the constructs. (B) Vector containing supernatant was transferred to
249
A375 cells and the number of eGFP-expressing colonies per ml of supernatant, as
250
determined by fluorescence microscopy, is shown. Data shown were generated from
251
triplicate experiments. Error bars show the standard error of the mean. Asterisks
252
indicate values significantly different (P-value less than 0.05) from pFV-eGFP as
9
253
determined by one-way analysis of variance with Dunnett’s post-test using Graphpad
254
Prism 5.0.
255
256
Fig.3: Transfer of HERV-K con vectors carrying HIV-1 and MLV packaging sequences.
257
(A) pCPX is a HERV-Kcon vector deleted between nt989 – 6639 of CHKCG and thus devoid
258
of its own sequences necessary for transduction serves as the negative control vector,
259
while p1800/6639 isthe positivecontrol vector. PCR amplified fragments carrying the
260
packaging sequences of HIV-1 (blue box in pCPX/HIV-1) and MLV (red box in pCPX/MLV)
261
were cloned between the PacI and XbaI sites of pCPX. (B) Data shown are from triplicate
262
transfection experiments. HERV-Kcon (open bars), MLV (red bars) and HIV-1 (blue bars).
263
Error bars show the standard error of the mean. One-way analysis of variance with
264
Dunnett’s post-test was employed using Graphpad Prism 5.0 to determine whether
265
vectors derived from pCPX containing foreign sequences were transduced by virions of
266
HERV-Kcon, MLV or HIV-1 to a greater extent than from empty pCPX. Significant
267
differences (P-value less than 0.05) are indicated by an asterisk.
268
10