HIV vectors: from base to translation, how do they work?

More than two decades have passed since genetically modified HIV was used for gene delivery. Through continuous improvements these early marker gene-carrying HIVs have evolved into safer and more effective lentiviral vectors. Lentiviral vectors offer several attractive properties as gene-delivery vehicles, including:

  1. sustained gene delivery through stable vector integration into host genome;
  2. the capability of infecting both dividing and non-dividing cells;
  3. broad tissue tropisms, including important gene- and cell-therapytarget cell types;
  4. no expression of viral proteins after vector transduction;
  5. the ability to deliver complex genetic elements, such as polycistronic or intron-containing sequences;
  6. potentially safer integration site profile; and
  7. a relatively easy system for vector manipulation and production.

Accordingly, lentivector technologies now have widespread use in basic biology and translational studies for stable transgene overexpression, persistent gene silencing, immunization, in vivo imaging, generating transgenic animals, induction of pluripotent cells, stem cell modification and lineage tracking, or site-directed gene editing.

Moreover, in the present high-throughput ‘-omics’ era, the commercial availability of premade lentiviral vectors, which are engineered to express or silence genome-wide genes, accelerates the rapid expansion of this vector technology. In the present review, we assess the advances in lentiviral vector technology, including basic lentivirology, vector designs for improved efficiency and biosafety, protocols for vector production and infection, targeted gene delivery, advanced lentiviral applications and issues associated with the vector system.


Lentiviruses are members of the viral family Retroviridae (retroviruses) that are characterized by their use of viral RT (reverse transcriptase) and IN (integrase) for stable insertion of viral genomic information into the host genome. Unlike other retroviruses, lentiviruses can replicate in non-dividing cells and cause slowly progressive diseases, including immunodeficiency, anaemia, pneumonitis and encephalitis, in their specific hosts (human, monkey, cat, horse, cow, goat and sheep).

Lentiviruses from different species have distinct properties in their genome structure, receptor usage and pathogenicity. Since the majority of lentiviral vectors are based on HIV-1, we start by providing basic information on this virus, including its genomic structure, viral proteins and life cycle.

HIV-1 genome

HIV has a single-stranded positive-sense RNA genome of approximately 9 kb in length that encodes nine viral proteins. The three largest open-reading frames encode its three major structural proteins: Gag, Pol and Env. The gag gene encodes viral core proteins. The pol gene encodes a set of enzymes required for viral replication.

Schematic representation of the HIV-1 viral genome

The env gene encodes the viral surface glycoprotein gp160. In addition to these major proteins, the viral genome also encodes the regulatory proteins Tat and Rev, which activate viral transcription and control the splicing and nuclear exports of viral transcripts respectively. Four other genes encode accessory proteins Vif, Vpr, Vpu and Nef. The viral genome is flanked by LTRs (long terminal repeats) that are required for viral transcription, reverse transcription and integration. The genome dimerization and packaging signal ‘’ is located between the 5 -LTR and the gag gene.

HIV-1 viral particles (virions)

HIV core proteins are encoded by gag and pol genes that are synthesized from the same transcripts by a ribosomal frameshift [10–12]. These make up the core structure of the infectious viral particle, called the virion (Figure 1B). The Gag-Pol precursor protein is also packaged and proteolytically cleaved into three viral enzymes, PR (protease), RT and IN, in virions. Env gp160 protein is cleaved into gp120 and gp41, the outer membrane proteins of HIV virions. Gp120 is referred to as the surface subunit or SU and gp41 is referred to as the TM (transmembrane) subunit.

Structure of HIV-1 virion

Both gp120 and gp41 are essential for normal infection of CD4 cells by wild-type virus. In many cases in vectors, the Env function is replaced with another protein to expand vector tropism. Once encapsidated, each virion contains Gag, Pol and Env proteins, accessory proteins Vif, Vpr and Nef, and two copies of viral genomic RNA.

HIV-1 infection

The natural HIV infection cycle is initiated by attachment of SU to its primary receptor CD4 and to its co-receptor CXCR4 (CXC chemokine receptor 4), expressed on T-lymphocytes or CCR5 on monocytes/macrophages, dendritic cells and activated T-lymphocytes. Upon receptor recognition, TM changes conformation to facilitate membrane fusion of HIV with the host cell, leading to viral entry (Figure 2). After cell entry, capsid proteins are uncoated, releasing the viral genome and MA (matrix protein), RT, IN and Vpr proteins into the cytoplasm. The positive sense RNA strand is converted into double-stranded DNA by viral RT. This proviral DNA is then imported into the nucleus and integrated into the host genome by viral IN.

HIV-1 production

Once proviral DNA is integrated, the LTRs capping the ends of the viral genome regulate transcription and polyadenylation of viral mRNAs. The LTR at the 5 -end of the genome acts as a combined enhancer and promoter for transcription by host cell RNA polymerase II. The LTR at the 3 -end of the genome stabilizes these transcripts by mediating their polyadenylation. Basal promoter activity by the 5 -LTR is minimal in the absence of viral transactivator Tat.

Initial transcription in the absence of Tat is inefficient and produces viral mRNAs that are multiply spliced into short transcripts. These short transcripts encode the non-structural proteins Tat, Rev and Nef that facilitate subsequent events in the viral life cycle. Newly synthesized Tat binds to TAR (transactivation-response element) on the 5 -end of HIV-1 mRNAs and transactivates and amplifies transcription of other structural viral proteins.

Meanwhile, Rev binds to the RRE (Rev-responsive element) on the viral transcripts to facilitate nuclear export of singly spliced or non-spliced viral transcripts and genomes. Singly spliced transcripts encode Env, Vif, Vpr and Vpu, whereas non-spliced viral RNAs are used for translation of Gag and Pol and as the genomic RNAs for progeny viruses. Exported viral genomes and proteins are assembled at the plasma membrane. After release from the host cell, multimerization of Gag and Gag-Pol activates the viral PR that converts these immature virions into mature infectious viruses.


Most currently available lentiviral vectors/packaging constructs are based on the second- or third-generation lentiviral vectors. In this section, we will review the transition of HIV-1 from a wild pathogenic virus to a transgene-carrying virus. We will then describe these highly sophisticated second- or third-generation lentiviral vectors that have been optimized for improved safety and infectivity.

Early HIV-1 vectors

The earliest lentiviral vectors were replication-competent viruses carrying transgenes. These were of use to lentivirologists to track viral replication in vitro and in vivo and as platforms to screen for anti-HIV-1 drugs. The first replication-competent HIV-1 vector was constructed by insertion of the CAT (chloramphenicol acetyltransferase) gene in the place of nef. To make these vectors safer, HIV vectors have evolved through a series of modifications to separate viral sequences needed for packaging and production from those encoding viral proteins. The first prototypes separated virus elements into two plasmids:

  1. a plasmid encoding HIV-1 proviral DNA with a deletion in the env gene; and
  2. a plasmid expressing Env.

Trans-complementation of Env protein from the separate plasmid allowed production of viruses that could undergo a single round of infection, but not a second round, since they do not carry the env gene. These early HIV vectors had transgenes inserted in nef or env with their expression driven by the 5 -LTR. Later, more sophisticated HIV-1-based vectors were generated carrying essential cis-acting elements for genome packaging, reverse transcription, and integration (LTRs, and RRE), but no viral proteins. In most cases, expression of the foreign genes in these vectors is driven by a heterologous internal promoter such as CMV (cytomegalovirus) or others.

Expanded tropism through pseudotyping with VSV-G (vesicular stomatitis virus envelope glycoprotein G)

Since HIV-1 Env recognizes human CD4 as a primary receptor, early HIV vectors could only infect human cells expressing CD4. In parallel to HIV vector development, Burns et al. ‘pseudotyped’ MLV (murine leukaemia virus)-based retroviral vectors by replacing the retroviral Env glycoprotein with the viral attachment protein of VSV-G. These VSV-G-pseudotyped retroviral vectors had two primary advantages over unmodified vectors. First, VSV-G is substantially more stable than retroviral or lentiviral envelopes, allowing pseudotyped viruses to be concentrated by ultracentrifugation to higher titres than ever before.

Secondly, although the receptor for VSV-G is controversial [26], it has been known that VSV-G binds the ubiquitous membrane component phosphatidylserine, allowing these vectors to transduce a markedly wider set of cells, even including nonmammalian cells (fish). Given these benefits, Akkina et al. used VSV-G to pseudotype HIV-1 vectors and demonstrated production of highly concentrated vectors that mediated highefficiency gene transfer into CD34 + haematopoietic stem cells. Most lentiviral vectors are now pseudotyped with VSV-G, giving them robust transduction into many cell types. This also increases that possibility of unintended transduction of users, so care needs to be taken in their use.

First-generation HIV-1-based lentiviral vectors with increased safety achieved through splitting vector components into three plasmids

Lentiviral vectors are derived from the pathogen HIV-1. Therefore there are safety considerations inherent to developing and using these as gene delivery vectors. Particular consideration must be given to the possibility of generating RCLs (replicationcompetent lentiviruses) with pathogenic potential when one intends to deliver a replication-defective gene delivery vector.

HIV-1 is a RG3 (risk group 3) virus. Therefore vectors derived from it are typically handled in the U.S.A. under BL3 (Biosafety Level 3) or modified BL2 + level containment in consideration of the finite risk of pathogen production. Although there are increased biosafety considerations with lentiviral vectors, unlike other retroviruses they can mediate stable gene transfer into both dividing and non-dividing cells, making them potent vectors for basic and translational research.

To reduce the likelihood of the production of RCLs in vector preparations, many laboratories have developed a number of ‘generations’ of lentiviral vector to reduce this risk. Lentiviral vector ‘generation’ is a loose terminology that does not consider the early replication-competent prototypes of HIV vectors described above as ‘first generation’. Rather, first-generation vectors are referred to as those vectors that first split the system into three separate plasmids to increase safety.

First-generation replication-deficient recombinant HIV-1 vectors are produced from three separate elements:

  1. a packaging construct;
  2. an Env plasmid encoding a viral glycoprotein; and
  3. a transfer vector genome construct.
Schematic representation of a first generation HIV vector

The packaging construct expresses HIV Gag, Pol and regulatory/accessory proteins from a strong mammalian promoter to generate viral particles. The Env plasmid expresses a viral glycoprotein, such as VSV-G, to provide the vector particles with a receptor-binding protein. These two plasmids have been specifically engineered without either a packaging signal or LTRs to avoid their transmission into vector particles and to reduce the production of RCL in vector preparations.

The transfer vector plasmid contains the transgene(s) and all of the essential cis-acting elements (LTRs, and RRE) for packaging/reverse transcription/integration, but expresses no HIV proteins. Since the transactivator Tat is not encoded by the transfer genome, the promoter activity by the 5 -LTR is minimal. Instead, transfer genomes use an internal promoter to express transgenes in transduced cells. This three-plasmid system allows the delivery of a gene of interest without expressing viral proteins in target cells.

Splitting the vector components into three plasmids means at least two recombination events are required to yield a replication-competent HIV-1-like virus during vector production. The use of VSVG, rather than HIV-1 Env, also reduces recombination, since it eliminates homologous sequences between the Env and transfer vector plasmids.

Second-generation lentiviral vectors with no viral accessory proteins (Vif, Vpu, Vpr or Nef)

First-generation lentiviral vectors provided a new level of safety for these potent gene-delivery vehicles. To increase safety further, second-generation vectors have been developed by modifying accessory genes in the system.

Schematic representation of a second generation HIV vector

HIV-1 Vif, Vpu, Vpr and Nef are called accessory proteins because they can be deleted without affecting viral replication in certain human lymphoid cell lines. However, these proteins are actually essential for efficient HIV-1 propagation/virulence in primary cells or in vivo. For example, lymphocytes are resistant to vif-deficient HIV-1 replication.

Vif is necessary to inactivate a host antiviral factor, APOBEC3G (apolipoprotein B mRNAediting enzyme-catalytic polypeptide-like 3G), to ensure efficient virus production. Similarly, Vpu neutralizes another cellular antiviral factor, called Tetherin. On the other hand, Nef promotes the degradation of host proteins, such as MHC class I and CD4, to augment virus production and facilitate immune evasion.

Therefore, although these accessory genes are important for HIV as a pathogen, they can be deleted in second-generation lentivectors. By replacing HIV-1 Env with VSVG, these second-generation vectors include only four of the nine HIV genes: gag, pol, tat and rev.

SIN (self-inactivating) vectors with a deletion in the U3 region of the 3 –LTR

Conventional lentiviral vectors integrate transgene cassettes flanked by two LTRs into the host genome. Under normal circumstances this should be a dead-end integration event.

Schematic representation of HIV vector constructs

However, if replication-competent recombinant lentiviruses are produced, they may be able to replicate in a similar manner to that of wild-type viruses. An alternative problem could arise if vector-transduced cells are subsequently infected by a wild-type lentivirus.

In this case, the wild-type virus can act as a helper virus to rescue the integrated vector into new viral particles to spread transduction beyond the original target cell.

Another serious issue is the undesired activation of cellular genes by integrated vectors. Since LTRs have an enhancer [binding sites for host transcription factors, including Sp1 (specificity protein 1) or NF-κB (nuclear factor κB)] and promoter regions, integration of LTRs into the genome can activate adjacent cellular genes. If semi-random integration of the transgene occurs near a proto-oncogene, these enhancers/promoters have the potential to activate transcription of these genes, moving that cell towards oncogenesis. Given these issues, SIN lentivectors have been developed. A standard vector genome is flanked by two LTRs that each contain three regions: U3, R and U5.

Integrated proviruses

U3 acts as a viral enhancer/promoter and R in the 3 -LTR acts as the polyadenylation signal. The 5 U3 and the 3 U5 element are therefore not present in mRNA from the provirus. Instead, the R region caps both ends. Duplication of LTR elements occurs during reverse transcription prior to integration when U3 in the 3 -LTR is copied and transferred to the 5 -LTR. As first shown in MLV vectors, if part of U3 in the 3 -LTR is deleted, its duplication will transfer the same deletion into the 5 -LTR’s promoter/enhancer region.

This deletion therefore results in transcriptional inactivation of potentially packageable viral genomes in the transduced cell. On the basis of success in MLV, this SIN approach was applied to HIV vectors by deletion of 3 -LTR elements, including its TATA-box-, Sp1-, NF-κB- and NFAT (nuclear factor of activated T-cells)-binding sites.

This SIN modification reduces the likelihood of:

  1. propagation of spontaneously produced replication-competent recombinant HIV-like viruses;
  2. insertional activation of cellular oncogenes by residual promoter activities of integrated LTRs;
  3. mobilization of integrated vectors by a wild-type virus; and
  4. transcriptional interference and suppression by LTRs.

Third-generation Tat-independent vectors from four plasmids

HIV-1 uses regulatory proteins Tat and Rev for viral transcription and nuclear export of intron-containing transcripts. Unlike its accessory proteins, Tat and Rev are absolutely required for HIV1 replication. To increase safety, third-generation vectors have been designed to be Tat-independent with Rev provided from a separate plasmid.Tat-independence is achieved by replacing the U3 promoter region of the 5 -LTR in the transfer vector with strong viral promoters from CMV or RSV (rous sarcoma virus).

The four plasmids used to generate third-generation vectors are:

  1. a packaging construct containing only gag and pol genes;
  2. a plasmid expressing Rev;
  3. an Env (VSV-G) plasmid; and
  4. a transgene plasmid driven by a heterologous strong promoter.

The enhancer/promoter region (U3) of 3 -LTR is also removed to add the SIN property. This vector system has only three of the nine genes of HIV, increasing its predicted biosafety. Since the vector elements are split into four plasmids, at least three recombination events are required to generate a replication-competent HIV-1-like virus.

Even if these occurred, the resulting viruses would have only HIV1 Gag, Pol, Rev and VSV-G proteins, with no active LTRs, Tat or accessory proteins. Increased safety of third-generation vectors is supported by the detection of no RCL within 1.4×1010 transducing units of vector produced from ten independent 14 litre production lots. When compared with other three-plasmid systems, the vector yields of third-generation vectors are typically lower.

Introduction of a cis-acting cPPT [central PPT (polypurine tract)] for increased vector transduction efficiency

During reverse transcription of HIV-1, plus strand DNA synthesis starts from the PPT and the cPPT. This leads to a plus strand overlap called the central DNA flap. It has been proposed that the central DNA flap enhances nuclear import of HIV-1 proviral DNA , although this is debated. In practice, introduction of cPPT into HIV-based vectors significantly increases vector transduction efficiency in vitro and in vivo.

Using WPRE [WHV (woodchuck hepatitis virus) post-transcriptional regulatory element] for increased transgene expression

Another cis-acting element that has been used to improve lentiviral vector expression is the WPRE sequence. WPRE increases the amount of unspliced RNA in both nuclear and cytoplasmic compartments. Introduction of WPRE into lentiviral vectors significantly increases transgene expression in target cells. Although increased expression is useful, the use of WPRE may raise safety concerns, since it contains a truncated form of the WHV X gene, which has been implicated in animal liver cancer. WPRE safety has subsequently been improved by a mutation of the open reading frame of the X gene.