Defining the molecular mechanisms of HIV-1 Tat secretion: PtdIns(4,5)P₂ at the epicenter

Abstract

The human immunodeficiency virus type 1 (HIV-1) transactivator of transcription (Tat) protein functions both intracellularly and extracellularly. Intracellularly the main function is to enhance transcription of the viral promoter. However, this process only requires a small amount of intracellular Tat. The majority of Tat is secreted through an unconventional mechanism by binding to phosphatidylinositol-4,5-bisphosphate (PtdIns(4,5)P₂), a phospholipid in the inner leaflet of the plasma membrane that is required for secretion. This interaction is mediated by the basic domain of Tat (residues 48–57) and a conserved tryptophan (residue 11). After binding to PtdIns(4,5)P₂, Tat secretion diverges into multiple pathways, which we categorized as: oligomerization-mediated pore formation, spontaneous translocation, and incorporation into exosomes. Extracellular Tat has been shown to be neurotoxic and toxic to other cells of the CNS and periphery, able to recruit immune cells to the CNS and cerebrospinal fluid, and alter the gene expression and morphology of uninfected cells. The effects of extracellular Tat have been examined in HIV-1-associated neurocognitive disorders (HAND), however, only a small number of studies have focused on the mechanisms underlying Tat secretion. In this review, the molecular mechanisms of Tat secretion will be examined in a variety of biologically relevant cell types.

Graphical Abstract

After the initial binding of residue 11 (light grey), a conserved tryptophan, inserts into the plasma membrane and stabilizes the interaction with PtdIns(4,5)P₂. The specific region of PtdIns(4,5)P₂ involved in this interaction is unknown. 3. After the interaction occurs, Tat is translocated extracellularly, but the exact process has yet to be elucidated. C) The process by which Tat is incorporated into exosomes is poorly understood. 1. Tat may bind to PtdIns(4,5)P₂ at the exosomal membrane, such as during the formation of multiple vesicular bodies. 2. During invagination of the exosome, Tat may interact with HIV-1 TAR (red) and be incorporated at a higher rate. 3. During invagination of the exosome, Tat may interact with host proteins (yellow) and be incorporated at a higher rate. (INT, Intracellular; EXT, Extracellular; EXO, Exosome)

Functional properties of Tat

Human immunodeficiency virus type 1 (HIV-1) encodes for the transactivator of transcription (Tat) protein. Tat is a small, basic, non-structured protein (86 or 101 amino acid residues in length) that is responsible for enhancing the transcription of HIV-1 by recruiting the host positive transcription elongation factor b (P-TEFb) to the RNA stem-loop transactivation response (TAR) element, encoded by the viral long terminal repeat (LTR) ^1,2. This, however, is not the only function of Tat, and many of its biological activities are related to the fact that the predominant proportion of Tat is transported extracellularly ³. Extracellular Tat has been shown to be neurotoxic and recruit additional immune cells to the central nervous system (CNS), causing low levels of chronic inflammation by activating bystander cells, such as endothelial cells, and causing the release of pro-inflammatory cytokines from immune cells, such as IL-1β from monocytes and macrophages ^4–8. Additionally, Tat produced in the periphery can also traffic to the CNS and cross the blood-brain barrier ⁹. These, among other factors (including other HIV proteins such as gp120, Vpr, and Nef), contribute to the development of HIV-1-associated neurocognitive disorders (HAND)¹⁰. Furthermore, the mortality associated with HIV-1 has decreased since the development of antiretroviral therapy (ART), but there has been an increase in the prevalence of HAND and even patients that are adherent to ART are still at risk ¹¹. While a comprehensive mechanism of HAND has yet to be elucidated, the secretion of viral proteins, such as gp120, Vpr, and Tat, have been demonstrated to alter behavior in vivo ¹² and has been suggested to be crucial for productive viral infectivity ¹³. ART does not block the secretion of Tat and there is currently no approved therapy targeting Tat ¹⁴. Additionally, it is unknown which cells are responsible for releasing Tat and the process by which Tat is secreted within the CNS remains poorly understood.

The secretion of Tat is characterized as unconventional, meaning it does not traffic through the endoplasmic reticulum (ER) or Golgi apparatus. The first report that examined the source of extracellular Tat originated in 1990, when it was demonstrated that Tat was released from infected H9 and COS-1 cells in the absence of lysis ¹⁵. Disruption of microtubules, actin cytoskeleton, Na⁺/K⁺-ATPase activity, endosomal-formation, and Golgi apparatus integrity in CD4+ T cells did not significantly alter the secretion of Tat ⁵. Further investigations have led to the current understanding of the unconventional secretion of Tat, which is initiated by binding to phosphatidylinositol-4,5-bisphosphate (PtdIns(4,5)P₂), a phospholipid component of the inner leaflet of the plasma membrane³. The reported abundance of PtdIns(4,5)P₂ varies ^16–19, but is approximately 5% of the total phosphoinositides in a cell ²⁰, however, it is only approximately 1% of the phospholipids in the inner leaflet of the plasma membrane ²¹. PtdIns(4,5)P₂ can be synthesized by a few pathways, such as by phosphorylation of phosphatidylinositol 4-phosphate by phosphatidylinositol 4-phosphate 5-kinase (PIP5K) or de-phosphorylation of phosphatidylinositol (3,4,5)-trisphosphate (PtdIns(3,4,5)P₃) by phosphatidylinositol 3,4,5-trisphosphate 3-phosphatase ²². The major synthetic pathway responsible for production of PtdIns(4,5)P₂ is generally accepted to be mediated by PIP5K. PtdIns(4,5)P₂ is also responsible for a number of cellular functions including: regulation of cytoskeleton dynamics, endocytosis and exocytosis, and G protein-coupled receptor activation, where it is a substrate for phospholipase C and plays a critical role in the activation of protein kinase C and Ca²⁺ release ²³.

The secretion mechanism of HIV-1 Tat has been frequently compared to fibroblast growth factor 2 (FGF2), a host cytokine that also binds to PtdIns(4,5)P₂ and is translocated extracellularly ²⁴. Furthermore, the mechanism of FGF2 secretion has been well characterized ^25,26. The mechanism of FGF2 secretion is generally accepted to be mediated by oligomerization and formation of a pore in the plasma membrane ^27,28. Other host proteins, such as IL-1β and galectins, are also unconventionally secreted, however, they are less well characterized compared to FGF2 ^26,29. The interaction between FGF2 and caspase-1 has been shown to be required for efficient secretion of both IL-1β and FGF2 ³⁰. Additionally, galectins and FGF2 have been detected within exosomes ³¹. If overlap between the secretion of host proteins exist, it can be rationalized that similarities exist for viral proteins as well.

Proteins that travel through the classical secretion pathway contain a stretch of 5–10 hydrophobic amino acids near the N-terminus, which mediates the initial transfer through the ER/Golgi apparatus ³². In this pathway, the protein will undergo folding, glycosylation, and insertion into COPII-coated vesicles, before reaching the plasma membrane and being secreted from the cell. Circumventing the ER/Golgi apparatus is essential for certain proteins because it allows for them to avoid modifications that would limit their biologic function ^28,32. This was demonstrated by experimentally forcing FGF2 through the ER/Golgi apparatus, by synthesizing a fusion protein of FGF2 and fibroblast growth factor 4 (FGF4) ³³. FGF4 naturally traffics through the ER/Golgi apparatus. The N-terminus signal peptide portion, derived from FGF4, caused the entire fusion protein to traffic through the ER/Golgi apparatus, which was then post-translationally modified by O-linked chondroitin sulfates and was secreted. This modification prevented interactions with the cell surface, which resulted in an accumulation of the secreted fusion protein in the supernatant ^28,32,33. Tat could potentially prevent modifications that limit its biological activity by alteration of the corresponding acceptor sequences. This, however, is not likely to occur, as several modification sites reside within the basic domain, such as methylation by PRMT6 and acetylation by p300 and GCN5 ^34–37. Additionally, not all modifications are detrimental, acetylation of Tat via p300 has been demonstrated to increase transactivation of the HIV-1 LTR ³⁷. However, alteration of the residues within the basic domain would more likely impact the secretion and extracellular functions of Tat more significantly than limitations caused by post-translational modifications.

It has been proposed that HIV-1 Tat, which also lacks a signal sequence, follows the same unconventional secretion pathway for exocytosis that is employed by FGF2, however, a direct comparison of the interaction kinetics of either Tat or FGF2 to PtdIns(4,5)P₂ demonstrated dissimilar binding between the two proteins and the phospholipid, suggesting that the secretion mechanism may be analogous rather than homologous ³⁸. Based on this information, we would propose that while there has been considerable evidence reported indicating the formation of membrane pores by oligomerization of Tat, it has also been shown that alternative mechanisms contribute, at least in part, to the secretion of Tat and must be further characterized. After binding to PtdIns(4,5)P₂, Tat secretion can diverge into several pathways that we propose to be categorized as: oligomerization-mediated pore formation, spontaneous translocation, and incorporation into exosomes (Figure 1). In this review, each of these pathways will be examined in detail, the gaps in published knowledge will be addressed, and evidence will be presented for these mechanisms occurring in a variety of cells that contribute to the development of HAND.

Figure 1. An overview of the secretion mechanisms of HIV-1 Tat.

Tat can be secreted from cells by various mechanisms. Interaction with PtdIns(4,5)P₂ allows Tat to be anchored or inserted into the plasma membrane. It has been proposed that Tat can be spontaneously translocated, though the final step has yet to be characterized. It may, however, be a precursor to oligomerization-mediated pore formation, which requires an increase in the concentration of cytosolic Tat. This is the most well characterized mechanism of Tat secretion. Finally, it has been suggested in the literature that Tat can be incorporated into multiple vesicular bodies and be incorporated into exosomes and be released. However, there is limited evidence on this mechanism and it requires more investigation. (INTRA, Intracellular; EXT, Extracellular; MVB, multiple vesicular body)

Oligomerization of Tat mediates the formation of pores in the plasma membrane

PtdIns(4,5)P₂ is required for the efficient secretion of HIV-1 Tat and although Tat has been demonstrated to interact with other phospholipids, disruption of the interaction with PtdIns(4,5)P₂ has been shown to attenuate secretion ³. This was determined by utilizing flotation gradient and lipid sedimentation assays to examine the interaction of Tat with phosphoinositides present in the plasma membrane ³. Tat bound with similar specificity to PtdIns(3,4,5)P₃, however, the interaction occurred with a lower affinity. Isothermal titration calorimetry (ITC) of the interaction with Tat and PtdIns(3,4,5)P₃ resulted in no detectable heat production ³⁹. Tat did not bind to PtdIns(3,4,5)P₃ in the ITC system, further supporting the low affinity interaction between the lipid and protein. In a different study, neomycin, which has been shown to inhibit protein binding to PtdIns(4,5)P₂, decreased the secretion of Tat in a dose-dependent manner, demonstrating the specificity of Tat for PtdIns(4,5)P₂ ⁴⁰. Further analysis of this interaction involved alanine substitutions, mutagenesis, and deletion studies of Tat, which determined that the basic domain (residues 48–57; GRKKRRQRRR) and a conserved tryptophan (residue 11; W) were required for efficient binding of Tat to PtdIns(4,5)P₂, which agreed with previous reported studies ^41,42. Overall, single alanine substitutions within the basic domain did not significantly alter secretion, such as substitutions at residue 50 or 51 alone, however, combinatorial substitutions, such as replacing residues 49 to 51 with alanine, were capable of ablating secretion. Alteration of W11 also impacted the stability of the interaction between Tat and PtdIns(4,5)P₂. Replacement of W11 with a either a hydrophobic or hydrophilic residue prevented insertion of the protein into the plasma membrane, and thus attenuated Tat secretion ⁴³.

HIV-1 Tat is a highly basic and hydrophilic protein; therefore, it does not easily cross the hydrophobic plasma membrane ^44–46. The direct translocation of hydrophilic proteins like Tat across the plasma membrane may occur by formation of a water pore ^32,38. The water pore is an opening through both bilayer leaflets ⁴⁶. The rate and quantity of protein translocation across the pore has been demonstrated to be dependent on the number of basic residues, specifically the abundance of arginine residues ⁴⁷. The basic domain of Tat contains two lysine and six arginine residues, causing the protein to be strongly cationic. Additionally, the arginine-rich motif has been demonstrated to be necessary and sufficient for translocation across the plasma membrane ^44,47. Studies on Tat internalization have demonstrated that all the cationic residues are necessary for translocation across the plasma membrane in both Tat peptides and alanine-substituted Tat peptides ³⁹. This is in direct opposition with the secretion of Tat, which can proceed with single alanine substitutions within the basic domain. The mechanisms of the secretion and internalization of Tat differ, though there are similarities, such as the basic domain of Tat being crucial for translocation through the plasma membrane. Furthermore, arginine is a charged amino acid, while alanine is hydrophobic, and substitution of varying amounts of arginine residues all impacted the translocation rate of Tat across the membrane, suggesting that the six arginine residues may be necessary for pore formation. Additionally, the importance of arginine side chains in membrane kinetics is not unique to Tat and has been associated with the opening and closing of voltage sensitive ion channels, which are responsible for transmembrane potentials ⁴⁵.

The pore formation of Tat is initiated by recruitment of the viral protein to the plasma membrane and interaction with PtdIns(4,5)P_2, which acts as a platform for multiple proteins involved in exocytosis (Figure 2A) ^38,45,48. Tat binds to PtdIns(4,5)P₂ with high affinity and multiple Tat proteins are recruited to the same location, resulting in oligomerization on the cytoplasmic side of the plasma membrane ³⁸. Interaction with PtdIns(4,5)P₂ has been shown to be crucial for Tat oligomerization and when PtdIns(4,5)P₂ is not present Tat is not recruited, and is unable to form a pore ^38,48. At low concentrations of Tat in the cytoplasm, the arginine and lysine side chains begin to interact with the phosphate and carbonyl groups of the plasma membrane, specifically PtdIns(4,5)P₂. As the concentration of Tat increases, to approximately one protein per eight lipids, the attractive forces between the arginine residues and phosphates will intensify and cause a gradual thinning of the plasma membrane bilayer ^44–47. Once this occurs, the arginine and lysine residues penetrate the bilayer, which will solvate the charged side chains ^45,49. This causes the formation of a pore, allowing the secretion of free Tat molecules. As the viral protein is secreted, the cytoplasmic concentration of Tat decreases, causing a reduction in the cumulative attractive forces between the arginine residues and phosphates, resulting in closure of the pore. The pore is transient in nature, which additionally prevents leakage of other cellular proteins ^44,45,47,49.

Figure 2. Detailing the mechanisms of HIV-1 Tat secretion.

A) The formation of pores in the plasma membrane has been shown to be mediated by Tat oligomerization. 1. The arginine and lysine side chains of Tat, within the basic domain, interact with the phosphate and carbonyl groups on PtdIns(4,5)P₂. 2. As the concentration of Tat increases to one peptide per eight lipids, the plasma membrane gradually thins and, along with water, causes the formation of a pore. 3. As the viral protein is released, the cytoplasmic concentration of Tat decreases, resulting in the pore closing. B) The spontaneous translocation of Tat is mediated by interaction with PtdIns(4,5)P₂. 1. The basic domain (light blue) of Tat (dark blue), residues 48-57, binds to PtdIns(4,5)P₂ (orange). 2.

The secretion of FGF2 by formation of a membrane pore has been demonstrated to be reduced by pharmacological inhibition of Na/K-ATPase, a pump responsible for generating membrane potentials, using ouabain ²⁷. ATP1A1, the α1-chain of Na/K-ATPase, was responsible for recruiting FGF2 to the plasma membrane, allowing for efficient secretion ⁵⁰. It was recently demonstrated that Tat had a similar mechanism and by inhibiting the Na/K-ATPase using ouabain, the secretion of Tat was impaired ¹³. Secretion was not completely ablated, however, suggesting that pore formation is not the sole route for Tat secretion. The evidence in the literature supports the formation of a membrane pore as the predominant mechanism of FGF2 secretion and we would propose that this likely holds true for Tat, however, alternative mechanisms may contribute as well.

Secretion of Tat is mediated by interaction with PtdIns(4,5)P₂ – spontaneous translocation

The spontaneous translocation of Tat across the plasma membrane is dependent upon interactions with PtdIns(4,5)P₂ (Figure 2B). The current understanding is that translocation can be initiated when the basic domain of Tat binds to the polar head of PtdIns(4,5)P₂ ^3,5. This interaction is then stabilized when the conserved W11 inserts into the plasma membrane, though it is not known what Tat is specifically interacting with to facilitate the interaction with the membrane ^38,43. W11 may bind to a hydrophobic tail, but it is unclear whether this interaction occurs with the same molecule of PtdIns(4,5)P₂ or with the tail of another phospholipid. After these events occur, Tat is transported extracellularly, however, this process has yet to be fully elucidated. To understand the mechanism more comprehensively, the secretion kinetics of Tat have been examined by altering pH and temperature. Tat secretion was determined to be temperature-dependent, reducing the temperature from 37°C to 16°C prevented secretion, and occur in the absence of cell lysis, as demonstrated by the lack of luciferase release, which was cotransfected into Jurkat cells along with Tat expression vectors ^5,41. While investigating the insertion of Tat into endosomal membranes, it was demonstrated that the biologically relevant pH of 5.3 was able to induce a conformational change in Tat ⁴³. This conformational change dramatically increased Tat insertion into the membrane by exposure of W11. It is important to note that the lipid monolayer utilized in these experiments was a model of the endosomal membrane for internalization of Tat and did not include PtdIns(4,5)P₂, thus the results may differ if the lipid composition was modeled to the inner leaflet of the plasma membrane. Additionally, a molecular dynamic analysis of Tat suggested that residue 11 is not accessible without acidification ⁵¹. However, it has been verified that a reduction in pH is not required for insertion of Tat into the plasma membrane for efficient secretion, as it was shown that the mere presence of a lipid bilayer could alter the conformation of Tat at a neutral pH ³⁸. Additionally, it has been demonstrated that acidification was not required for Tat secretion, as Tat was observed to insert into a lipid monolayer at a pH of 7.0 in the presence of PtdIns(4,5)P₂ ^3,5. These distinctions, and others, between the intracellular and extracellular interaction of Tat and the plasma membrane demonstrates the complexity of this viral protein and factors that may contribute to spontaneous translocation. Furthermore, there is conflicting evidence on the order of the initiation for spontaneous translocation of Tat, which also applies to the other mechanisms. There are reports of W11 initiating the interaction with PtdIns(4,5)P₂, although the majority appear to accept the basic domain is responsible for initiation ^3,38,45. After the initial interaction with PtdIns(4,5)P₂, the process by which Tat becomes extracellular remains to be determined. In addition, the kinetics of this interaction, such as the rate at which this occurs and if specific concentrations of Tat are required for this mechanism still need to be elucidated. Furthermore, it has yet to be conclusively shown that this mechanism is separate from oligomerization-mediated pore formation. Based on the lack of evidence relevant to the connection between these mechanisms, we would propose that they should be characterized separately until demonstrated otherwise.

The interaction of Tat with PtdIns(4,5)P₂ does not always result in secretion, since other cellular pathways can also be altered. Interestingly, only Tat capable of binding to PtdIns(4,5)P₂ exerts an effect on the exocytosis and secretion of the viral protein from neuroendocrine cells ⁴⁸. Overexpression of PIP5K, which is the major enzyme synthesizing PtdIns(4,5)P₂, rescued, to some degree, the overexpression of Tat, suggesting that Tat may sequester PtdIns(4,5)P₂. Tat also inhibited the binding of annexin A2, which interacts with PtdIns(4,5)P₂, providing in vivo evidence for direct binding to PtdIns(4,5)P₂. The co-expression of Tat in PC12 cells with a drug-inducible type IV 5-phosphatase or the PtdIns(4,5)P₂ 5-phosphatase, synaptojanin, reduced the levels of plasma membrane PtdIns(4,5)P₂. Reduction of PtdIns(4,5)P₂ in both cases altered the plasma membrane distribution of Tat, which became largely cytosolic ⁴⁸. Furthermore, Tat expression had significant effects on the actin cytoskeleton, which plays a vital role in maintaining the morphology of the cell. Feedback loops between the actin cytoskeleton and the plasma membrane has been proposed to have further effects on the organization of the plasma membrane phosphatidylinositol lipids in the cell and may influence exosome formation.

Incorporation of Tat into exosomes

Exosome-mediated trafficking of Tat is still a novel concept with only a limited number of studies reported in the literature investigating the mechanism. Evidence for this process is limited to the detection of Tat in exosomes isolated from transfected U-373MG cells, an astrocytic cell line ⁵². The results in this report suggested evidence for higher HIV-1 activation with exosome-associated Tat when compared to Tat not associated with an exosome. Interestingly, the basic domain of Tat did not affect packaging into exosomes, however, this result has not been examined in other biologically relevant cellular phenotypes. Furthermore, residue 11 was not examined, allowing for speculation on the importance of this residue for incorporation into exosomes.

Exosomes are small vesicles (30–100nm) that are released from cells through endocytic mechanisms. The biogenesis of these vesicles commences with the formation of early endosomes following endocytosis. Early endosomes mature into late endosomes, or multi-vesicular bodies (MVBs), marked by the presence of intraluminal vesicles (ILV) in the lumen of the endosome ⁵³. Inward budding of endosome membrane forms ILVs and permits uptake of diverse cytosolic content into the vesicle and transmembrane, causing exosomes to have unique surface markers including: CD9+, CD63+, ALIX and TGS101 ^54–56. Tat incorporation and recruitment to exosomes can be rationalized by the interaction with PtdIns(4,5)P₂ during exosome biogenesis. PtdIns(4,5)P₂ is enriched in exosomes and Tat has been detected within exosomes secreted from transfected cells ⁵². In addition to PtdIns(4,5)P₂ binding, HIV-1 Tat may associate with exosomes through other mechanisms. DNA, RNA, and protein, such as Tat, can be incorporated during the sorting of various ILVs ^57–59. These ILVs can then either fuse with lysosomes for destruction or fuse with the plasma membrane to secrete exosomal vesicles that affect bystander cells ^52,60,61. Tat can potentially enter the exosome during ILV formation, a process which uptakes contents from the cytosol (Figure 2C). Additionally, Tat may associate with other cellular content, such as RNA, to promote incorporation into exosomes. If the exosomal membrane is enriched in proteins that bind HIV-1 Tat, these interactions could potentially lead to enrichment of Tat on the inner or outer exosomal membrane. An important function of Tat, as mentioned earlier, is its ability to bind to TAR and recruit elongation factors for proviral transcription. It has been shown that the core domain of Tat (residues 38–48) can bind to TAR RNA ^62–64. Evidence for TAR RNA elements found in exosomes has suggested another potential mechanism for HIV Tat incorporation into exosomes. Exosomes derived from HIV-1-infected cells or patients have been shown to contain TAR, which has been suggested to enhance HIV-1 replication in the recipient cell, via downregulation of apoptosis ⁶⁵. The incorporation of TAR into the exosomes potentially provides a binding site to tether Tat for subsequent release and transport.

Moreover, HIV-1 Tat has been shown to bind to noncoding RNAs, specifically microRNAs (miRNA) that are also found in exosomes. RNAs are a required element for Tat to bind and inhibit Dicer function during miRNA biogenesis, highlighting the importance of the interaction between Tat and RNA as a mediator of cellular function ⁶⁶. Within the basic domain of Tat, residue 51 has been shown to have functional roles in miRNA binding ⁶⁷. Additionally, Tat has been demonstrated to interact and inhibit the activity of over 300 miRNAs, with a unique subset of these miRNAs binding to Tat with high affinity ⁶⁸. Furthermore, the hairpin structure of HIV-1 TAR interacts with Dicer to yield TAR-derived miRNAs found in infected cells ⁶⁹. During an HIV-1 infection, the cellular profile of miRNA expression is altered ^70–74. These virus-associated miRNAs and virus-mediated miRNA changes may have unique interplay with Tat and influence loading into exosomes. Cells transfected with HIV-1 Tat secrete specific miRNAs, notably miR-132 and miR29b ^72,75. Furthermore, during the formation of ILVs when miRNAs, proteins, and other contents are packaged into exosomes, the ability of Tat to bind to miRNA could promote Tat-miRNA packaging into vesicles for release. Tat has also been shown to interact with the Y-box 1 protein, a critical factor involved in the sorting of non-coding RNA into exosomes ^76–78. This interaction may support the packaging of Tat into exosomes and further suggests that Tat may be able to alter the incorporation of miRNA and other non-coding RNA into exosomes.

Evidence of HIV-1 Tat incorporation into exosomes and delivery into surrounding cells offers new avenues for investigation, particularly in HAND-related neuropathogenesis. Several studies suggest that exosomes can facilitate horizontal transfer of cargo and stimulate intracellular signaling pathways ^58,79,80. Exosomes have been shown to be involved in the regulation of many processes, including neurodegenerative diseases, such as Alzheimer’s and Parkinson’s disease ⁸¹. While the entire mechanism remains unclear, there is emerging evidence for exosome-mediated neuropathogenesis in HAND, specifically in the packaging of Tat. Many of these released vesicles contain DNA, RNA, and proteins, which may also confer pro-neoplastic and inflammatory effects ^61,82,83. These extracellular vesicles carry the genetic and proteomic profile that mirror the cell it is secreted from ⁵⁹. These mechanisms highlight the unique, indirect modes of transferring aberrant factors in HIV-1 CNS infection, a process that still requires more specific diagnostics and targeted treatment. We rationalize that exosomal trafficking of DNA and protein, specifically Tat, from HIV-1-infected cells to healthy cells could potentially lead to chronic inflammation and subsequent worsening of HAND. It has been demonstrated that Nef, another HIV-1 protein, was secreted via exosomes, supporting the rationale that Tat may be incorporated in exosomes as well ⁸⁴. Furthermore, RNA contained within exosomes is stable, suggesting that the membrane of an exosome could protect the contents and potentially even increase the stability of HIV-1 Tat protein during transport ^59,85. While there is limited evidence for the secretion of Tat in exosomes, we would propose that this mechanism may occur. Previous work suggested that FGF2 may be enriched in extracellular vesicles, specifically exosomes ⁸⁶. However, quantification of FGF2 secretion via plasma membrane blebbing and release in exosomes derived from MVBs concluded that this was not the predominant mechanism for the efficient secretion of FGF2 ³¹. However, recent work has demonstrated that Tat was not present in the exosomes isolated from the media of transfected HEK293T cells ⁸⁷. The addition of a c-Myc nuclear localization signal to the C-terminus of Tat was required for detectable incorporation into exosomes. Furthermore, it was shown that Tat was not present in exosomes isolated from J1.1 or Jurkat cells ⁸⁸. However, as previously stated, there has been data published that has demonstrated the presence of Tat within exosomes derived from U373.MG cells. Given these observations, the exact mechanisms that are operative to facilitate Tat secretion may be dependent on the type of cell involved in the secretion of Tat. To answer this question conclusively, additional secretion studies will be required to clearly demonstrate the exact mechanisms that may be involved in the secretion of Tat across a number of pathogenically relevant cellular phenotypes that may ultimately contribute to the development and severity of HAND.

Conclusion

PtdIns(4,5)P₂ is relatively abundant in the mammalian plasma membrane, thus Tat secretion may occur in a variety of cell types. The secretion of Tat appears to be dependent on the intracellular concentration of Tat, thus restricting these mechanisms to cells that have active viral transcription. The investigations discussed in this review were predominantly performed in primary CD4+ T cells and T cell lines. CD4+ T cells are one of the main targets for HIV-1 infection and are the predominant HIV-1-susceptible cell in the periphery ⁸⁹. Monocytes and macrophages are also susceptible and permissive to HIV-1 ⁹⁰. Additionally, studies suggest that astrocytes are susceptible to HIV-1 infection in the CNS and although histology of patient samples have detected Tat within astrocytes, it is still unclear whether infection of these cells occurs frequently enough to alter the microenvironment of the CNS ^91,92. These cell types contribute to the development of HAND and there is limited literature on Tat secretion alterations that are cell-specific. While studies that have begun to define these three secretion mechanisms of Tat (Table 1), there is still additional research required to completely elucidate them. After the initial interaction of Tat with PtdIns(4,5)P₂ there are many questions that remain to be addressed. What causes the release of Tat from binding to PtdIns(4,5)P₂ and allows it to be spontaneously translocated? Do these mechanisms occur at different rates in different cells? Do different cells have a preferential mechanism? Do these mechanisms occur simultaneously? Are there additional mechanisms associated with Tat secretion that have not yet been elucidated, such as incorporation into microvesicles? Is a single molecule of Tat able to be secreted and then interact with the plasma membrane of the same cell to be re-internalized? Does the full-length Tat protein (101 residues) exhibit the same secretion kinetics as the truncated protein (86 residues) used to examine and elucidate these processes? These are but a few of many unanswered questions.

Table 1. Evidence for the specific mechanisms of HIV-1 Tat secretion by a variety of model systems.

Enzyme-linked immunosorbent assay (ELISA); Fourier-transform infrared spectroscopy (FTIR); Giant unilamellar vesicle (GUV); Confocal laser scanning microscopy (CLSM); Synchrotron X-ray scattering (SAXS); and Surface plasmon resonance (SPR)

MECHANISM	MODEL SYSTEM(S)	METHOD(S)	REFERENCE
Spontaneous translocation	Jurkat	Pharmacological characterization; ELISA; Immunofluorescence	Rayne F. et al., Cell. Biol. Int., 2010
	Primary human CD4+ T-cell; Jurkat; Liposomes	Liposome sedimentation assay; ELISA; Immunofluorescence	Rayne F. et al., EMBO J., 2010
	Jurkat; Liposomes; Lipid monolayer	Fluorescence spectroscopy; Immunofluorescence	Yezid H. et al., J. Biol. Chem., 2009
	COS-1	Western blot; Gel shift assay	Chang HC. et al., AIDS, 1997
Exosome	Primary mouse astrocytes; U373.MG; SHSY-5Y; HEK293T; TZM-bl	Western blot; Exosome fractionation	Rahimian P. et al., J. Neurovirol., 2016
	HEK293T (No Tat present in exosomes)	Western blot; Exosome fractionation	Tang X. et al, JCI Insight, 2018
	J1.1; Jurkat (No Tat present in exosomes)	Western blot;	Narayanan A. et al., J Biol Chem., 2013
Pore formation	In silico	Molecular dynamic simulations	Herce H. et al., PNAS, 2007
	Liposomes	Flotation gradient; FTIR spectroscopy; Carboxyfluorescein	Zeitler M. et al., JBC Papers, 2015
	GUV	Light microscopy; CLSM; Single molecule microscopy	Ciobanasu C. et al., Biophys. J., 2010
	GUV	Confocal microscopy; SAXS	Mishra A. et al., Angew. Chem. Int. Ed., 2008
	In silico	Molecular dynamics simulations	Huang K. et al., Biophys. J., 2013
	In silico	Molecular dynamics simulations	Herce H. et al., Biophys. J., 2009
	Primary human CD4+ T-cells; CHO K1; psgA-745; HEK293T; HeLa; U2OS	SPR binding assay; Phosphoimaging; Western blot	S. Agostini. et al., EBioMedicine., 2017
Undefined	Primary cells derived from Kaposi’s sarcoma lesions on HIV-1 infected patients; H9; Jurkat; COS-1; HLM1	Western blot; Immunostaining	Ensoli B. et al., J. Virol., 1993
	Primary human macrophages; Primary rat hippocampal neurons; U-937; NG-108 mouse neuroblastoma	Western blot; ELISA	Johnston JB. et al., Ann. Neurol., 2001
	Primary human astrocytes; Rat C6 glioma; SVGA; HeLa	ELISA; Immunofluorescence	Chauhan A. et al. J. Biol. Chem., 2003

While several studies have addressed some of these questions, there is still controversy regarding the isoforms of Tat and whether they are truly representative of what is occurring within HIV-1-infected individuals. Tat is encoded by two exons, which are alternatively spliced to produce the full-length 101-residue protein ⁹³. There are, however, 72-residue and 86-residue isoforms of Tat that are commonly utilized in experimental systems, as well as various length peptides. Investigations have used expression constructs ^94,95 and recombinant proteins ^96–98 of the truncated 86-residue isoform of Tat to examine various biological processes. These points prompt the question that the experimental results one obtains may differ if performed with the full-length protein. Interestingly, the truncated isoforms are sometimes even referred to as the “full-length” protein. The first HIV-1 isolates (LAI, formerly LAV, and HTLV-IIIB) contained a pre-mature stop codon in the second exon of Tat, which encoded for an 86-residue protein ⁹⁹. It was later discovered that full-length Tat was 101-residues in length. Additionally, it is now understood that the full-length protein is more prevalent in clinical isolates ¹⁰⁰, and that variability within the second exon causes functional alterations, such as changes in viral gene expression and replication in specific cell populations ^101,102. Furthermore, the full-length protein has not been studied as frequently as the truncated isoforms. Moreover, the majority of the functional domains are within the first exon ¹⁰³, meaning the 72-residue and 86-residue isoforms of Tat transactivate the HIV-1 LTR similarly to the 101-residue full-length protein in most cell types ^104,105. Previous investigations studying the mechanism of Tat secretion have commonly used the 86-residue isoform ^5,38, however, the full-length protein has not yet been examined. Furthermore, the exact residues that comprise the basic domain of HIV-1 Tat are not consistent in the literature. The basic domain of Tat has been reported as either residues 49-57 ¹⁰⁶ or 48-57 ¹⁰⁷, though others, such as 47-57, have been used ¹⁰⁸. A single residue could alter or even ablate the secretion of Tat ³, suggesting that the full-length protein may yield different results. Tat is considered to be non-structural, the secondary structure of the protein remains a controversial topic, and the conformation of the protein is flexible ¹⁰⁹. Thus, we propose that it is likely the chemical properties of the peptide that are more relevant to secretion. Certain single residue substitutions may not alter the overall charge in the basic domain significantly enough to alter secretion ⁵¹. It has been suggested that only resides 49 to 51 and residue 11 are required for interaction with PtdIns(4,5)P₂ ⁵, however, utilization of different Tat isoforms between investigations makes comparisons difficult to interpret.

Combinatorial mutations are reasoned to alter the function of Tat considerably. These mutations are caused by several factors, including the error rate of HIV-1 reverse transcriptase, in conjunction with selective immune pressures ^103,110. The impact of genetic variability on the function of Tat has been well documented in the context of LTR transactivation ^94,111, but there are comparatively fewer publications examining the effect of genetic variation on Tat secretion. In silico analysis of Tat genetic signatures has been demonstrated to be correlated to co-receptor utilization ¹¹². Tat sequences derived from patients diagnosed with HIV-1-associated dementia ⁷ and those without dementia have demonstrated differences in transactivation of the HIV-1 LTR ¹¹¹. Furthermore, the individual influence of specific amino acid residues, such as the dicysteine motif C30C31 in Tat from HIV-1 subtype C ^113,114, have been the subject of debate regarding their contribution to neuroprotection. Correlating the genetic variability of Tat to neurocognitive impairment and alterations in secretion kinetics has yet to be examined. This information would provide insight into the contribution of extracellular Tat and the severity of HAND, potentially leading to more comprehensive diagnostic tools for those chronically infected with HIV-1.

Synopsis.

The HIV-1 transactivator of transcription (Tat) protein causes a myriad of effects to bystander cells, such as apoptosis and gene expression alterations, once secreted from infected cells. Tat is accepted as a crucial component of HIV-1-associated neurocognitive disorders (HAND) by causing neurotoxicity within the CNS. The mechanisms associated with Tat secretion are still poorly understood. This review examines gaps in published knowledge and uses existing evidence to categorize the mechanisms of secretion and their relevance to cells influenced during HAND.