healthmedicinet

HIV-1 genomic RNA binds to eEF1A in infected cells

To examine whether eEF1A interacts with HIV-1 genomic RNA, RC-co-IP experiment was
performed in HIV-1 infected TMZ-bl cells using an anti-eEF1A antibody in the presence
/absence of an HIV-1 reverse transcriptase inhibitor, nevirapine. Nevirapine is a
non-nucleoside RT inhibitor of enzymatic activity. The RNA was extracted from the
IP products followed by RT-PCR to examine the levels of HIV-1 RNA. A PCR without reverse
transcription reaction was performed as a control to detect DNA contamination. There
was a large amount of HIV-1 RNA detected in samples from HIV-1 infected cells compared
to infection with heat-inactivated virus (p??0.05), which are defective for viral entry, indicating that HIV-1 RNA co-immunoprecipitated
with eEF1A (Fig. 1a). Similar levels of RNA were detected from the infected cells incubated with nevirapine,
suggesting that the association of HIV-1 RNA with eEF1A is not dependent on RT activity.
Samples analyzed by qPCR, which omit the cDNA synthesis step, showed low levels of
viral DNA that were similar to the heat-inactivated samples indicating that the RNA
samples were not DNA contaminations (Fig. 1a). Importantly, RC-co-IP using a control antibody anti-eIF3A (Fig. 1a), did not specifically capture HIV-1 RNA suggesting that the interaction between
eEF1A and HIV-1 RNA in cells was specific.

Fig. 1. eEF1A binds to HIV-1 genomic RNA. a TMZ-bl cells were incubated with HIV-1 for 2 h at 4 °C and 2 h at 37 °C in the presence
or absence of nevirapine followed by RC-co-IP using anti-eEF1A or anti-eIF3A antibodies
as indicated. The level of RNA extracted from IP product was measured by RT-PCR targeting
HIV-1 5’UTR. Data are presented as means?±?SD from 3 independent experiments. TMZ-bl
cells were also transfected with biotin-labeled RNAs derived from HIV-1 genomic 5’UTR,
HIV-1 RT and luciferase sequences followed by RC-co-IP. The RNAs in cell lysates and
recovered from IP using anti-biotin antibody (b, upper panel) or anti-eEF1A antibody (b, lower panel) were detected by dot blot using streptavidin-peroxidase. The blot signals were quantified
using ImageQuant programme and the relative PI unit of IP product to lysate product
are presented in c. The biotin-labelled d 5’UTR, e RT or f luciferase RNAs were immobilized on the biosensors coupling with streptavidin. The
associations of RNAs with eEF1A were measured with OctetRed system using 90nM or 30nM
purified eEF1A protein and referenced using 90 nM of BSA in kinetic buffer. The dissociation
was measured by moving the biosensor to wells containing kinetic buffer only. g The association and dissociation of immobilized 5’UTR, RT and luciferase RNA with
90 nM of eEF1G was also measured as a control. Data are representative of 3 independent
experiments. * indicates p??0.05

eEF1A directly binds to the 5’UTR of HIV-1 RNA

The association of eEF1A and HIV-1 genomic RNA detected from infected cells could
be due to indirect binding through virus proteins. To determine whether eEF1A specifically
binds HIV-1 RNA in the absence of the viral proteins, equivalent amounts of biotin-labeled
RNAs corresponding to the 5’UTR (the 5’UTR RNA used corresponded to nucleotides 1
to 325 of the HIV RNA genome), a similarly sized RNA from the RT encoding region (RT
RNA) or a biotin-labelled luciferase RNA were transfected individually into TZM-bl
cells followed by a RC-co-IP assay using an anti-eEF1A antibody. The RNA extracted
from the IP products were subjected to a dot blot for biotin labelled RNA detection.
As shown in Fig. 1b (upper panel), the RNAs extracted from each lysate contained a similar level of biotin
labelled RNA. However a measurably larger amount 5’UTR RNA was recovered from anti-eEF1A
co-IP product than RT or luciferase RNAs (Fig. 1b, lower panel and Fig. 1c). The result suggests that 5’UTR RNA interacts with eEF1A in the absence of viral
protein.

5’UTR RNA binds to eEF1A in biolayer Interferometry (BLI) assay

While co-IP assays can detect associations between molecules, they cannot determine
if an association between molecules occurs by direct or indirect binding. To determine
if 5’UTR RNA has a direct interaction with eEF1A without any other cellular protein,
an in vitro BLI assays was employed using purified protein. The BLI method uses the interference
pattern of white light reflected from two surfaces on a biosensor probe: a layer of
immobilized molecule of interest, such as an RNA or protein, and an internal reference
layer. When the biosensor is immersed in a solution containing a molecule of interest,
any change in the number of molecules bound to the biosensor tip due to interaction
causes a shift in the interference pattern that can be measured. Streptavidin-coated
biosensors were saturated with biotin-labelled 5’UTR, RT or luciferase RNAs individually.
The association of each RNA with eEF1A was examined using the OctetRed system by incubating
each biosensor into kinetic buffer containing 90 nM and 30 nM of eEF1A. The 5’UTR
RNA strongly bound with eEF1A at 90 nM (Fig. 1d), while RT RNA showed a weak binding (Fig. 1e), and no binding was detected with luciferase RNA (Fig. 1f). Previously we showed that both eEF1A and eEF1G associated with the HIV-1 RTC 12]. However using BLI assay, no binding was detected between each RNA with 90 nM of
eEF1G protein (Fig. 1g), indicating that the association between 5’UTR RNA and eEF1A is specific. The results
are consistent with the RC-co-IP results and confirm a direct and specific interaction
between 5’UTR of HIV-1 genomic RNA and eEF1A.

The nt 106 to 224 of 5’UTR RNA is important for interaction with eEF1A

5’UTR of HIV-1 genomic RNA contains several well-defined RNA elements with important
roles in HIV-1 replication. These elements include TAR, PolyA loop, TLE, PBS, and
the stem-loops 1 to 3 (SL1 to SL3) (Fig. 2a) 15], 18]. To better define the eEF1A binding region in the 5’UTR RNA, two truncations of the
5’ UTR RNAs were synthesized. The first truncation was made by removing the stem-loops
SL1-SL3 (designated 5’UTR-?SL1-3). The second RNA contained only TAR and polyA regions
(designated TAR?+?polyA). The interaction of the truncated 5’UTR RNAs, 5’UTR-?SL1-3
and TAR?+?polyA with eEF1A were examined using BLI assay by immobilizing biotinylated
RNAs onto biosensors as described earlier. The 5’UTR-?SL1-3 RNA showed a similar binding
profile with eEF1A as the intact 5’UTR RNA, and the maximum responses (nm) of the
interaction with 90 nM eEF1A is 0.32 nm (no significant difference compared to the
intact 5’UTR), whilst the interaction of TAR?+?polyA with eEF1A was significantly
reduced (p??0.05) (Fig. 2b). The results suggest that RNA sequences after polyA and before SL1-3, which includes
the TLE and PBS (nt 106 to 224), are important for optimal interaction with eEF1A.

Fig. 2. The stem-loop formed by nucleotide 142 to 170 in 5’UTR of HIV-1 genomic RNA is important
for interaction with eEF1A. a The predicted secondary structure of HIV-1 5’UTR is formed by TAR, polyA, tRNA anticodon-like
element (TLE), PBS and stem-loop 1, 2, 3 (SL1, 2, 3). The arrows indicate the sites
of truncated RNA. b Biotin labelled wild-type, truncated and mutated 5’UTR RNAs were immobilized on biosensors.
The maximum responses (nm) during a 600 s incubation of each biosenor with 90 nM of
eEF1A protein using OctetRed system are shown. c The local secondary structures of wild-type, loop and bulge mutated RNAs and their
minimum free energy (?G) were predicted using Mfold 21]. All data sets are presented as mean?±?SD from at least 3 independent experiments
and * indicates p??0.05

Alterations of the stem-loop structure that contains TLE in 5’UTR RNA impair eEF1A
binding

The region of the viral genome from nt 106 to 224 contains a highly conserved RNA
stem-loop structure 17], which contains a TLE 15]. To examine whether this RNA stem-loop is important for the interaction with eEF1A,
two mutants of 5’UTR were made in two highly conserved sequence clusters 19], 20]. One mutant was created by changing CCC of nt 142 to 144 in the bulge region to GGG
(designated Bulge-M) and another mutant was generated by changing CCC of nt 150 to
152 in the loop region to GGG (designated Loop-M) (Fig. 2a). RNA-folding analysis shows that introductions of these mutations alter the local
RNA structures formed by nt 142 to 170 (Fig. 2c) 21]. BLI assay were performed to examine binding of the mutated 5’UTR RNAs with eEF1A
protein as described above. The binding of the two mutant 5’UTR RNAs with eEF1A were
significant reduced compared to wild-type RNA (p??0.05) (Fig. 2b), indicating that the RNA stem-loop structure formed by nt 142 to 170 are important
for binding by eEF1A.

Association of eEF1A with reverse transcriptase was reduced in cells infected with
viruses that contain the bulge and loop mutations

The Bulge-M and Loop-M mutations were introduced into a HIV-1 proviral plasmid DNA
and the virus was made by plasmid transfection into HEK293T cells. An equivalent amount
of each mutant and wild-type virus, normalised to CAp24 levels, was used to infect
TZM-bl cells. Association of eEF1A with HIV-1 RT, a surrogate marker of HIV-1 reverse
transcription complex (RTC), was determined by PLA at 2 h post-infection. PLA is a
modified fluorescent in situ hybridization method to detect protein-protein interactions that produces fluorescent
foci if the two PLA antibodies are in proximity. Heat-inactivated virus, which is
defective for viral entry, was used to measure non-specific foci made by the PLA antibodies.
There was a significant reduction in the number of foci present in cells infected
with the mutant compared to wild type virus (P??0.01), while inactivated wild type, which are defective for viral entry, virus
produced very few foci (Fig. 3a). The levels of virus entry into cells, as determined by measurement of genomic RNA
in the cytoplasm at 2 h of post-infection, were similar indicating that reduced levels
of eEF1A-RTC association in mutated virus infection were not due to differences in
viral entry (Fig. 3b). As expected, heat-inactivated virus sharply reduced levels of viral RNA in cells.
The results suggest that the RNA stem-loop structure altered by the mutations is important
for interaction between eEF1A and RT.

Fig. 3. Bulge and loop mutations in 5’UTR of HIV-1 genomic RNA resulted in reduced association
of eEF1A and RT in virus infected cells. a TZM-bl cells were infected with WT, Loop-M, Bulge-M virus as indicated equal to 100 ng
CAp24. Association of eEF1A with HIV-1 RTC in virus infected TZM-bl cells was determined
by proximity ligation assay using anti-eEF1A and anti-RT antibodies at 2 h of post-infection.
A red foci represents a positive signal that was visualized using a DeltaVision Core
imaging system and analysis was performed from more than 200 cells. b The levels of genomic RNA at 2 h of post-infection were determined by RT-PCR and
possible DNA contamination was examined by PCR without RT

The Bulge-M and Loop-M mutations in 5’UTR of HIV-1 genome affect late steps in HIV-1
reverse transcription

The effect of the Bulge-M and Loop-M mutations on reverse transcription and replication
was examined by infecting Jurkat cells with equivalent amount of either wild-type
or mutant viruses. The total cellular nucleic acids was extracted from cytoplasm at
4 h of post-infection followed by qPCR detection of the early (minus strand strong-stop
DNA, ssDNA) and late (2
^d
strand transfer DNA) reverse transcription products. While the levels of early DNA
from the three virus stocks infection were similar (P??0.05), the 5’UTR RNA mutations resulted in a significant reduction in the amount
of late RT product detected (p??0.05) (Fig. 4a). The levels of late DNA from the Bulge-M and Loop-M mutated virus infections dropped
3-4 fold compared with wild type virus infection (Fig. 4a). Jurkat cells individually infected with wild type or mutant viruses were monitored
for virus replication by measuring CAp24 in culture supernatant for 2 weeks. Both
mutant viruses showed a significant reduction in replication kinetics compared to
wild type virus (Fig. 4b). The results indicate that the structural changes in the RNA introduced by the two
mutations, which affect eEF1A binding to the RNA, lead to reduced efficiency of reverse
transcription as observed by reduced levels of late DNA products compared to early
DNA products. Hence, the mutations affect the replication of HIV-1 in Jurkat cells.

Fig. 4. HIV-1 with bulge and loop mutations in 5’UTR of HIV-1 genomic RNA undergo reduced
levels of reverse transcription and replication in Jurkat cells. a Jurkat cells were infected with equivalent amount of wild type, bulge and loop mutated
HIV-1. Viral DNA was extracted from cytoplasm at 4 h post-infection and analysed by
quantitative PCR measuring early and late reverse transcription DNA products. b Virus replication was monitored for 14 days by measuring CAp24 in culture supernatant
at 3, 7 and 14 days post-infection. Data are presented with mean?±?SD of 3 independent
experiments