The mutation of Transportin 3 gene that causes limb girdle muscular dystrophy 1F induces protection against HIV-1 infection

The causative mutation responsible for limb girdle muscular dystrophy 1F (LGMD1F) is one heterozygous single nucleotide deletion in the stop codon of the nuclear import factor Transportin 3 gene (TNPO3). This mutation causes a carboxy-terminal extension of 15 amino acids, producing a protein of unknown function (TNPO3_mut) that is co-expressed with wild-type TNPO3 (TNPO3_wt). TNPO3 has been involved in the nuclear transport of serine/arginine-rich proteins such as splicing factors and also in HIV-1 infection through interaction with the viral integrase and capsid. We analyzed the effect of TNPO3_mut on HIV-1 infection using PBMCs from patients with LGMD1F infected ex vivo. HIV-1 infection was drastically impaired in these cells and viral integration was reduced 16-fold. No significant effects on viral reverse transcription and episomal 2-LTR circles were observed suggesting that the integration of HIV-1 genome was restricted. This is the second genetic defect described after CCR5Δ32 that shows strong resistance against HIV-1 infection.

Author summary TNPO3 has been described as a key factor in the infection by the human immunodeficiency virus (HIV-1), the causative agent of AIDS. In 2013, a relationship between a genetic defect in TNPO3 gene and a rare muscle disease named Limb Girdle Muscular Dystrophy 1F (LGMD1F), with an autosomal dominant transmission, was discovered.
LGMD1F patients show a heterozygous single nucleotide deletion in the TNPO3 gene that generates a TNPO3_mut protein. Our results demonstrate that cells from patients with

Introduction
Productive HIV-1 infection requires the interaction with cellular co-factors at virtually all the steps of the viral replication cycle [1]. Viral entry depends on fusion of viral and cellular membranes through successive interactions with CD4 receptor combined with CXC chemokine receptor type 4 (CXCR4) or CC chemokine receptor type 5 (CCR5) [2]. Once the core is released into the cytosol, the reverse transcriptase converts the viral RNA genome into a double-stranded copy DNA (cDNA) and the capsid (CA) uncoating process is initiated. HIV-1 cDNA gains access to the nucleus through the cellular nuclear transport machinery located at the nuclear pore, in the form of a pre-integration complex (PIC). These PICs consist of viral cDNA and other HIV-1 components like integrase (IN), matrix, nucleocapsid, CA and viral protein R (Vpr), as well as various host proteins, such as the high mobility group protein B1 (HMGB1), barrier to autointegration factor 1 (BAF1), lamina-associated polypeptide 2α (LAP2α) and lens-epithelium derived growth factor (LEDGF/p75) [3][4][5][6][7]. Several cellular import factors, including importin-7, importin-α3 and Transportin 3 (TNPO3, also called TRN-SR2) have also been involved in HIV-1 nuclear import [8]. Apart from its implication in nuclear import of the viral PIC, it has been confirmed that N-terminal end of TNPO3 protein act as a direct binding partner of HIV-1 IN [9]. Interaction with the viral CA has also been documented [10,11] and nearly 30 CA-mutants able to modify HIV-1 dependence on TNPO3 have been described [12].
TNPO3 is a member of the karyopherin β superfamily of proteins [13,14] that imports into the nucleus mostly serine/arginine (SR)-rich proteins. Within these proteins are essential pre-mRNA splicing factors such as serine/arginine-rich splicing factor 1 (SRSF1), SR-rich splicing factor 2 (SRSF2, also known as SC35) and cleavage and polyadenylation-specific factor 6 (CPSF6) [15,16]. The interaction between HIV-1 CA and CPSF6 impedes interferon (IFN)mediated innate responses, allowing HIV-1 to escape from immune sensing and favouring infection. In fact, HIV-1 virions carrying CA mutation N74D that cannot interact with CPSF6 trigger innate sensors that induce an antiviral state against HIV-1 infection in macrophages [17]. Moreover, as HIV-1 is highly dependent on the cellular splicing machinery [18], modifications in TNPO3-mediated nuclear import may indirectly affect HIV-1 replication through changes in the post-transcriptional mRNA maturation [19][20][21][22]. TNPO3 has been identified as a HIV-1 co-factor in two independent genome-wide siRNA screens [23,24] and as a specific binding partner of HIV-1 IN in a yeast two-hybrid screen [25]. These results support the idea that TNPO3 may be essential for HIV-1 life cycle along with other fundamental proteins such as CPSF6. However, its precise role during HIV-1 nuclear import and viral integration is not fully understood [26].
LGMDs comprise a group of genetically heterogeneous disorders characterized by a progressive and predominantly proximal muscle weakness with histological signs of muscle degeneration and regeneration [27]. In 2001, a novel form of LGMD classified as LGMD1F was reported, affecting 32 individuals in one large Spanish kindred spanning six generations [28]. The genetic defect of this autosomal recessive disease was identified as a single adenosine nucleotide deletion in TAG stop codon of one allele of TNPO3 gene, common to both protein isoforms encoded by this gene. As a result, the cells from these patients may synthesize both TNPO3_wt and TNPO3_mut proteins forms, being TNO3P-mut an extended form of TNPO3 with fifteen additional amino acids in the C-terminal end. Because the cargo-binding domain of TNPO3 resides in this part of the molecule [29], this function might be altered in the mutated protein [30,31]. Being TNPO3 a co-factor of HIV-1 replication [23][24][25], the susceptibility to HIV-1 infection of peripheral blood mononuclear cells (PBMCs) isolated from LGMD1F patients was analyzed. Our data revealed that the mutation of TNPO3 present in patients with LGMD1F protected PBMCs from HIV-1 infection. Therefore, this is the second genetic defect described so far after CCR5-Δ32 deletion [32,33] that is able to confer resistance to HIV-1 infection.

Clinical characteristics of LGMD1F patients
Twenty-three patients with LGMD1F were recruited for this study. All these patients belong to a Spanish/Italian family that shares a common old ancestor born in south-eastern Spain [28], specifically to generations III, IV and V in the family tree ( Fig 1A). These patients have been closely followed up at the University Hospital La Fe (Valencia, Spain) and show a wide variety of clinical features (Table 1). Most patients included in this study presented onset symptoms such as difficulties in climbing stairs, rising from sitting, running or fatigue. Thirteen patients showed scapula-humeral and pelvic-femoral weakness and eight of them also presented hand and leg weakness and/or atrophy. The rest of patients showed pelvic-femoral weakness, hand atrophy and leg weakness. Two patients remained asymptomatic when this study was performed. Only three patients presented with grades > 6 in the Vignos score, which is given to individuals who need a long leg brace for walking or standing. Eight patients were graded as 3-4 in the Brooke score and were unable to elevate their shoulders [34,35]. Average levels of creatine kinase (CK) were 3.2-fold higher than the normal range. This human disease is caused by a deletion in the long arm of chromosome 7 (7q32.1) which compromises the TNPO3 gene. LGMD1F patients show a heterozygous single nucleotide deletion (c.2771del) in exon 23 that generates a 15 amino acid extension of the C-terminus of the protein (Fig 1B).

Allelic discrimination of wt and mut TNPO3 gene
The expression pattern of wt and mut variants of TNPO3 was analysed by RT-qPCR in PBMCs isolated from all LGMD1F patients and compared to twenty-seven healthy donors (labelled as CT). All patients revealed co-dominant expression of each allele (Fig 2A and 2B), in comparison with dominant expression of TNPO3_wt gene in healthy individuals. In order to know whether the longer protein encoded by TNPO3_mut allele was co-expressed with TNPO3_wt allele, protein extracts from four patients and two healthy controls were analyzed by immunoblotting using an antibody against TNPO3 that recognized both forms of the protein. Similar levels of TNPO3_wt and TNPO3_mut isoforms were observed in LGMD1F patients, whereas only one band corresponding to TNPO3_wt was observed in healthy controls ( Fig 2C).

Expression of surface and activation markers
The expression levels of the HIV-1 receptor CD4 and the co-receptors CCR5 and CXCR4 were analyzed by flow cytometry in order to exclude an expression defect in cells from LGMD1F patients. No significant difference with healthy controls was found (S1A Fig). PBMCs from LGMD1F patients and controls were then activated with antiCD3, CD28 and IL2 for 48 hours. The expression of activation markers CD25 and HLA-DR was also analyzed but no significant difference was observed (S1B Fig), thus excluding an activation defect influencing cellular susceptibility to HIV-1 infection. LGMD1F patients display a mutation located on the long arm (q) of human chromosome 7 (7q32.1). This mutation is a heterozygous single nucleotide deletion (c.2771del) located at exon 23 of TNPO3 gene, which encodes a 15 amino acid extension of the C-terminus of this protein.

Localization of CPSF6 protein in LGMD1F patients
TNPO3 is involved in nuclear import of splicing factors such as ASF/SF2, SC35 and CPSF6 [29]. Because it has been shown that TNPO3 knockdown induces the accumulation of CPSF6 -a predominantly nuclear protein-in the cytoplasm [12] we analyzed if the mutant form of TNPO3 from LGMD1F patients resulted in relocalization of CPSF6, a predominantly nuclear protein, to the cytosol. We observed relatively equal levels of CPSF6 protein expression in the nucleus, but PBMCs from LGMD1F patients showed higher level of cytoplasmatic CPSF6 (S2 Fig).

Resistance to HIV-1 infection in CD4+ T cells from patients with LGMD1F
In order to evaluate the susceptibility to HIV-1 infection, we analyzed the kinetics of viral replication in activated PBMCs isolated from seven LGMD1F patients and seven healthy individuals. PBMCs were infected ex vivo by spinoculation with NL4.3-Renilla and NL4.3_N74D-Renilla, a CA mutant in which the nuclear import is independent of TNPO3 [17]. The production of Renilla was measured in cell lysates several days post-infection as relative light units (RLUs). The values were normalized with the total protein taking into account the cell viability during infection. There were no significant differences in cell survival after infection with either virus (S3A and S3B Fig, respectively). Low HIV-1 replication was observed in PBMCs from LGMD1F patients 3-7 days after infection with NL4.3-Renilla ( Fig 3A) but no significant difference in the replication of NL4.3_N74D-Renilla virus was observed ( Fig 3B).
To confirm these results, activated PBMCs isolated from twenty-two LGMD1F patients and twenty-seven healthy individuals were infected ex vivo with NL4.3-Renilla strain by spinoculation. The production of Renilla was measured in cell lysates 5 days post-infection and results were normalized with total protein and viability. Average virus replication in all LGMD1F patients was reduced 18-fold compared with controls ( ���� p<0.0001) (Fig 3C).
In order to determine whether the impairment in HIV-1 infection was due to the presence of TNPO3_mut in LGMD1F patients and not to other potential restrictive activity, PBMCs isolated from seven LGMD1F patients and seven healthy donors were activated and infected The mutation of Transportin 3 gene that causes LGMD1F induces protection against HIV-1 infection  Fig 3D). In contrast to the results obtained using a wt HIV-1 clone, no reduction in infectivity was observed between LGMD1F and control patients. On the contrary, an increase in HIV-1 replication was observed in six out of seven patients in comparison with controls (p<0.001).

Dissection of HIV-1 replication cycle in PBMCs from LGMD1F patients
PBMCs isolated from twenty patients and twenty-six healthy controls were activated with antiCD3/CD28 and IL-2 for 48 hours and then infected with NL4.3-Renilla by spinoculation. At 5 hours post-infection, DNA was extracted and synthesis of HIV-1 strong stop (R/U5) and full-length (R/ATG-gag) reverse transcriptase products, which represent early and late reverse transcriptase transcripts, respectively, were quantified by qPCR ( Fig 4A). No significant difference in the efficiency of reverse transcription in PBMCs from LGMD1F patients and healthy controls was detected.
In order to monitor the nuclear import of viral DNA, the accumulation of circular DNA intermediates was determined by measuring episomal 2-LTRs by ultrasensitive digital PCR 24 hours after infection. No significant difference was detected in the number of episomal 2-LTR circles between PBMCs from LGMD1F patients and healthy controls (Fig 4B).
To assess viral integration, infected cells were incubated for 5 days, DNA was extracted and proviral copies were quantified by qPCR. Proviral integration was on average 16-fold lower in PBMCs from LGMD1F patients than in controls ( ���� p<0.0001) (Fig 4C). These data indicate an integration defect in cells carrying TNPO3 mutation.

CD4+ T cells isolated from patients with
LGMD1F and healthy controls were infected with fluorescently labeled particles (HIV-IN-eGFP) [7,36]. The presence of IN-eGFP in PICs allows quantitative analysis of the number of PICs and their intracellular location by confocal microscopy. When HIV-1 infected CD4+ T cells were fixed 10 hours post-infection, no significant differences were detected between both groups, neither in the number of uninfected cells ( Fig  5A), nor in the number of cytoplasmic and nuclear PICs per cell (Fig 5B and 5C, respectively). However, after 24 hours of infection, the number of cells without PICs was 1.7-fold higher in CD4+ T cells from patients with LGMD1F than in healthy controls (p<0.001) (Fig 5A). The number of cytosolic PICs was 3.1-fold reduced in CD4+ T cells from patients with LGMD1F compared to the healthy controls ( Fig 5B) (p<0.01). This was not due to higher PIC translocation to the nucleus, as there was no significant difference in the number of nuclear PICs ( Fig  5C and 5D).

Validation of TNPO3_mut as host factor refractory to HIV replication
HIV-1 infectivity was examined in HeLaP4 cell lines stably expressing TNPO3_wt or TNPO3_mut form. These cell lines were validated by western blot, immunocytochemistry and RT-PCR (S4 Fig). Next, we compared the luciferase signal of HIV-fLuc VSV-G in a cell line containing TNPO3_wt which was transduced with an empty vector (control shRNA + empty standard error of the mean (SEM). The RLUs mean value in NL4. 3   vector). As described before, depletion of TNPO3 (TNPO3 shRNA + empty vector), resulted in a three-fold drop in luciferase activity, reflecting a decreased viral infectivity (Fig 6) (p<0.01) [7,25]. Back-complementation with TNPO3_wt in stable knock-down cells (TNPO3shRNA + TNPO3_wt) restored the luciferase activity level to that of cell lines containing endogenous TNPO3. Back-complementation with the mutant form of the protein (TNPO3 shRNA+TNPO3_mut) resulted in lower recovery of HIV-1 infection. These experiments supported that TNPO3_mut was not able to rescue HIV-1 replication in these cell lines thus validating its role as a defective host factor impairing HIV infection.

Discussion
HIV-1 infection remains incurable despite efficient antiretroviral treatments that tackle HIV-1 enzymes and proteins. One potential strategy to develop new therapeutic targets can be based on the study of the interaction between viral proteins and their cellular cofactors. In this regard, TNPO3 and other importins have been previously described as essential cellular proteins for HIV-1 infection [23][24][25]37]. However the exact mechanism of action of TNPO3 still remains a matter of controversy [38]. Some studies suggest that TNPO3 participates in the nuclear import of PICs [9,25,39,40] whereas other authors propose that TNPO3 promotes HIV-1 infection though the interaction with HIV-1 CA [41][42][43] or indirectly through the interaction of CPSF6 with HIV-1 CA [12]. Besides an indirect role for TNPO3 in viral integration through its interaction with CA and/or the IN and their respective cellular partners such as CPSF6 or LEDGF/p75 [38] has been proposed. These cellular host factors may affect the nuclear landscape of HIV-1 infection [44] by targeting the viral genome to silent or actively The mutation of Transportin 3 gene that causes LGMD1F induces protection against HIV-1 infection transcribed chromatin [45] opening a new perspective in the mechanisms of HIV-1 latency and reactivation.
Of note, nuclear transport has not been widely studied as a potential target to HIV-1 infection and only recently new hits impeding TNPO3-IN interactions blocking HIV-1 nuclear transport have been described [8]. Besides this role in HIV-1 infection, TNPO3 is also linked to a rare muscular dystrophy termed LGMD1F. The genetic cause for LGMD1F was found to be an adenosine deletion in the stop codon of TNPO3 gene, which leads to the addition of 15 amino acids to the C-terminal end [30,31]. The effect of this elongation for the functionality of TNPO3 protein is still unknown but previous data indicate that TNPO3_mut shows a perinuclear distribution whereas TNPO3_wt predominantly localizes inside the nucleus. This suggests that TNPO3 mutation could affect the subcellular localization of the protein [31]. Because the cargo-binding domain of TNPO3 resides in the C-terminal end, the additional 15 amino acids may also alter its cargo binding properties and subsequently influence alternative splicing [30,31]. Apart from these preliminary results, the mechanism by which TNPO3 mutation affects its function and causes LGMD1F remains undetermined. Interestingly, other rare muscular diseases known as laminopathies are related to changes in the nuclear lamin architecture, a fibrous structure located below the nuclear membrane which functions include among others the regulation of the nuclear transport [46,47]. Besides, other muscle diseases like myotonic dystrophies are due to mutations in splicing factors. For these reasons the term "splicepathies" has been proposed to design these genetic diseases [48,49]. In summary, the biomedical interest of TNPO3 is first, as an essential cellular protein in the HIV-1 cycle and second through a specific mutation, as the genetic cause of the ultra-rare disease LGMD1F. In this context, we propose that the understanding of HIV-1 biology in lymphocytes from LGMD1F patients represents a tool to get a better insight into the mechanisms of nuclear transport and to understand the pathogenic mechanisms leading to muscle disease such as muscular dystrophies that share common physiopathology pathways.
Our results demonstrate that resistance to HIV-1 infection observed in CD4 lymphocytes from patients with LGMD1F is directly related with the mutation in TNPO3 gene because the infectivity was not affected when we infected with an HIV-1 clone carrying a CA-mutation (N74D), in which the nuclear transport is independent of TNPO3. As previously described [17], the efficacy of infection is lower with N74D clone as compared to a wt virus, which explains lower luciferase levels. Interestingly, infection with N74D virus was not equal in the PBMCs of all LGMD1F patients as in some of them the infection with the mutant virus was higher than in controls, suggesting that PBMCs from some patients with LGMD1F can develop compensatory mechanism to overcome the deficiency in TNPO3-mediated transport. By unknown reasons, age of onset and severity of clinical symptoms are highly variable in LGMD1F patients, despite all of them carry the same mutation, pointing to the presence of alternative mechanisms that compensate the deficit in TNPO3 function.
Due to the central role of TNPO3 in HIV-1 nuclear import, we hypothesized that HIV-1 infection might be impaired in PBMCs from LGMD1F patients.
LGMD1F is an autosomal recessive disease and accordingly, both TNPO3_wt and TNPO3_mut form are co-expressed in similar quantity. The TNPO3_mut allele was expressed and translated as a 15 amino acids longer protein than TNPO3_wt [31]. However despite the presence of 50% of the normal protein, high resistance to HIV-1 infection ex vivo was observed in these cells, with a reduction of more than 18-fold on average in the production of viral proteins. These results suggest that in patients with LGMD1F TNPO3_mut was interfering with the normal function of TNPO3_wt. It has been proposed that TNPO3 multimerizes in order to carry out its nuclear import functions [29,40]. If TNPO3_mut cannot multimerize or interferes with regular multimerization of TNPO3_wt, this could explain the severe restriction of viral infection observed in LGMD1F cells despite the co-dominant expression of both TNPO3_wt and TNPO3_mut.
The cytosolic location of CPSF6 in lymphocytes from patients with LGMD1F confirms the functional impact of the mutated protein "in vivo" for the first time and correlates with previous work showing that TNPO3 knockdown causes CPSF6 to accumulate in the cytoplasm [12]. Relocalization of CPSF6 to the cytosol could decrease its availability to bind to HIV-1 CA and contribute to the increased resistance to infection observed in LGMD1F patients.
To confirm these data we generated HeLaP4 cell lines with stable knock-down of endogenous TNPO3 that were trans-complemented with either TNPO3_wt or TNPO3_mut form associated to LGMD1F. As described before, depletion of endogenous TNPO3 resulted in decreased viral infectivity that was recovered with the back-complementation of TNPO3_wt. However, expression of the TNPO3_mut form of the protein resulted in lower recovery of HIV-1 infection, which supports that the TNPO3_mut form found in LGMD1F patients restricts HIV-1 infection.
In order to define the step at which TNPO3_mut impaired HIV-1 replication, the viral cell cycle was analyzed in PBMCs from LGMD1F patients in comparison with healthy controls. There were no significant differences between the cells of patients and controls in the formation of early and late reverse transcripts 5 hours after the infection. It has been previously described that TNPO3 depletion leads to a reduction in nuclear 2-LTR circles during HIV-1 infection [7,25,50], although other studies did not corroborate these results [10,11,51]. This discrepancy has been attributed to the different primer sets used to detect the intranuclear viral DNA products [12]. In the present study we could not find differences between controls and LGMD1F patients at the level of the formation of episomal 2-LTR circles 24 hours postinfection ( Fig 4B). Accordingly, a direct role for TNPO3_mut on the entry of viral DNA into the nucleus cannot be concluded. However, integration was decreased more than 16-fold in PBMCs from LGMD1F patients in comparison to controls 5 days after infection suggesting an impairment of HIV-1 integration due to the TNPO3 defect leading to deep impact on viral replication.
When using fluorescently labeled viruses (HIV-IN-eGFP), no differences were detected in the cellular distribution at 10 hours after infection. However at 24 hours post-infection the number of infected cells, as defined by IN-eGFP detection, was significantly decreased in patients with LGMD1F. Unexpectedly, only cytoplasmic complexes were reduced, suggesting a decreased stability of the complexes that was not a consequence of higher translocation to the nucleus, as there was no increase in the number of nuclear PICs. The change in subcellular distribution of CPSF6 could not explain this observation since an increase in CA stability due to cytosolic CPSF6 location after silencing TNPO3 has been proposed [12]. One potential hypothesis explaining this sharp decrease in cytosolic PIC numbers could be related to the induction of innate immune responses leading to CA degradation. It has been described that in macrophages HIV-1 evades innate immune recognition in macrophages through specific recruitment of cellular factors to the CA, including TNPO3 and CPSF6 to the CA [16]. A deficient binding of TNPO3 and CPSF6 could increase CA sensing and induction of IFN-mediated mechanisms in lymphocytes of LGMD1F patients.
Finally, because integration targeting of the HIV-1 genome to silent or actively transcribed chromatin is linked to nuclear import mechanisms and cellular factors as LEGDF/p75 and CPSF6 we cannot rule out that entry through TNPO3-independent mechanisms in patients with LGMD1F leads to different selection of chromosomal integration sites [52].
Our data confirm for the first time "in vitro" that adequate TNPO3 function is essential for HIV-1 replication infection, but they also provide insights into the role of TNPO3 in LGMD1F. It should be noted that although LGMD1F patients show progressive muscle weakness as the disease advances, they do not show any type of immunodeficiency or higher susceptibility to infectious diseases. Therefore, the presence of TNPO3_mut is not affecting all types of cells in a similar way, being the muscle cells clearly more affected. Moreover, PBMCs from all patients showed similar resistance to HIV-1 infection, while there is a wide variability in the muscular clinical symptoms that affects these patients, although all of them share the same TNPO3 mutation. One possibility that merits further investigation is that TNPO3_mut could be affecting the nuclear import of essential factors for alternative splicing such as SC35 and CPSF6 [15,16]. It has been described that skeletal muscle is the tissue with the highest number of differentially expressed alternative exons [53,54], including isoforms of myogenic transcription factors, metabolic enzymes and myofibrillar proteins [55].
LGMD covers a group of genetically determined disorders that varies depending on factors such as age of onset, rate of disease progression, distribution of muscle weakness and genetic causes. Such array of factors implies different steps of muscle development and neural and hormonal influences that can affect differently the highly complex pattern of muscle-specific transcripts processed by alternative splicing [56]. Therefore, the analysis of muscle cells from LGMD1F patients could determine whether this disease might be caused by defects in TNPO3-mediated import of splicing factors involved in the alternative splicing of essential proteins for muscular contraction.
In conclusion, TNPO3_mut protein expressed in LGMD1F cells is the second genetic defect leading to strong HIV-1 restriction in humans. Importantly, the first genetic defect shown to produce HIV-1 restriction, the CCR5 delta32 deletion, blocks entry of R5-tropic but not X4tropic strains. TNPO3 mutation described here acts at a post-entry step in the virus life cycle, and may therefore be independent of viral tropism. These findings increase our understanding of the role of TNPO3 in HIV-1 infection, and support further characterization of LGMD1F as a splicing disease.

Patients and controls
Twenty three patients with diagnosed LGMD1F and twenty-seven healthy donors were recruited for this study. Muscle strength was graded using the Modified Medical Research Council (MMRC) scale. The upper and lower extremity functions were assessed using Brooke and Vignos scores, respectively [34,35]. LGMD1F patients were recruited at the Hospital de La Fe (Valencia, Spain) and healthy donors were recruited at the Centro Regional de Transfusión from the Complejo Hospitalario de Toledo (Toledo, Spain).

Ethics statement
All individuals gave informed written consent and this study was approved by the Institutional Ethical Committee Board of Hospital de La Fe (2016/0388) and Instituto de Salud Carlos III (Madrid, Spain; CEI PI 22_2017-v3).

Quantitative PCR for allelic discrimination
Total genomic DNA and RNA were isolated from PBMCs using DNA/RNA Mini Kit (Qiagen). SNP genotyping assay was designed for detecting simultaneously TNPO3_wt and TNPO3_mut forms using the following primers and probes: forward primer (5´-GCGAGAC TTCACCAGGTTGTT-3´); reverse primer (5´-CTGGGTGACAGGCACAGT-3´); TNPO3_wt probe: (TNPO3 deletion-V, VIC, 5´-CAGGAGTGTGAGCTATCGA-3´); and TNPO3_mut probe (TNPO3 deletion-M, FAM, 5´-AGGAGTGTGAGCATCGA-3´). cDNA was synthesized from 200 ng of RNA by using GoScript Reverse Transcription System (Promega), following manufacturers' instructions. RT cycling conditions were as follows: 5 min at 25˚C; 1h at 45˚C; and 15 min at 70˚C. SNP genotyping was also performed using 50 ng of genomic DNA and TaqMan Universal Mix (Applied Biosystems). Analyses were performed in triplicate per sample using StepOne Real-Time PCR system (Applied Byosistems) with standard cycling conditions. Results for the allelic discrimination were represented as ΔRn, being Rn the ratio between the fluorescent emission intensity of the reporter dye and the passive dye.

Immunoblotting assays
Whole protein extracts were obtained as described previously [57] and protein concentration was determined by Bradford method [58]. Ten micrograms of protein extracts were fractionated by SDS-PAGE and transferred onto Hybond-ECL nitrocellulose paper (GE Healthcare). Subsequently, the membranes were blocked and incubated with an anti-TNPO3 antibody (Abcam). Analysis was performed by chemiluminescence using a BioRad Geldoc 2000 (Bio-Rad Laboratories, Madrid, Spain).

HIV-1 infection
Infectious supernatants were obtained from calcium phosphate transfection of HEK293T cells (provided by the existing collection of the Instituto de Salud Carlos III, Madrid, Spain) with plasmid pNL4.3-Renilla, which contains the HIV-1 proviral clone pNL4.3 with the nef gene replaced by renilla luciferase gene. The pNL4.3_N74D-Renilla clone was generated introducing the N74D mutation in the nucleotide position 1405 of the previously described plasmid pNL4.3-Renilla [59]. PBMCs were isolated from blood samples by centrifugation using a Ficoll-Hypaque gradient (GE Healthcare) and then activated for 3 days with purified antihuman CD3 (clone OKT3), CD28 (clone CD28.2) (Biosciences, San Diego, CA) and 300 U/ml interleukin-2 (IL-2) (Chiron, Emeryville, CA). Activated cells were infected with 1 ng p24 of NL4.3-Renilla or NL4.3_N74D-Renilla per million of cells by spinoculation. Briefly, after 30 minutes of gently rotation at room temperature, cells were centrifuged at 600xg for 30min at 25˚C and extensively washed with PBS1X. Infected cells were cultured for 5 days with IL-2. Renilla activity (RLU) was quantified at different time points in the cell lysates with Renilla Luciferase Assay System (Promega) and measured with a Sirius luminometer (Berthold Detection Systems, Oak Ridge, TN). Data were normalized for protein concentration measured with the Bradford method [58] and cell viability was measured with the CellTiter-Glo Luminescent Cell Viability assay (Promega).

Quantification of early and late reverse transcription
At 5 hours after infection, DNA was extracted using QIAamp DNA Blood Mini Kit (Qiagen) and quantified with Nanodrop 2000C (Thermo Scientific). Strong stop DNA was quantified using primer pairs specific for R and U5 regions of the HIV-1 long terminal repeat (LTR), as described [60]. Serial dilutions of genomic DNA from 8E5 cell line, which contains a single integrated copy of HIV-1 [61], were used as a standard curve. The ccr5 gene was used as endogenous control. qPCR was performed in triplicate in StepOne Real-Time PCR system (Applied Biosystems) using standard cycling conditions.

Quantification of episomal 2-LTR circles by digital PCR
Genomic DNA was extracted from PBMCs 24 hours after infection with NL4.3-Renilla, using DNeasy Blood and Tissue kit (Qiagen), and quantified using the Nanodrop-1000 spectrophotometer (Nanodrop). The samples were measured by the QuantStudio 3D Digital PCR System (Life Technologies). The reaction mixture for Digital PCR (dPCR) is as follows: cDNA, QuantStudio 3D Digital PCR Master Mix v2, 300nM of C1_2LTR primer, 300nM C4R_71 primer, 250nM 2nr4nr_FAM probe, 0,5x CCR5_VIC probe and H 2 O for a final volume of 14.5 μl. The digital PCR reaction mix was loaded onto the QuantStudio 3D digital PCR chips, according to the manufacturer instructions. The thermal cycling and amplifications were performed in a ProFlex 2xFlat PCR system, according to the manufacturer protocol: 96˚C for 10 min, followed 39 cycles of 55˚C for 2 min and 98˚C for 30 sec, 55˚C for 2 min and a stabilization phase at 22˚C. The chips were transferred to a QuantStudio 3D digital PCR instrument for imaging and the final analysis of the files generated was carried out using the cloud-based.

Quantification of proviral integration by TaqMan qPCR
Total DNA from PBMCs infected for 5 days with NL4.3-Renilla was extracted using DNeasy Blood and Tissue kit (Qiagen) as described above and the integrated HIV-DNA was measured by nested Alu-HIV-LTR PCR [62][63][64] with modifications [65]. Briefly, 50 ng DNA were used for the first conventional PCR (C1000 Thermal Cycler, Bio-Rad) with 10x TaqMan Buffer, dNTPs, Alu-1 primer, Alu-2 primer, L-M667 primer and Platinum Taq DNA polymerase (Roche); 2x TaqMan Universal PCR Master (Promega), AA55M primer, Lambda T primer and MH603 probe for the second quantitative PCR (StepOne, Applied Biosystems). A standard curve of integrated HIV-DNA from 8E5 cell line using serial dilutions was prepared as reference and CCR5 was used as housekeeping gene.

Analysis of HIV-1 PIC
HIV-1 particles containing fluorescently labeled IN (HIV-IN-eGFP) were generated by Vprmediated trans-incorporation as described before [7,36,66]. For infection with HIV-IN-eGFP, PBMCs were previously incubated with CD3:CD28 bispecific monoclonal antibody (NIH AIDS Reagent Program, Division of AIDS, NIH from Drs. Johnson Wong and Galit Alter) during 5 days for CD4+ T cell enrichment. CD4+ T cells were infected with HIV-IN-eGFP by spinoculation. At 2 hours post-infection cells were washed and further incubated for 5 or 19 hours. Next, the cells were plated in poly-D-lysine chambers and allowed to adhere, to reach a total infection time of 10 h and 24 h, respectively. Cells were fixed for 15 min with 4% (v/v) paraformaldehyde and permeabilized during 5 min with 0.1% (v/v) Triton-X100. Nuclei were immunostained with lamin A/C antibody (Santa Cruz Biotechnology) and secondary anti-Mouse IgG (H+L) Alexa Fluor 555 conjugate (ThermoFisher Scientific) diluted in blocking buffer (1% (w/v) BSA and 0.1% (v/v) Tween-20 in PBS). Imaging of the cells was performed using a laser scanning microscope (Fluoview FV1000, Olympus, Tokyo, Japan). An in-house MatLab routine (MatWorks) was used to determine the localization and number of IN-eGFP complexes [7]. In short, IN-eGFP complexes and the nuclear lamin were determined automatically using an intensity threshold based on the triangle algorithm. Based on the nuclear lamin staining, IN-eGFP complexes were divided into cytoplasmic or nuclear compartments and the percentage of nuclear IN-eGFP complexes was calculated. Typically, data were collected from 90 cells.
HeLaP4 cells, a kind gift from Dr. P. Charneau (Institut Pasteur, France), and HeLaP4 cells stably depleted of TNPO3 [7,25] were back-complemented with TNPO3_wt or TNPO3_mut through stable transduction with lentiviral vectors. Briefly, 3x10 4 HeLaP4 cells were plated in a 96-well plate the day before transduction. The next day, cells were transduced with a dilution series of vector containing cassettes coding for TNPO3_wt or TNPO3_mut. As a control, cells were transduced with an empty cassette vector (empty vector). Transduced cells were selected with 160 μg/ml hygromycin.
To determine the effect of the TNPO3_mut on the viral infectivity in cell lines, 1.5x10 4 HeLaP4 cells were seeded per well in a 96-well plate. The next day, the cells were infected with a three-fold dilution of a single-round HIV-fLuc VSV-G [68,69]. At 72 h post-infection, cells were lysed in buffer (50 mM Tris, 200 mM NaCl, 0.2% (v/v) NP40 and 5% (v/v) glycerol) and analysed for firefly luciferase activity (ONE-Glo Promega GMBH, Mannheim, Germany). Chemiluminescence was measured with a Glomax luminometer (Promega). Readouts were normalised for protein content as determined by a BCA protein assay. Data are represented as relative infectivity compared to a cell line expressing endogenous TNPO3_wt (control shRNA) and are means of at least two independent experiments. A Kruskal-Wallis test was used to evaluate statistical significance. Error bars represent the standard deviation.

Statistical analysis
Statistical analysis was performed using GraphPad Prism 5.0 Software (GraphPad). Comparisons between LGMD1F patients and healthy individuals were made with one-way analysis of variance (ANOVA) using Tukey´s Multiple Comparison Test to describe statistical differences among groups. The number of nuclear and cytoplasmic PIC was plotted in a scatter plot and a Mann-Whitney test was used to determine statistical significance. Differences were considered statistically significant when �� p < 0.01, ��� p < 0.001 and ���� p < 0.0001. PBMCs of LGMD1F patients and controls were activated for 3 days with purified anti-CD3, anti-CD28 and IL-2. Intracellular expression was confirmed by immunofluorescence using a monoclonal antibody against CPSF6 and a secondary antibody conjugated to Alexa 488 (green). DAPI was used for nuclear staining (blue). The results are representative of those observed in four independent patients and four controls. Bars indicate 5 μm.