Preterminal protein, the achilles heel of adenoviridae: Implications for adenoviral infections

Abstract

BACKGROUND

Adenoviruses pose a serious health risk particularly in the absence of any clinically approved treatment. As adenoviral infections are quite frequent and recent outbreaks manifest more virulent variant strains, the need to develop an effective treatment remains a priority. The adenoviral protein, preterminal protein (pTP), is one of the key common products of the viral lifecycle as it is necessary to initiate viral replication and hence the infection process. This makes pTP a potential chemotherapeutic target in the search for and development of an effective treatment for adenoviral induced infections. Here we report, for the first time, that glycosylation of pTP in situ prevents binding to ssDNA in vitro.

AIM

To explore whether specific structural tailoring of the adenoviral protein pTP, imparts the potential to scupper the viral replication process.

METHODS

All chemicals used were of reagent grade. Overexpression of pTP was achieved using the ‘BAC to BAC’ expression system. The presence and relative concentration of the protein was determined throughout the incubation period by the Bradford assay. The pTP was identified by MALDI-TOFF and sodium dodecyl sulphate polyacrylamide gel electrophoresis. For the removal of the aminosugar, a deglycosylase enzyme kit from PROZYME was used. Purification of cloned pTP (6xHis) was done with a ssDNA cellulose column followed by a Ni-NTA column. His-tags were excised with the Tobacco etch virus protease. Protein fractionation was performed with a fraction collector coupled to a UV detector (280 nm) from Pharmacia.

RESULTS

The pTP overexpressed in insect cells (Spodoptera frugiperda) (> 96 hours), is unable to bind to ssDNA in vitro. Treatment of this unbound protein with a deglycosidase enzyme that is specific for the removal of truncated unsubstituted O-linked Galβ(1-3)GalNAc-α1 disaccharides bound to Thr or Ser in a glycoprotein, restores binding to ssDNA. Data is presented as a linegraph for both the glycosylated and the deglycosylated proteins. Each point represents the mean of triplicate experiments (from different batches). Means and standard deviation were calculated and plotted on a line graph (with error bars).

CONCLUSION

The finding that glycosylation of cloned pTP in situ prevents binding to ssDNA in vitro could aid in the development of an effective treatment of adenoviral infections and/or as an adjunct to complement other anti-adenoviral chemotherapeutic strategies.

Key Words: Thomsen-Friedenreich antigen; Adenoviral; Preterminal protein; Spodoptera frugiperda cells; Protein priming mechanism; Glycosylation

Core Tip: A key objective in antiviral research is to recognize appropriate targets within the viral life cycle that are independent of normal host cellular functions. Adenoviruses (Ad) utilize a ‘unique’ protein priming mechanism that requires a viral protein, preterminal protein (pTP), for the initiation of replication. Therefore, it appears that pTP could potentially be a target for chemotherapeutic intervention. The finding that glycosylation of cloned pTP (in situ) in insect cells (Spodoptera frugiperda) prevents binding to ssDNA (in vitro) could prove useful in developing an effective treatment that facilitates the generation of an abortive adenoviral replication complex that would adversely affect human Ad virulence and propagation.

INTRODUCTION

Human adenoviruses (HAd), serotypes A to F, represent pervasive and clinically significant human pathogens that are responsible for many respiratory diseases (common cold, pharyngitis, and pneumonia), tonsillitis, cystitis, conjunctivitis (pink eyes) and gastroenteritis. They also infect the urinary tract, lymph system and heart. These infections (serotype dependent) can range from mild to severe and in the case of respiratory infections (they infect the mucous membranes), affect individuals with typical flu-like symptoms such as coughing, wheezing, fever and runny nose[1-3].

Characteristically, human adenoviral infections are unresponsive to most medications currently available on the market and treatment is mainly supportive and focused on relieving the symptoms. Whilst HAd infections are largely subclinical in the immunocompetent, there is significant morbidity and mortality linked to HAd infections in immunocompromised patients [cancer, organ transplants and human immunodeficiency virus (HIV) infected individuals], pediatrics, individuals from overpopulated areas and underdeveloped communities. Suffice it to say that the availability of effective antiviral therapies would have significant clinical benefits for humans. To date there is no Food and Drug Administration (FDA) approved antiviral for the treatment of HAd infections and attempts at overcoming this hurdle have met with limited success thus far, in the search for an effective therapeutic approach. The broad-spectrum anti-viral, Cidofovir, a nucleoside analogue, clinically approved by the FDA for treating cytomegalovirus induced retinitis in Acquired Immune Deficiency Syndrome (AIDS) patients, are efficacious against HAd serotypes tested in clinical trials but with varying results including reports of nephrotoxicity[1,3,4]. The phospholipid conjugate of Cidofovir, Brindofovir, has been shown, in phase 1 clinical trials, to have increased oral bio-availability and an improved side effect profile (reduced levels in kidney) when compared to the parent molecule. Disappointingly, phase 2 clinical trials were statistically no different to the placebo[5].

Alternative therapeutic approaches such as immunotherapy are promising but are labour intensive, not cost effective and limited with respect to efficacy against the different HAd strains. Cidofovir and intravenous immunoglobulin have been used clinically as a combination therapy to treat adenoviral hepatitis in a liver transplant patient with some success[6].

However, it remains a major challenge to develop antivirals and other treatment regimens that are effective against a range of HAd serotypes. It stands to reason that knowledge of key aspects of the life cycle of adenoviridae could prove invaluable in facilitating development of novel HAdenoviral therapies with high specificity and/or efficacy[7]. A specific area of the adenoviral infectious process that has been receiving a lot of attention over the years in pursuit of this elusive goal, is the adenoviral DNA replication cycle. The present study presents a brief overview of the relevant salient features of the adenoviral DNA replication cycle and the pertinent experimental procedures (and results) involved in acquiring preterminal protein (pTP), our molecule of interest. It concludes with a discussion on the findings and potential impact a specific post-translational modification, such as glycosylation, would have on the in vivo DNA binding properties of pTP and ultimately on the viral DNA infection process.

Background and rationale

Here we present an abridged version of the adenoviral DNA replication process, emphasising the pivotal role played by the viral polypeptide, pTP.

Adenoviral replication machinery: The HAd genome is a linear double-stranded DNA molecule of approximately 25-36 kilo-base pairs with inverted terminal repeats (ITRs) [e.g. adenoviruses (Ad)5 = 103, Ad2 = 102] (Figure 1). Ad replication origins[2] are located at the ends of the genome within the ITRs. Terminal protein (TP) is covalently attached to the 5’ end of the mature virion. This 55 kDa polypeptide is cleaved from a 80 kDa precursor, pTP, by an adenoviral protease at the end of the replication cycle[7]. TP protects the viral genome from nuclease degradation. TP-DNA, in cell culture, has also been shown to have a higher infectivity titre than the ‘naked’ viral genome[7].

Figure 1 Adenovirus 2 replication machinery. Adenoviral polymerase, preterminal protein and DBP (+ adenosine triphosphate) are essential for the replication process. Ad: Adenovirus; AdPol: Adenoviral polymerase; ATP: Adenosine triphosphate; PTP: Preterminal protein; NF: Host nuclear factor; Mu: Viral core protein; TP: Terminal protein.

The C-terminal end of precursor pTP forms a duplex with dCMP, that serves as a primer for initiation of viral replication at specific vDNA sequences within the host cell nucleus. In contrast, the N-terminus is known to bind both the adenoviral polymerase (AdPol) and the viral template genome (at base pair positions 9-18) at the start of the replication process, the latter, presumably to help stabilise the replication complex[8,9]. A more inclusive figure of the HAd genome binding regions is presented in the discussion. Adenoviral DNA replication occurs late in the infection cycle and is a prerequisite for expression of late viral genes, virion assembly[9] and ultimately viral progeny and propagation[9,10].

Mechanism of replication: There is general consensus that Ad deploy a protein priming mechanism to activate DNA replication (Figure 2). The aim of this study is to explore whether the removal of specific endogenously added amino sugar residues from cloned viral pTP, expressed in insect Spodoptera frugiperda (Sf-9) cells, restores binding to single stranded DNA in vitro[11-13]. The polypeptide, pTP, is a key component of the infection process and is covalently bound via the hydroxyl (OH) group of a specific Serine₅₈₀ residue, to the α phosphoryl group of the terminal dCMP of the fledgling nascent strand to form the heterodimer primer, Adpol-_NpTP_C-dCMP, a reaction catalysed by Adpol[14]. The 3^’OH of dRibose of this primer serves as a starting point for the synthesis of the nascent strand.

Figure 2 Proposed mechanism (abridged) of adenoviral replication. //: Some steps omitted for clarity.

The linking of the second nucleotide adenosine monophosphate and third thiamine monophosphate, is also catalysed by Adpol[14-16]. Base pairing with the 2^nd 3’G₄T₅A₆ (at positions 4, 5, and 6) trinucleotide on the template strand guides the synthesis of the first CAT trinucleotide. This CAT triplet then jumps back to positions 1, 2, and 3 on the template strand, base pairing with the 3’G₁T₂A₃ trinucleotide[16]. The synthesis of the 2^nd CAT triplet of the viral nascent strand by Adpol, then proceeds through the 2^nd 3’G₄T₅A₆ (TP) triplet. By the time the 7^th nucleotide of the nascent strand is synthesised, Adpol has completely separated from pTP[16,17], a separation process which began at the synthesis of the 1^st trinucleotide. Displacement of the extant complementary strand by the replication fork, from the template strand, occurs as Adpol continues to process the nascent strand[14,18].

Figure 2 is a schematic representation of the initiation of viral DNA replication that illustrates how the 4-6bp template G₄T₅A₆ trinucleotide guides the synthesis of the 5’pTP_C-CAT trinucleotide which then jumps/slides back to pair with the template 3’G₁T₂A₃ at positions 1, 2 and 3, after which chain elongation by AdPol proceeds[19-23].

METHODS AND MATERIALS

Preparation

For a more detailed account of the preparatory materials and protocols used, refer to Addendum A.

Materials

Cloned pTP (6xHis-tag), recombinant baculovirus, Sf-9 insect cells (as described previously[9,10]), MALDI–TOF MS (ABSciex 4800), ssDNA cellulose column, nickel column, peristaltic pump and fraction collector (Pharmacia), Bradford reagent, Invitrogen NuPAGE precast gels and arecombinant endo-O-deglycosidase kit[24]. All chemicals were of reagent grade and purchased from Sigma-Aldrich, unless otherwise indicated. GlycoPro^TM endo-O-glycosidase [recombinant from Streptococcus pneumonia (S.pneumonia)] obtained from PROZYME Inc., (Addendum C).

Purification of pTP

Adenoviral replication provides a convenient alternative model to study human DNA replication in addition to providing insight into key areas of the adenoviral life cycle. An absolute requirement for such an investigation is the need for homogeneous protein of sufficient quantities. Fulfilling this criteria for pTP, our protein of interest, involved cloning and overexpression of the viral polypeptide in Sf-9 insect cells using the ‘BAC to BAC’ expression system, as described previously[9,10] followed by extraction and purification. Infected cells were harvested after 96 hours and collected by centrifugation, washed twice with phosphate-buffered saline (10 mL) and stored at -70 °C until required. The purification of pTP was performed using a ssDNA cellulose column (regions within the N-terminus are known to bind to HAd DNA) followed by a Ni-NTA (binds histidine with high affinity) column as described previously[10]. The presence and concentration of protein were determined throughout the incubation period using the Bradford assay. During the purification, pTP was identified by sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) for both column (ssDNA and Ni-NTA) steps.

RESULTS

Optimal production of ‘active’ pTP by Sf-9 occurred at 72 hours of incubation. The overall yield of ‘active’ recombinant pTP becomes progressively less beyond this point. At 96 hours, the inactive (inability to bind ssDNA) portion is nearly 95% of the total rec-pTP and appears in the flow through as most of the protein did not bind to the DNA column. When this unbound pTP was passed through the Ni-NTA column there was nearly complete retention > 95% (graph not shown), confirming that the ssDNA flow through protein was indeed the cloned his-tagged pTP. Repeating the DNA binding step after elution with imidazole from the Ni-NTA column (this was done twice and with different batches) again resulted in no adherence to the ssDNA column. The identity of ‘unbound’ pTP from these purification steps was also confirmed by MALDI-TOFF MS and SDS-PAGE in every instance (Figure 3).

Figure 3 MALDI-TOFF MS of unbound preterminal protein. Inset (A) shows a 12% sodium dodecyl sulphate polyacrylamide gel electrophoresis chromatogram of pooled unbound and bound preterminal protein (1:1). Preterminal protein (30 µg) appears at 76.283 kDa.

Suspecting that the protein was post-translationally modified by the host, it was decided to test whether the viral polypeptide was glycosylated. A recombinant endo-O-deglycosidase kit was purchased and unbound fractions pooled and treated with the enzyme from S.pneumonia (Figure 4). The protocol supplied was used with minor modifications[24]. This enzyme is specific for unsubstituted Galβ(1-3)GalNAc-α1 disaccharides O-linked to Ser or Thr residues of glycopeptides.

Figure 4 Selectivity of the recombinant endo-O-deglycosidase (spec. activity: ≥ 12 U/mg) for unsubstituted Gal β(1-3)GalNAc-α1 disaccharides O-linked to Ser/Thr residues of glycopeptides (Cleavage site: |).

Results depict that the binding activity of unbound pTP can be successfully restored upon treatment with the deglycosidase (Figure 5).

Figure 5 Elution profiles of recombinant preterminal protein with respect to the ssDNA column. Experimental values are means of triplicate experiments.

It would thus appear that pTP is post translationally modified by the insect cells through glycosylation with Gal(β1-3)GalNAc-α1. This fact is not surprising given that the Sf-9 insect cell system is chosen as an expression system for recombinant human proteins precisely because of the presence of similar post translational modification systems such as glycosylation and phosphorylation. Gal(β1-3)GalNAc-α1, also known as the Thomsen-Friedenreich antigen (TF-antigen), is a tumour associated carbohydrate antigen (TACA) and known recognition marker associated with a number of metastasizing organotropic cancers (e.g. prostate cancer) and infections caused by some bacteria and viruses (including HAd) across the eukaryotic kingdom. This molecule serves an important role as mediator of cellular communication between affected cells in such instances[25].

Though it has been shown that aberrant glycosylation can promote metastasis and invasiveness in some human cancers[25], TACAs do also have roles that are beneficial to the host (not only in cancers) as is shown here. The fact that Sf-9 cells link the TF-antigen to a toxic viral protein to prevent it from binding to its genome, emphasises the importance this molecule plays in the natural defense systems of many living organisms. Potentially glycosylation could also account for the extremely low yields (± 15%) of ‘free’ pTP [even after increasing the Tobacco etch virus (TEV) protease incubation period from 4 hours to 8 hours] for the His-tag removal reaction. Yields for both 72 hours and 96 hours incubations were consistently below 20%. The TEV protease cleavage site is identified as having a serine residue[26]. It is possible that this Ser residue at the cleavage site is glycosylated and would consequently impede catalysis (Figure 6).

Figure 6 Tobacco etch virus protease 7 amino acid recognition site. Cleavage site: -x-Gln⤓ Ser-R₁.

Based on these findings it appears that His-tagged recombinant proteins, incorporating a Ser residue at the active site, are best not expressed in Sf-9 cells.

DISCUSSION

Selecting glycosylation as a strategy to modify the structure of pTP has a number of potential benefits for the invaded host: It allows for identification (tagging) of pTP. The protein is in this way targeted for destruction. It also maintains pTP in a soluble state so as to prevent precipitation (most likely due to protein-protein interactions) at relatively high concentrations (deglycosylation resulted in precipitation in vitro) which would cause cell death. The protein is also prevented from binding to host and viral DNA, a host cell directed counter activity that is beneficial to the ‘host’.

Traversing the nuclear envelope in moving from the rough endoplasmic reticulum (RER) to the nucleus is bound to be hampered or prevented. The natural product, rotenone, has recently been shown, using plaque assays, to significantly reduce adenoviral replication (IC₅₀ = 3.9 nM). This drug acts by preventing movement of the virion from the endosomes to the nucleus[27].

Glycosylation thus appears to be a specific response to an event that threatens the continued existence of the insect cell host and as such can be viewed as a natural survival strategy. It is thus entirely feasible that glycosylation, as a defence mechanism to invasion by HAd in vivo, could effectively combat the viral onslaught by targeting three susceptible areas for direct inhibition of viral replication.

The meaning represented by ‘a’ in Figure 7: Undermining the protein priming mechanism through glycosylation of serine₅₈₀ and as a consequence inhibit formation of the pTP_C-dCMP duplex primer (pTP-Ser-OH+dCMP) that is so vital for viral replication and propagation.

Figure 7 Potential sites a, b, c of direct inhibition, as a result of glycosylation, during initiation of adenoviral replication. Arrows 1 and 2 signify the jumping back step of the CAT trinucleotide at the 5’end of the nascent strand (refer to mechanism in text).

The meaning represented by ‘b’ in Figure 7: Causing steric hindrance at the ‘jumping back’ step of the pTP-CAT trinucleotide given that a highly regulated interaction between _NpTP_C and vDNA would be an absolute requirement for correct positioning of template (TP-GTA) and nascent (pTP-CAT) strand base pairs.

The meaning represented by ‘c’ in Figure 7: Preventing the association between pTP_N and adenoviral DNA at the ORI (bp 9-18) on the template strand during the formation of the pre-initiation complex.

The results of this study provide evidence that an in vivo glycosylated pTP has differential affinity for ssDNA (in vitro). Furthermore, by implication, translocation from the site of synthesis in the cytosol (RER) into the nucleus is likely to be hindered. It will be these factors which could ensure that viral replication does not progress (in vivo). Nonetheless, a glycosylated pTP (there are other potential Ser and Thr glycosylation sites on pTP[28]) will have a negative impact on base pairing between template, TP-GTA and newly synthesised nascent pTP-CAT. Interaction of Adpol with viral DNA (also with pTP_N) is required for vDNA replication, particularly at initiation[14]. It thus stands to reason that the stability and alignment of the interaction between pTP and AdPol, if altered by an event such as glycosylation, are likely to have a detrimental effect on initiation and progression of vDNA replication. This will likely reduce the efficiency of Adpol-catalysed viral DNA synthesis and/or result in abortive virion assembly, preventing viral propagation.

Agents that control the activities of glycosidases have been found to have therapeutic effects not only against many carbohydrate-related diseases but as shown, potentially also against adenoviral infections. In fact, glycosidase inhibitors have been used chemotherapeutically, in a clinical setting, for a few decades already, to treat a range of unrelated diseases such as diabetes, influenza, HIV infections, tumor metastasis, and glycosyl sphingolipid storage diseases[29,30].

The fact that adenoviral infections are generally localised, facilitates tissue specific targeting in the form of a reaction cocktail, constituted of ingredients that would include the endo-O-glycopeptide, in addition to agents that would inhibit ‘on site’ glycosidases. It follows then that a potential regimen to treat, for example, HAd induced conjunctivitis, could include a mixture of molecules as part of an eye drop formulation (localised chemotherapy) consisting of: (1) N-Acetylgalactosaminyltransferase (to add the sugar to pTP)[31]; (2) The substrate Gal(β1-3)GalNAc-α1[30] (can also be used in isolation); (3) An endo-O-glycosidase inhibitor such as galactostatin or deoxynojirimycin to prevent removal of the sugar from pTP[29]; and (4) Cidofovir (a deoxycytidine analogue) to inhibit the viral polymerase AdPol[7].

Inhibition of viral glycosidases and/or stimulation of glycosyl transferases may be viewed as complementary to other antiviral therapeutic strategies (adjunctive), bearing in mind that different viruses require different treatment regimes. One only has to consider the chemotherapeutic approach used in the treatment of HIV/AIDS (e.g. dolutegravir the integrase inhibitor), to appreciate the feasibility of such an approach.

CONCLUSION

The finding that glycosylation by Sf-9 cells of cloned pTP in situ prevents binding to ssDNA in vitro could thus aid in the development of an effective treatment, in isolation or as an adjunct together with another anti-viral therapy, for adenoviral infections. Glycosylation could potentially be afforded multiple roles in the preservation of a living organism. In this particular instance, a known TACA, Gal(β1-3)GalNAc-α1, the TF-antigen, is harnessed by Sf-9 larval cells as an integral part of a natural survival strategy to safeguard against the toxic effects of pTP. Firstly, the 3-dimensional structure of pTP is altered, which impedes subsequent binding to host DNA. Secondly, this would most likely also hamper movement across the nuclear envelope from the site of synthesis on the rough endoplasmic reticulum. Thirdly, the fact that Sf-9 cells tag an alien protein, such as pTP, with a known ubiquitous mammalian recognition element[27] provides corroborative evidence for the exploitation of the TF-antigen as a target in the development of an antigen-specific therapeutic vaccine. This could be directed at a range of organotropic cancers and other pertinent viral and bacterial infections. It is envisioned that our laboratory, in the near future, will pursue the testing of various formulations of the cocktail ingredients (as outlined above), on virus induced localised infections, such as keratoconjunctivitis, in animal models such as the rat.