Peptides derived from evolutionarily conserved domains in Beclin-1 and Beclin-2 enhance the entry of lentiviral vectors into human cells

Autophagy-related proteins such as Beclin-1 are involved in an array of complex processes, including antiviral responses, and may also modulate the efficiency of gene therapy viral vectors. The Tat-Beclin-1 (TB1) peptide has been reported as an autophagy-inducing factor inhibiting the replication of pathogens such as HIV, type 1 (HIV-1). However, autophagy-related proteins are also essential for the early steps of HIV-1 infection. Therefore, we examined the effects of the Beclin-1 evolutionarily conserved domain in TB1 on viral transduction and autophagy in single-round HIV infection or with nonreplicative HIV-1–derived lentiviral vectors. TB1 enhanced transduction with various pseudotypes but without inducing the autophagy process. TB1 augmented the transduction of human CD34+ hematopoietic stem/progenitor cells while maintaining their capacity to engraft in vivo into humanized mice. TB1 was as effective as other transduction additives and functioned by enhancing the adhesion and fusion of viral particles with target cells but not their aggregation. We also found that the N-terminal L1 loop was critical for TB1 transduction–enhancing activity. Interestingly, the Tat-Beclin-2 (TB2) peptide, derived from the human Beclin-2 protein, was even more potent than TB1 in promoting viral transduction and infection. Taken together, our findings suggest that the TB1 and TB2 peptides enhance the viral entry step. Tat-Beclin peptides therefore represent a new family of viral transduction enhancers for potential use in gene therapy.

The lysosomal degradation pathway of autophagy has a dual role in HIV, type 1 (HIV-1) 3 replication and pathogenesis (1).
In this study, we evaluated the effect of TB1 on single-round infections with HIV-1 and also on LVs pseudotyped with various envelope glycoproteins; namely, the Chikungunya virus glycoprotein (CHIKV-G), the modified gibbon ape leukemia virus glycoprotein (GALVTR), the modified RD114 feline endogenous retrovirus (RD114TR), and the vesicular stomatitis virus glycoprotein (VSV-G), the latter being broadly used in gene therapy. Surprisingly, we observed that the TB1 peptide was able to strongly promote viral infection or transduction with all LV pseudotypes tested, either on cell lines or hCD34ϩ hematopoietic stem/progenitor cells (HSPCs). Consequently, This work was supported by the Association Française contre les Myopathies (AFM/Telethon). The authors declare that they have no conflicts of interest with the contents of this article. This article contains supplemental Figs. S1-S9, Table S1, and Materials and Methods. 1 To whom correspondence may be addressed: Généthon, UMR_S951, 1bis rue de l'Internationale, 91000 Evry, France. Tel.: 33-1-69-47-34-40; E-mail: galy@genethon.fr. 2 To whom correspondence may be addressed: TxCell SA, Allée de la Nertière, les Cardoulines, 06560 Valbonne-Sophia Antipolis, France. cro ARTICLE the potential toxicity of TB1 has been evaluated on human HSPCs either in vitro (CFC assay) or in vivo (humanized NSG mice). We also investigated which steps of the viral life cycle are targeted by TB1. Finally, this study was extended through the design of various TB1 variants and a new peptide called Tat-Beclin-2 (TB2), a fusion of the Tat (47-57) transduction peptide with the human Beclin-2 ECD 249 -266 (12).

Low doses of Tat-Beclin-1 strongly improved cell line transduction with various lentiviral pseudotypes and with HIV-1
To evaluate the effect of the TB1 peptide on LV transduction, we used the human colon carcinoma cell line HCT116, which is routinely employed in our laboratory, to titer LV pseudotyped with the VSV-G envelope (VSV-G-LVs). Using a low concentration of LV, we found that the TB1 peptide enhanced the transduction of HCT116 cells in a dose-dependent manner and up to 8-fold compared with the control Tat-Scrambled (TS) peptide (Fig. 1A). Further increasing the TB1 peptide concentration was associated with a bell-shaped dose-response curve. At the same molarity, neither the Tat peptide alone nor the Beclin-1 domain alone were capable of promoting lentiviral transduction (supplemental Fig. S1). The TB1 effect was not saturable over a one-log concentration of VSV-G-LV (corresponding to a multiplicity of infection of 0.5 to 5), reaching up to 84% of transduction efficiency (Fig. 1B). The improvement in transduction efficiency, evaluated by transgene expression levels (i.e. GFP), was also confirmed by proviral DNA integration following quantification by qPCR of vector copy numbers per cell (supplemental Fig. S2).
A great advantage of LVs for gene transfer is their capacity to be pseudotyped with numerous heterologous envelope glycoproteins for specific cell targeting (13). Certain hematopoietictropic LV pseudotypes (GALVTR-LV and RD114TR-LV) require the use of culture additives to promote efficient transduction. These additives include Polybrene, protamine sulfate, or the recently identified Vectofusin-1 peptide (14,15). As shown in Fig. 1C, TB1 was capable to promote cell line transduction with GALVTR-and RD114TR-LVs to an extent comparable with other culture additives.
The enhancing effects of TB1 on LV transduction are in apparent contradiction with reports that describe TB1 as an inhibitor of HIV-1 and Chikungunya virus replication (5). This prompted us to verify the effect of TB1 on the early phase of live virus infection using a single-round infection assay in the MAGIC 5B permissive cell line. TB1 efficiently promoted the infection of wild-type HIV-1 (pNL4.3 molecular clone) ( Fig.  2A) and also of HIV-1 pseudotyped with the Chikungunya GP (CHIKV-LV), although not statistically (Fig. 2B). TB1 is therefore defined as a lentiviral transduction enhancer of HIV-1 vectors or viruses bearing a large panel of envelope glycoproteins.

Tat-Beclin-1 promotes safe lentiviral transduction of hematopoietic stem/progenitor cells
An important goal for the use of LVs is to achieve clinically relevant levels of transduction of HSPCs for ex vivo gene therapy approaches. As shown in Fig. 3A, highly purified VSV-G-LV particles were used to transduce human CD34ϩ HSPCs. In the presence of TB1, a 2-fold increase in lentiviral transduction was observed (Fig. 3, A and B). The optimal dose of TB1 to promote HSPCs was defined to be around 10 M, twice the concentration used on cell lines (data not shown). Because it has been shown previously that high doses of TB1 are triggering a specific kind of cell death called autosis (16), safety studies were performed on HSPCs to examine the effects of TB1. Incubation of hCD34ϩ cells with increasing concentrations of TB1 led to partial cell death only at concentrations above 20 M Tat-Beclin-1 (5 M), Vectofusin-1 (6 g/ml), protamine sulfate (4 g/ml), or Polybrene (3 g/ml). In all three panels, transduction efficiencies were evaluated after 3 to 4 days by monitoring GFP expression. All data are expressed as the average of three independent experiments performed in duplicate Ϯ S.E. (supplemental Fig. S3). A colony-forming cell (CFC) assay was used to evaluate the effect of TB1 on the hematopoietic differentiation of hCD34ϩ cells. As shown in Fig. 3C, TB1 did not affect the absolute number of each type of colonies compared with the scrambled control or with cells treated with RetroNectin, a fibronectin fragment peptide used in gene therapy protocols to preserve HSPCs during transduction and to enhance transduction by co-localizing viral particles and target cells (17,18). We previously reported the enhancing effects of RetroNectin in our system (19).

Beclin-derived peptides promote viral transduction
To extend our safety studies to an in vivo system, TB1treated HSPCs were injected into the immunodeficient NSG mouse model, and the engraftment efficiency was evaluated after 12 weeks. As shown in Fig. 4A, the engraftment of hCD34ϩ cells in the bone marrow was comparable (around 20%) between all conditions, either in the absence or presence of TS, TB1, or the control RetroNectin. Similarly, transduced GFPϩ hCD34ϩ cells were present in the bone marrow (Fig.  4B), with a slightly better efficiency than RetroNectin, although the difference was not statistically significant. Comparable engraftment of transduced cells was observed in the spleen, thymus, and blood (data not shown). Altogether, this encouraging safety profile of TB1 indicates that it is compatible with clinical applications involving lentiviral gene transfer.

Tat-Beclin-1 acts on the adhesion and fusion steps of LVs with target cell membranes
In terms of timing, TB1 pretreatment was found to be inefficient for viral transduction, and addition of TB1 only 3 h after the start of transduction led to a 50% drop in TB1 efficacy (supplemental Fig. S4), suggesting that TB1 acts at the early steps of viral transduction. Because LV entry into target cells is a ratelimiting step, we investigated whether low doses of TB1 were capable of enhancing adhesion and fusion of LVs with target cell membranes. For that, the BLAM-LV fusion assay was used (19) and showed that TB1 promoted efficient viral fusion (Fig.  5A). Next, by quantifying the number of viral particles interacting with target cells at 4°C, it was shown that the level of viral adhesion was augmented in the presence of TB1, reaching levels comparable with those of Vectofusin-1 (Fig. 5B). This

Beclin-derived peptides promote viral transduction
increase in viral adhesion was the same whether cells were preincubated with TB1 at 37°C or not (data not shown). Interestingly, TB1 also promoted viral transduction after the adhesion of viral particles as efficiently as Vectofusin-1, described previously to act on the viral fusion step (Fig. 5C). Hence, this enhancement in viral fusion is, at the same time, the consequence of an increase in viral adhesion and post-adhesion steps. However, this increase in viral entry was not caused by TB1-induced aggregation of viral particles, contrary to Vectofusin-1 (Fig. 5D). The effect was also not caused by an increase in cell surface expression of retroviral receptors (supplemental Fig. S5).

The effect of TB1 on lentiviral transduction seems to be independent of the autophagy process
To better understand the mechanism of action of TB1, we sought to investigate whether it involved the activation of the autophagy process. To monitor autophagy at the single-cell level, HEK293T cells were transfected with a plasmid expressing the mCherry-eGFP-LC3 fusion protein to mark the autophagolysosomes in the cells (supplemental Fig. S6). Next, cells were incubated with TB1 (5 M) or EBSS starvation solution and analyzed with an imaging flow cytometer. As expected, EBSS increased the number of autophagolysosomes (mCherry spots) per cell (Fig. 6A). On the contrary, the number of autophagolysosomes in TB1-treated HEK293T cells was comparable with that under control conditions in the absence of peptide (Fig. 6A, None). The autophagy level was also determined by monitoring LC3 lipidation using immunoblot experiments. As expected, an increase in the LC3-II form was observed in the presence of EBSS, but the results obtained with TB1-treated cells and control TS-treated cells were comparable. Consistently, the level of p62/SQSTM1, which is known to be degraded by autophagic flux, decreased under starvation conditions (EBSS) but did not vary between cells treated with TS or the TB1 peptide (Fig. 6B). In parallel, we confirmed that TB1 is able to induce autophagic flux at high concentration, as described previously (5) (supplemental Fig. S7). Finally, the use of various PI3K inhibitors, which are known to block the autophagy process, did not impact TB1-induced lentiviral transduction (supplemental Fig. S8). Altogether, these data suggest that the improvement of lentiviral transduction observed in the presence of TB1 is certainly not the consequence of an induction of autophagy flux.

Critical role of the N-terminal L1 loop of Beclin-1 for efficient promotion of viral transduction
To better define the critical molecular determinants in the TB1 peptide necessary for efficient promotion of viral transduction, numerous peptides overlapping the human Beclin-1 protein from position 250 to 300 were designed (Fig. 7A). As a reference, the Beclin-1 domain of TB1 (267-284) is also represented. All of these peptides were tested for their capacity to promote the infectivity of VSV-G-LV pseudotypes. As shown in Fig. 7B, the more potent TB1 variant is Tat-Bec (250 -282), corresponding to the ␣ helix1/loop1/␤1 region. This peptide is three times more efficient than TB1 at only 2.5 M. Interestingly, a gradual loss in viral transduction efficiency is observed when the peptide variants are less and less encompassing the N-terminal region of the L1 loop, suggesting that this short domain is critical for viral infectivity improvement.

The Tat-Beclin-2 peptide derived from human Beclin-2 is more efficient than TB1 for lentiviral transduction enhancement
In 2013, He et al. (12) identified a new mammal-specific protein called Beclin-2 (12). Beclin-2 behaves in autophagy like Beclin-1 but also plays a major role in an additional lysosomal degradation pathway. Sequence alignment of the human Beclin-1 and Beclin-2 proteins shows a high degree of homology between the ECDs. Therefore, the Tat-Beclin-2 (TB2) peptide, a fusion of the Tat (47-57) peptide with human Beclin-2 ECD 249 -266 , was designed (Fig. 8A). Because TB1 contains three mutations (H275E, S279D, and Q281E), the Tat-BecWT peptide corresponding to the fusion of the Tat (47-57) peptide with wild-type human Beclin-1 ECD 267-284 was also designed and is represented in the sequence alignment ( Fig. 8A and supplemental Table S1). These three peptides were tested for their capacity to promote lentiviral transduction over a large range of concentrations. As shown in Fig. 8B, all peptides promoted lentiviral transduction but with a large variability in their optimal doses: 3 M for TB1, 1 M for Tat-BecWT, and 500 nM for TB2. As few as 100 nM TB2 increased lentiviral transduction by 3-fold, from 10% to 30%, whereas TB1 and Tat-BecWT had no

Beclin-derived peptides promote viral transduction
effect at this concentration (Fig. 8B). Similarly tested as in Fig. 2, TB2 efficiently promoted the infection of wild-type HIV-1 (Fig.  8C) and also of CHIKV-LV (Fig. 8D). Tat-Scr2, a scrambled version of TB2, behaved as a negative control in the concentration range for TB2 activity. At higher concentration, Tat-Scr2 promoted lentiviral transduction, but only in HCT116 cells (Fig. 8B), not in MAGIC 5B (Fig. 8C) or HEK293T cells (Fig.  8D), possibly as a result of interactions between this peptide and specific membrane components of HCT116 cells. Interestingly, like TB1 and Tat-Bec(250 -282), TB2 is capable of promoting viral transduction after viral adhesion (supplemental Fig. S9), excluding a simple nonspecific effect of TB2 on viral adhesion. In conclusion, the TB2 peptide is a potent enhancer of lentiviral transduction at very low doses.

Discussion
Our results show that the TB1 peptide, at low doses, can be a potent enhancer of the entry of LV into target cells without inducing apparent autophagy in the cells. These results are not inconsistent with the more complex effects TB1 can exert as a potent inducer of autophagy and as an efficient antiviral agent on replicative viruses at higher doses (5), considering that different conditions are involved. Here we observed that a short exposure of cells to TB1 efficiently promoted the transduction of cell lines and HSPCs with various non-replicative HIV-1derived lentiviral pseudotypes (VSV-G-LV, RD114TR-LV, GALVTR-LV, and CHIKV-LV) as well as HIV-1 infection in vitro in single-round assays. Such findings are compatible with the notion that the replication of various enveloped viruses (i.e. HIV-1, VSV, CHIKV, and influenza virus) requires the expression of autophagy-related factors (3, 20 -24). Our new findings regarding TB1 peptide properties are compatible with applications in gene therapy. The safety profile of TB1 in HSPCs, the lack of effect on differentiation of hCD34ϩ cells in vitro, and the possibility to engraft TB1-exposed hCD34ϩ cells into the bone marrow of humanized NSG mice support the use of TB1 as a transduction additive to promote CD34ϩ cell transduction with LV in ex vivo gene therapy protocols. Additional preclini-

Beclin-derived peptides promote viral transduction
cal evaluation, including toxicity studies, will be necessary in this regard.
The use of numerous TB1 variants, encompassing the human Beclin-1 protein from position 250 to 300 (␣1-L1-␤1-␤2-L2), suggests that the N-terminal region of the L1 loop is critical for viral infectivity improvement. This critical domain is described as the docking site of HIV-1 Nef, allowing the virus to modulate autophagy through this specific interaction with Beclin-1. The L1 loop is also the target of the Golgi-associated protein GAPR-1 (5). GAPR-1 belongs to the cysteine-rich secretory proteins, antigen 5, and pathogenesis-related 1 proteins (CAP) superfamily, found in a remarkable range of organisms (25). GAPR-1 could be seen as an innate immune effector, triggered upon viral infection. However, we never observed any negative impact of GAPR-1 overexpression on transduction efficiencies (data not shown). It cannot be excluded that other cellular factor(s), critical for transduction enhancement, interact with the L1 loop of Beclin-1.
In the viral cycle, TB1 acts on the adhesion and post-adhesion step, leading to an increase in viral fusion, a phenomenon that is not the consequence of an increase in the surface expression or the receptor density of various retroviral receptors (supplemental Fig. S5). Furthermore, contrary to transduction enhancers forming peptide nanofibers (26), TB1 does not pro-mote viral entry through the induction of viral particle aggregation (Fig. 5D). It is unlikely that TB1 or TB2 exert some nonspecific cationic effect capable to improve viral adhesion. Indeed, neither the Tat peptide alone nor the Beclin-1 domain alone was capable of promoting lentiviral transduction. Furthermore, TB2 is a lot more efficient than TB1 in promoting lentiviral transduction, and Tat-Scrambled is inefficient, but all of these peptides contain the same number of cationic charges. All of these features are rather reminiscent of a recently described transduction enhancer called ␤-Deliverin (27). ␤-Deliverin seems to promote viral adhesion/fusion by modulating endosomal maturation and trafficking. In the past, some host factors related to the autophagy process have been described as modulators of viral entry. For instance, the UV radiation resistance-associated gene protein (UVRAG) has been shown to regulate the VSV entry by promoting the formation of a fusogenic SNARE complex involving vesicle-associated membrane protein 8 (24). However, that study exclusively described a mechanism that allowed enveloped viruses to interact with late acidic endosomes. In our study, Tat-Beclin peptides efficiently promote not only lentiviral pseudotypes that are pH-dependent, like VSV-G-LVs or CHIK-LVs (28), but also pH-independent pseudotypes (e.g. GALVTR-LVs (29) or RD114TR-LVs (30)), for which fusion occurs at the cell surface and/or in early endosomes. Further investigations will be needed to determine whether the relief of viral entry exerted by Beclin ECD-derived peptides could be a consequence of the modulation of endosomal maturation and trafficking.

Beclin-derived peptides promote viral transduction
In conclusion, autophagy-related factors, like Beclin-1 or UVRAG proteins, play a crucial role not only in the autophagy process but also in cellular endocytic membrane trafficking and fusion events. In fact, many host autophagy molecular factors have to be seen in a broader way as endomembrane trafficking regulators. Therefore, modulation of Beclin-1 and 2 functions using Beclin ECD-derived peptides could be of high interest for the development of more efficient vaccines or gene therapy protocols through an optimal virus and viral vector entry.

Viral vector production and titering
LVs were generated as described previously (32). Briefly, HEK293T cells were transiently transfected using calcium phosphate transfection with four plasmids: the gagpol (pKLgagpol) and rev (pKrev) expression plasmids, a transfer plasmid (pCCLsin.cPPT.hPGK.eGFP.WPRE), and a plasmid encoding either the VSV-G (pMDG) envelope glycoprotein (GP), the GALVTR GP (pBA.GALV/Ampho-Kana), or the RD114TR GP (pHCMV-RD114TR). After 24 h of production, raw viral supernatants were harvested, filtered (0.45 m), and frozen at Ϫ80°C. The purification of GFP-expressing VSV-G-LVs, through several membrane-based and chromatographic steps, has been described previously (31). Infectious titers were determined in HCT116 cells using either detection of GFP by flow cytometry (FACSCalibur, BD Biosciences, Le Pont de Claix, France), with titers expressed as transducing units (TU) per milliliter (33), or using qPCR with titers expressed as infectious genome (ig) per milliliter (31). The same qPCR approach was used to measure integrated vector copy numbers per cell following transduction. CHIKV-LVs were produced by cotransfection of the self-inactivating transfer plasmid pHRЈS-INcPPT-SGW (34), the lentiviral packaging plasmid psPAX2 (a gift from Didier Trono, Addgene, 12259), and the pCAGGS-EnvCHIKV plasmid encoding CHIKV GP (35). After 48 h of production, viral supernatants were harvested, filtered (0.22 m), and stored at Ϫ80°C. Viral inputs were calibrated to infect around 10% of the cell culture. Transduction by CHIKV-LVs was monitored by quantification of GFP expression in total cell lysates using an Infinite F200PRO fluorometer (Tecan, Salzburg, Austria), and values were normalized according to protein content (BCA assay, Pierce).
HIV-1 stocks were prepared from the culture supernatant of HEK293T cells 48 h after transfection with the pNL4.3 provirus

Beclin-derived peptides promote viral transduction
using TurboFect reagent (Thermo Fisher Scientific). After filtration (0.45 m), viral supernatants were concentrated by ultracentrifugation (25,000 rpm for 2 h at 4°C). HIV-1 and LV physical titers were obtained by measuring HIV-1 p24 capsid contents using commercial ELISA kits from Fujirebio/Innogenetics and PerkinElmer Life Sciences. The levels of viral infection were determined by quantifying ␤-galactosidase activity using the Galacto-Star ␤-galactosidase reporter gene assay system (Thermo Fisher Scientific) according to the instructions of the manufacturer. Briefly, 0.2.10 6 MAGIC 5B cells were infected in the presence of different peptides for 6 h. Cells were then washed five times and cultured for 72 h in DMEM and 10% FCS containing 1 M AZT. Cells were then lysed in 200 l of lysis buffer provided by the manufacturer. 50 l of cell lysate was used in each assay in duplicate. Results were normalized by quantifying the total protein content using a Bradford assay (Sigma-Aldrich).

Human CD34؉ cell culture and transduction
Umbilical cord blood (UCB) samples were obtained after uncomplicated births and in accordance with international ethical principles and French national law under declaration DC-201-1655 to the French Ministry of Research and Higher Studies. Human CD34ϩ cells were isolated by immunomagnetic selection (Miltenyi Biotec). The survival rate of fresh or frozen hCD34ϩ cells was evaluated using the trypan blue exclusion method. Next the preactivation of hCD34ϩ cells was performed overnight as described previously (19). Preactivated cells were plated in 96-well plates, and transduction was initiated by adding the desired amount of LV particles mixed with or without the peptides of interest. At 6 h post-transduction, reactions were diluted by adding differentiation medium to each well. After 4 -6 days, cellular mortality and transduction efficiency were evaluated, respectively, by 7-amino-actinomycin D labeling and measurement of GFP expression using flow cytometry (FACSCalibur, BD Biosciences).

Viral pulldown assay
The pulldown of LV particles in presence of culture additives was adapted from a protocol described previously (36). Briefly, the VSV-G-LV supernatant was diluted to 100 ng/ml of p24 with X-Vivo20 medium equilibrated at room temperature. Next, 1.5-ml tubes were loaded with 500 l of LV suspension in the absence or presence of the indicated culture additive (10 m). After homogenization, samples were centrifuged at low speed (15,000 ϫ g) for 5 min at room temperature. Then the supernatant was discarded, and the pellet was suspended in 100 l of fresh medium and frozen at Ϫ20°C. For each condition, the amount of pelleted p24 was evaluated using a commercial HIV-1 p24 ELISA kit as described above.

Adhesion, post-adhesion transduction, and BLAM-LV fusion assay
The protocol for LV adhesion to target cells has been described previously (37). Briefly, viral supernatants were incubated for 3 h at 4°C in HCT116 cells in the absence or presence of culture additives. Next, cells were washed three times with cold 1ϫ PBS and lysed in 1ϫ PBS containing 1% Triton X-100 and a protease mixture inhibitor, Complete (Roche Diagnostics). p24 contents in lysates were evaluated using a commercial HIV-1 p24 ELISA kit, and data were normalized to total protein content using the DC protein assay (Bio-Rad). For the postadhesion transduction assay, after incubation of 2.5 h at 4°C in HCT116 cells, viral supernatants were washed twice with cold PBS. Next, the indicated peptides were added to the wells, and transduction was performed at 37°C for 6 h. Transduction efficiencies were evaluated by monitoring GFP expression after 3 days. The BLAM-LV assay has been described extensively previously (19).

Western blot
Cell lysates were loaded in 4 -20% precast gels (Bio-Rad) and transferred to PVDF membranes. After a blocking step for 1 h at room temperature in PBS containing 0.5% casein, membranes were incubated overnight at 4°C with anti-LC3B antibody (Sigma-Aldrich) or anti-p62/SQSTM1 antibody (Ozyme) in blocking buffer. After three washes with PBS supplemented with 0.05% Tween, the membranes were incubated for 1 h at room temperature with peroxidase-coupled secondary antibody. Upon extensive washes, membranes were incubated with Luminata Western HRP substrate (Millipore, Molsheim, France). Proteins were detected by chemiluminescence and imaged with a G-box camera (Syngene imaging system, Cambridge, UK). The expression level of GAPDH was used as a loadingcontrol.TheformofLC3conjugatedtophosphatidylethanolamine, named LC3-II, presents a higher electrophoretic mobility in gels and is used to determine the autophagy level. Autophagic flux can be monitored by quantification of the p62/ SQSTM1 level.

Autophagy assay based on imaging flow cytometry (ImageStream)
Using the transient calcium phosphate transfection method, HEK293T cells were transfected with the pBABE-puro-mCherry-eGFP-LC3B expression plasmid (38), a gift from Jayanta Debnath (Addgene plasmid 22418). Next, cells (5 ϫ 10 5 cells/well) were incubated in the absence or presence of Tat-Beclin-1 (5 M) or in Earle's balanced salt solution (EBSS) for 6 h at 37°C, 5% CO 2 . Next, cells were washed in 1ϫ PBS, fixed (1.2% paraformaldehyde), and analyzed using ImageStream (Amnis Corp., Seattle, WA). Images of cells were acquired in a bright-field channel, in an mCherry-fluorescence channel, and finally in the SSC channel (742 nm) using a ϫ40 objective and the lowest flow velocity to optimize sensitivity. After acquisition, images were treated with Ideas software. Focused images were gated on a histogram displaying Gradient_RMS feature values in the bright-field channel between 40 and 90. Then a scatterplot of area versus aspect ratio (in the brightfield channel) was used to gate on single cells and remove doublets of cells. When gated on single cells, the spots on each cell were counted in the mCherry channel using the automatic spot counting wizard included in the Ideas software. Data were exported as text files and processed with GraphPad Prism 5.

Conditioning and reconstitution of NSG mice
Mice were housed in an accredited facility (CERFE, Evry, France) in a specific pathogen-free environment (French National Agreement C91.228.101). Experiments were performed according to the institutional and international guidelines for animal care and use. Experimental procedures were performed in accordance with French and European directives, approved by the local ethical committee (C2EA-64) and Ministry of Research for Gene-modified Organism Studies (Agreement 5244-CAI). NSG mice 3-4 weeks of age were conditioned intraperitoneally with two doses of busulfan (25 mg/kg). Human CD34 ϩ HSPCs from cord blood transduced with VSV-G-LV were injected intravenously into mice 24 h after the last dose of busulfan. Twelve weeks post-injection, the levels of engraftment and hematopoietic reconstitution were evaluated by immunophenotyping of human cells.