Recent advances in discovery and functional analysis of the small proteins and microRNA expressed by polyomaviruses

Abstract

The polyomavirus family consists of a highly diverse group of small DNA viruses isolated from various species, including humans. Some family members have been used as model systems to understand the fundamentals of modern biology. After the discovery of the first two human polyomaviruses (JC virus and BK virus) during the early 1970s, their current number reached 14 today. Some family members cause considerably severe human diseases, including polyomavirus-associated nephropathy (PVAN), progressive multifocal leukoencephalopathy (PML), trichodysplasia spinulosa (TS) and Merkel cell carcinoma (MCC). Polyomaviruses encode universal regulatory and structural proteins, but some members express additional virus-specific proteins and microRNA, which significantly contribute to the viral biology, cell transformation, and perhaps progression of the disease that they are associated with. In the current review, we summarized the recent advances in discovery, and functional and structural analysis of those viral proteins and microRNA.

1. Polyomaviruses

The first family member of the Polyomaviridae was isolated from mice during the late 1950’s and named mouse polyomavirus, (MPyV) (Stewart et al., 1958). Following this discovery, the second family member, simian virus 40 (SV40) was isolated as a contaminant of poliovirus vaccines prepared from the monkey kidney cell cultures (Sweet and Hilleman, 1960). More than a decade later, two new members were isolated, but they were of human origin, the JC virus (JCV) and BK virus (BKV) in 1971. Later on, many diverse family members were isolated from a wide range of animal species including hamster, monkey, mouse, cattle, rabbit, fish, parakeet and rat (Imperiale, 2007; Van Doorslaer et al., 2018). The polyomaviruses and papillomaviruses were originally classified under the genus of the family Papovaviridae, but in the year 2000, this family was split into two families, Polyomaviridae and Papillomaviridae by the International Committee on the Taxonomy of Viruses. In recent years, however, this classification was updated (Johne et al., 2011; Moens et al., 2017)

The first human polyomavirus, JCV, was isolated from a human brain biopsy tissue obtained from a patient with a demyelinating disease known as, progressive multifocal leukoencephalopathy (PML) (Padgett, 1971) through the co-culturing of the biopsy samples with a tissue culture system. However, the second human polyomavirus, BKV, was isolated from a kidney transplant recipient with polyomavirus-associated nephropathy (PVAN) (Gardner, 1971). These two human polyomaviruses are now known to cause severe illnesses in a subset of immunocompromised individuals. In fact, JCV, for example, causes a fatal demyelinating disease of the human brain known as PML. JCV asymptomatically infects a large number of the world’s population (20–40%) during childhood (Kean et al., 2009) and establishes a persistent infection in the kidneys (latent infection) and perhaps in B cells in the archetype form and this number significantly increases during the adulthood (60–70%) (Berger, 1987, 2000; Ferenczy et al., 2012; Monaco et al., 1998). As described in more detail later in this review, the configuration of the regulatory region of the genome of the archetype strain significantly differs from those of the well-known strains, such as Mad-1 and Mad-4. Under immunocompromised conditions, the latent archetype strain of JCV gains an opportunity to reactivate itself and undergoes deletions and duplications within its regulatory region, allowing the new recombinant viruses to become more virulent, such as Mad-4 and Mad-1 (Frisque and White, 1992). The new strains then travel to the brain by an unknown mechanism and lytically infect oligodendrocytes and astrocytes, leading to the initiation of the PML (Ferenczy et al., 2012; Frisque and White, 1992). In addition to the kidneys and B cells, the hematopoietic progenitor cells and tonsillar stromal cells were also considered to harbor JCV, and thus serve as additional sites for JCV latent infection (Monaco et al., 1998). PML is a rare and fatal disease in the normal healthy population. However, it generally manifests itself under immunosuppressive conditions such as AIDS (Berger, 1992; Berger et al., 1992). Additionally, JCV infections were also reported to be associated with a brain disease called “granule cell neuronopathy” (Dang et al., 2012a; Du Pasquier et al., 2003; Soleimani-Meigooni et al., 2017; Wuthrich et al., 2009, 2016). Moreover, close to two decades ago, PML cases started occurring relatively high proportions in some Crohn’s disease and multiple sclerosis (MS) patients who were treated with immunomodulatory drugs (antibodies, such as Natalizumab and Efalizumab). These antibodies were designed to target specific cell surface receptors on B and T cells (Kleinschmidt-DeMasters and Tyler, 2005; Langer-Gould et al., 2005; Van Assche et al., 2005). The intent of using these immunomodulatory drugs was to prevent the infiltration of B cells into the skin and extravasation of T cells into the brain, respectively, and thus to reduce the harmful effects of these immune cells in target organs. The primary target of Efalizumab is the integrin molecule, CD11-alpha, on both T and B cells in order to block the interaction of these cells with the intercellular adhesion molecules (ICAM) to prevent lymphocyte infiltration into target organs (Gottlieb et al., 2002). On the contrary, Rituximab, another immunomodulatory antibody, interacts with the CD20 receptor on B cells and causes B cell death (Tobinai et al., 1998). Natalizumab, another immunomodulatory antibody, is known to interact with the α4 integrin on either T or B cells (Kleinschmidt-DeMasters and Tyler, 2005). Alpha4 integrin heterodimerizes either with integrin β1 (α4β1) or integrin β7 (α4β7) on both B and T cells. These complexes (α4β1 and α4β7) then serve as target molecules for vascular cell adhesion molecules (VCAM) on endothelial cells and thus prevent T-cell infiltration into the brain. Patients treated with any of these drugs are vulnerable to developing immunomodulation in specific subsets of immune cells.

Consequently, such immunosuppressive conditions allow JCV to reactivate itself from latency to induce PML. It is important to note that one may suggest that the loss of immune surveillance in those treated individuals would be sufficient to induce PML. However, data indicate that additional unknown factors play a role in reactivation of JCV in immunosuppressed individuals, leading to the development of PML. It has been suggested that two main factors play a significant role in this outcome, including 1) the occurrence of the specific mutations within the JCV VP1 gene (Gorelik et al., 2009, 2011) and 2) the genetic makeup of those individuals.

Similarly, BK virus (BKV), the second polyomavirus isolated from humans, infects kidney epithelial cells and is highly prevalent in the human population. When reactivated, BKV causes a disease known as polyomavirus-associated nephropathy (PVAN) in individuals who are also treated with highly immunosuppressive drugs, which creates a significant problem in allograft patients and may result in transplant failure.

Following the isolation of JCV and BKV, due to the availability of new technologies, such as digital transcriptome subtraction and rolling circle amplification techniques, additional polyomaviruses were discovered in recent years, raising the total number to 14 (Table 1). Despite this rapidly inflated polyomavirus number, it turned out that only several of those viruses are known to be associated with a specific human disease. In addition to JCV and BKV, only four other polyomaviruses are categorized as the disease-causing polyomaviruses. For example, Merkel cell polyomavirus (MCPyV) is associated with a Merkel cell carcinoma (Feng et al., 2008), Trichodysplasia Spinulosa-associated polyomavirus (TSPyV) induces trichodysplasia spinulosa (van der Meijden et al., 2010), the human polyomavirus-6 (HPyV6) has been implicated in a chronic inflammatory disorder, Kimura disease, that involves subcutaneous tissues (Nguyen et al., 2017; Rascovan et al., 2016; Schowalter et al., 2010; Schrama et al., 2014) and the human polyomavirus 7 (HPyV7) was detected in a pruritic rash case in transplant recipients (Ho et al., 2015; Schowalter et al., 2010). The rest of the human polyomaviruses, including Washington University polyomavirus (WUPyV) (Gaynor et al., 2007), Karolinska Institute polyomavirus (KIPyV) (Allander et al., 2007), human polyomavirus-9 (HPyV9) (Scuda et al., 2011), human polyomavirus-10 (HPyV10) (Buck et al., 2012), Malawi human polyomavirus (MWPyV) (Siebrasse et al., 2012), MX human polyomavirus (MXPyV) (Yu et al., 2012), St. Louis Polyomavirus (SLPyV-11) (Pastrana et al., 2013; Lim et al., 2013), human polyomavirus-12 (HPyV12) (Korup et al., 2013) and New Jersey polyomavirus (NJPyV) (Mishra et al., 2014) have yet to be linked to a particular human disease. Table 1 illustrates the involvement of the human polyomaviruses in various human pathologies.

Table 1.

Currently known human polyomaviruses.

Full Name	Abbreviations	Source of isolation	Time of discovery	Associated disease
JC virus	JCPyV	Brain, PML patient	Padgett (1971)	Progressive Multifocal Leukoencephalopathy
BK virus	BKPyV	Urine, Kidney transplant recipient	Gardner (1971)	BK virus-associated nephropathy, Hemorrhagic cystitis
Karolinska Institute Polyomavirus	KIPyV	Nasopharyngeal Tissue	Allander et al. (2007)	Respiratory disease
Washington University Polyomavirus	WUPyV	Nasopharyngeal Tissue	Gaynor et al. (2007)	Respiratory disease
Merkel Cell Polyomavirus	MCPyV	Skin lesions, Merkel cell carcinoma patient	Feng et al. (2008)	Merkel cell carcinoma
Human Polyomavirus 6	HPyV6	Normal skin	Schowalter et al. (2010)	Keratoacanthoma; Kimura disease
Human Polyomavirus 7	HPyV7	Normal skin	(Ho et al., 2015; Nguyen et al., 2017; Schowalter et al., 2010)	Pruritic and dyskeratotic dermatosis, Trichodysplasia spinulosa
Trichodysplasia Spinulosa-Associated Polyomavirus	TSPyV	Trichodysplasia Spinulosa, skin lesions	van der Meijden et al. (2010)
Human Polyomavirus 9	HPyV9	Blood, skin, urine, kidney transplant recipient	Scuda et al. (2011)	Unknown
Malawi Polyomavirus	MWPyV	Stool, a patient with warts, hypogammaglobulinemia, infections, and myelokathexis (WHIM) syndrome	Buck et al. (2012)	Unknown
Human Polyomavirus 12	HPyV12	Resected human liver tissue	Korup et al. (2013)	Unknown
St. Louis Polyomavirus	STPyV	Stool samples, healthy child	(Lim et al., 2013; Pastrana et al., 2013)	Unknown
MX Polyomavirus	MXPyV	Stool samples from children	Yu et al. (2012)	Unknown
New Jersey Polyomavirus	NJPyV	Biopsy from the swollen endothelial cells, Pancreatic transplant recipient.	Mishra et al. (2014)	Unknown

2. Polyomavirus genome and its expression

2.1. Genomic structure

Polyomaviruses have a small, circular, double-stranded DNA genome encased in an icosahedral virion (Imperiale, 2007; Saribas et al., 2010). The viral DNA is composed of both “noncoding control region (NCCR) and coding regions (Fig. 1A). NCCR regions contain the DNA regulatory elements (promoter/enhancer elements) and the origin of the viral DNA replication. The regulatory region drives the expression of the early and late coding regions and initiates viral DNA replication along with the viral and host proteins. The coding regions (early and late), on the other hand, contain coding sequences for large and small viral proteins. Each coding region produces a single pre-mRNA transcript on each direction, which undergoes an alternative splicing (cis- and trans-splicing) process to produce these proteins. In the archetype forms of polyomaviruses, the noncoding regulatory elements display hyper-variability concerning deletions and duplications. For example, the JCV archetype strain undergoes intricate recombination events within its regulatory region, resulting in more effective virulent strains such as Mad-1 and Mad-4. The archetype strain itself, for example, contains only one 98-bp repeat within its noncoding regulatory region. Still, this 98-bp region is split by two unique insertional blocks, 23-bp and 66-bp in length (Safak et al., 1999b). During the recombination process, these insertional blocks are sometimes entirely removed. Then the remaining parts of the 98-bp region are rejoined and duplicated, as seen in the production of one of the prototype strains of JCV known as the Mad-1 strain. Of course, many other recombinant strains are also produced during the transition from the archetype to the more virulent strains (Ferrante et al., 2003; Reid et al., 2011). For example, the Mad-4 strain of JCV contains two 84 bp tandem repeats rather than 98 bp repeats where the 23-bp and 66-bp insertional blocks are completely removed (Ferrante et al., 2003; Reid et al., 2011; Safak et al., 1999b). In addition, the occurrence of the rearranged forms within the noncoding regulatory region has been reported for BKV, too (Barcena-Panero et al., 2012; Randhawa et al., 2003).

Fig. 1.

Circular map of JC virus. (A) JCV genome contains an NCCR and a bidirectional coding region (early and late). Numbering is according to the JCV Mad-1 nucleotide strain (Accession: J02226.1). The positions of the recently discovered open reading frames (ORFs) on the late coding region due to either cis- or trans-splicing were designated as ORF1, ORF2, ORF3, ORF4 and ORF5. (B) Illustration of the trans-splicing events for the ORF1 and ORF2. (C) Illustration of the two main cis-splicing events that occur on the viral late pre-RNA transcripts. (D) Illustration of the mechanisms of the cis-splicing events to create an ORF3 protein. (E) Illustration of the mechanisms of the cis-splicing events to create ORF4 protein. (F) Illustration of the mechanisms of the cis-splicing events to create the ORF5.

2.2. Regulatory proteins encoded by the early coding region of JC virus

Transcription of all polyomaviruses occurs in early and late directions and is regulated by 1) the cis-acting DNA elements located within the noncoding region and 2) trans-acting cellular and viral transcription factors that directly or indirectly interact with these cis-acting DNA elements. Transcription extends in a bidirectional manner from the transcription initiation sites near the DNA replication’s origin to produce the early and late pre-mRNA transcripts. These transcripts then undergo cis-and trans-splicing to produce a wide range of mRNA species, upon which their translation produces various regulatory and structural (capsid) proteins.

For the sake of clarity of our descriptions, we will describe the gene expression patterns of the monkey (SV40), the human (JCV, BKV, MCPyV, and TSPyV), and mouse (MPyV) polyomavirus coding regions rather than describing it for all polyomaviruses. Among those, the expression patterns of the JCV coding regions will be described in detail, and the deviations from those for other polyomaviruses will also be described. Two major regulatory proteins, the small tumor antigen (Sm t-Ag) and the large tumor antigen (LT-Ag), were encoded by the JCV early coding regions (Fig. 1A). It should be noted here that these two proteins are also universally expressed by all other polyomaviruses from their early coding regions. In 1995, Trowbridge et al. reported the discovery of additional alternatively spliced variants of LT-Ag (Trowbridge and Frisque, 1995), named T′ proteins (T′165, T′136, T′135, T′147, and T′152) (Prins and Frisque, 2001). The function of the T′ proteins was not studied in great detail in viral replication and transcription. However, some experimental evidence suggests a role for them in cell transformation through interaction with two tumor suppressor proteins, p107 and p130 (Bollag et al., 2006) (Table 2).

Table 2.

Small viral proteins produced by polyomaviruses due to alternative splicing.

Virus name	Early/late gene expression	Putative Protein	No. of amino acids	Expected size (kDa)	References
TSPyV	Early	Middle T	332	38	(van der Meijden et al., 2015)
TSPyV	Early	Tiny T	85	10	(van der Meijden et al., 2015)
TSPyV	Early	Small T*	197	24	(van der Meijden et al., 2015)
TSPyV	Early	21 kT	184	21	(van der Meijden et al., 2015)
MCPyV	Early	57 kT	432	48	Carter et al. (2013)
MCPyV	Early	ALTO	248	27	Carter et al. (2013)
MPyV	Early	Middle T	421	46	(Soeda et al., 1979)
MPyV	Early	Tiny T	75	8	Riley et al. (1997)
JCPyV	Early	T’165	165	18	Prins and Frisque (2001)
JCPyV	Early	T’136	136	15	Prins and Frisque (2001)
JCPyV	Early	T’135	135	15	Prins and Frisque (2001)
JCPyV	Early	T’152	152	17	Prins and Frisque (2001)
JCPyV	Early	T’147	147	16	Prins and Frisque (2001)
JCPyV	Late	ORF1	58	6.4	(Saribas et al., 2018)
JCPyV	Late	ORF2	72	7.9	(Saribas et al., 2018)
JCPyV	Late	ORF3	70	7.7	Saribas et al. (2023)
JCPyV	Late	ORF4	173	19.0	Saribas et al. (2023)
JCPyV	Late	ORF5	278	30.6	Saribas et al. (2023)
JCPyV	Late	Agnoprotein	71	7.8	(Del Valle et al., 2002)
SV40PyV	Early	17 kDa	135	17	Zerrahn et al. (1993)
SV40PyV	Late	Agnoprotein	62	6.8	(Gilbert et al., 1981)
BKPyV	Early	TruncTAg	ND	17–20	Abend et al. (2009)
BKPyV	Late	Agnoprotein	68	7.5	(Rinaldo et al., 1998)

ND: Not determined.

Note: All polyomaviruses universally express large and small T antigens and they are not included in this table.

The LT-Ag is the largest and most prominent regulatory protein of all polyomaviruses and is required to initiate the viral DNA replication and transactivation of the viral late promoter. It also plays a critical role in auto-regulating its own promoter to regulate the JCV early gene expression (Safak et al., 1999a). Our recent proteomic studies showed that LT-Ag of JCV, beside targeting two tumor suppressor proteins, p53 and pRb, targets a wide range of cellular proteins and protein complexes, including, Smc5/6 complex, PP4–PP1 complex, E3-Ubiquitin-protein ligase, V-ATPase, ribosomal proteins, actin-myosin network, and others (Saribas and Safak, 2020). Little is known about the function of the Sm t-Ag. However, the previous studies showed a regulatory role for Sm t-Ag in viral replication cycle. It also promotes the S-phase entry to facilitate viral DNA replication (Khalili et al., 2008) along with LT-Ag. Our recent proteomic studies also showed that JCV Sm t-Ag uniquely targets a variety of proteins and protein complexes in cells, including chromatin-remodeling proteins, mitochondrial proteins, PP2A complex proteins, Zn-binding proteins, heterogeneous ribonuclear proteins, chaperone proteins, actin-myosin network, ribosomal proteins and others (Saribas and Safak, 2020). These proteomic studies also revealed common host protein targets for both JCV LT-Ag and Sm t-Ag in cells, including ribosomal/RNA binding proteins, actin-myosin network, and others (Saribas and Safak, 2020).

2.3. Regulatory proteins expressed by the early coding regions of the other polyomaviruses

Early coding regions of several other polyomaviruses also express alternatively spliced forms of proteins. Mouse polyomavirus (MPyV) early coding region expresses two additional proteins besides its Sm t-Ag and LT-Ags, one of which is called middle T antigen (420 aa long protein). It has membrane-anchored features as well as well-known oncogenic properties. It activates the Src-family protein tyrosine kinases in a PP2A-dependent manner to stimulate the Src signaling pathway during cell transformation (Fluck and Schaffhausen, 2009; Ichaso and Dilworth, 2001; Zhou et al., 2011) (Table 2). Another small protein encoded by the mouse polyomavirus early coding region is called “tiny T antigen”, which stimulates the ATPase activity of Hsp70 through its J domain and localizes to both cytoplasm and nucleus (Riley et al., 1997) (Table 2). The simian polyomavirus isolate-40, SV40, also encodes a small regulatory protein from its early coding region called 17K protein (17 kD protein). Although its precise function is unknown (Rundell et al., 1980; Yang et al., 1979), it was reported to contribute to the transformation of both human and rodent cells (Boyapati et al., 2003; Zerrahn and Deppert, 1993; Zerrahn et al., 1993). Following the discovery of JCV T’ proteins, the expression of a 17–20 Kd protein from the early coding region of BKV was reported by Abend et al., in 2009 (Abend et al., 2009). This protein is a truncated form of BKV LT-Ag (truncTAg) (Table 2). TruncTAg is expressed in both BKV-transformed and lytically infected cells. It primarily localizes to the nucleus (Abend et al., 2009) and is believed to be involved in cell transformation (Abend et al., 2009). Another human polyomavirus, MCPyV, which has oncogenic properties in humans and is known to be associated with human Merkel cell carcinoma, was also reported to encode two proteins from its early coding region besides its LT-Ag and Sm t-Ag proteins. One of these is named the Large T Open Reading Frame (ALTO) (a 27 Kd protein) (Carter et al., 2013) and the other is called 57 kT protein (DeCaprio, 2017). It was suggested that ALTO does not appear to be required for the replication of MCPyV. However, it may play an auxiliary role for the virus life cycle. It should also be mentioned here that it is more likely that the early coding region of some or all of the other polyomaviruses may encode additional splice variants of the early mRNA, resulting in the generation of additional novel proteins, but the identity of those that have yet to be discovered.

2.4. Generation of novel proteins by the trans-splicing of the early transcripts of polyomaviruses

So far, we have described the generation of early regulatory proteins for JCV, BKV, MCPyV, MPyV, and SV40 as a result of a cis-splicing process, where the introns are removed, and exons are rejoined within a single pre-mRNA transcript (Black, 2003). Previous studies also showed that another type of alternative splicing occurs for some of the pre-mRNAs expressed from the early coding region of polyomaviruses called trans-splicing, where the joining of the exons occurs after the removal of the introns from the two different pre-mRNA transcripts to create novel viral proteins. Such a splicing event was first described for trypanosomes in 1986 and later in the nematode, Caenorhabditis elegans (Krause and Hirsh, 1987; Sutton and Boothroyd, 1986). Previous reports demonstrated that this type of splicing also occurs between the viral transcripts. Eul et al. first reported that the SV40 early large T antigen pre-mRNA transcripts (Eul et al., 1996a, 1996b; Eul and Patzel, 2013) undergo a trans-splicing process generating a 147 nt duplication of the SV40 LT-Ag exon 2 between nucleotides 4571-4425, corresponding amino acids 83–131. This event generates an open reading frame for a 100 kD LT-Ag protein with high transforming activity in rodent cells (Eul and Patzel, 2013). This novel protein is now called “super LT-Ag” and surprisingly contains a binding site for pRb, a critical target site for LT-Ag to transform cells.

2.5. Regulatory proteins expressed by the late coding region of polyomaviruses

In contrast to the early coding region, the late coding regions of the polyomaviruses are known to encode primarily the structural capsid proteins such as VP1, VP2, and VP3 (Frisque et al., 1984), although not every human polyomavirus encodes VP3 protein. However, an exception to these types of expression has also been known for a long time (Frisque et al., 1984; Saribas et al., 2016), where a small regulatory protein, named agnoprotein, is also encoded by only three polyomaviruses including JCV, BKV and SV40 from their late coding regions (Frisque et al., 1984; Saribas et al., 2016). In addition to the agnoprotein expression, the generation of a wide range of alternatively spliced variants of the JCV late pre-mRNA was also previously reported by Shishido-Hara et al. (2000). However, whether these variants encode any protein was unknown during that time.

The main capsid protein of the polyomaviruses, VP1, forms the outer shell of the virions and plays a critical role in the spread of the virus from cell to cell by mediating the attachment of the viral particles to the cell surface receptors. The minor capsid proteins, VP2 and VP3, are localized to the inner portion of the capsid and play critical roles in the deposition of the viral genome into the capsids during virion biogenesis. Besides these three capsid proteins, the SV40 late transcripts also encode a novel regulatory “late protein” called “VP4” (Daniels et al., 2007). The VP4 expression, like VP3 expression, is generated from a downstream AUG start codon within the VP2 coding region at amino acid Met228. It is a small protein (13.9 kD), and its expression of VP4 was demonstrated both using the reticulocyte lysates in vitro and SV40-infected cells in vivo (Daniels et al., 2007). Although its precise function is unknown, it is thought to contribute to the release of the SV40 virions from the infected cells (Giorda et al., 2012; Raghava et al., 2011, 2013; Tange et al., 2011). A recent study by Henriksen et al., however, failed to detect the expression of VP4 in SV40-infected cells. This was perhaps due to the technical difficulties associated with their detection system (Henriksen et al., 2016).

We now know that the genome of the polyomaviruses can generate novel splice variants from their late transcripts (Shishido-Hara et al., 2000), although the coding capacity of some of those has been recently discovered. For example, our group recently reported several novel open reading frames (ORFs) associated with various novel splice variants of JCV late transcripts (Saribas et al., 2018a, 2023). These variants are generated by trans-splicing and customary cis-splicing events. These variants include ORF1, ORF2, ORF3, ORF4, and ORF5, two of which are produced by a trans-splicing (ORF1 and ORF2), and the remaining three, ORF3, ORF4, and ORF5, are generated by cis-splicing. During the generation of ORF1 and ORF2, the following events take place: Two varying lengths of the 5′-short coding regions of VP1 are spliced into between the coding regions of agnoprotein and VP2 after replacing the intron 1 (Fig. 1AB), allowing a frameshift occurring within the VP2 coding region and terminating with a stop codon (Fig. 1AB). ORF1 and ORF2 encode small proteins, 58 and 72 amino acids long, respectively, and are detected both in PML patient samples in vivo and in the infected cells in vitro (Saribas et al., 2018a). Although both proteins share a common coding region with VP1, they also harbor their own unique coding sequences at their C-terminal region (Fig. 1B). The effect of these proteins on viral replication cycle was analyzed by mutational analysis on the viral background and results showed a significantly less efficiency in their replication rate compared to WT virus suggesting that these ORF1 and ORF2 play important regulatory roles in the viral life cycle (Saribas et al., 2018a). The cellular distribution studies with immunocytochemistry showed that ORF1 and ORF2 proteins do not follow a particular distribution pattern in cells. That is, they show a relatively uniform distribution pattern for the cytoplasm and nucleus. The coding capacity of the additional splice products of the JCV late coding region was also further analyzed by our group through RT-PCR studies, resulting in the discovery of three additional novel ORFs encoding ORF3, ORF4, and ORF5 (Saribas et al., 2023).

2.5.1. ORF3 splicing pattern and its cellular distribution

The open reading frame for ORF3 protein coding region results from the alternative splicing of the JCV VP1 coding region as shown in Fig. 1CD (Saribas et al., 2023). Such a splicing pattern is not only detected in cells infected with JCV in tissue culture but also in the total RNA isolated from JCV-infected patient brain tissue samples, confirming the authentic expression of ORF3 during the JCV infection cycle in vitro and in vivo. This splicing creates an open reading frame encoding a small 70 aa long protein, which contains two distinct regions: 1) the C-terminus of ORF3 (a 29 aa long region) is unique to ORF3 itself because it is produced as a result of a frameshift occurred at the exon junctions within VP1 coding region and 2) The N-terminus of ORF3 (aa 1–41) is identical to that of VP1 (Fig. 1D). The stable expression of this protein was explored by western blotting (Saribas et al., 2023) and immunocytochemistry (Fig. 2A) (Saribas et al., 2023) and was demonstrated its detectable expression by both assays. ORF3 protein shows no unique expression pattern in glial and non-glial cells. Both cell types show a nuclear and cytoplasmic distribution pattern (Fig. 2A). The ORF3 protein sequence is shown below.

Fig. 2.

Immunocytochemistry (ICC) analysis of ORF3 and ORF4 protein expression. (A) Analysis of the subcellular distribution patterns of ORF3 protein (Saribas et al., 2018a) in SVG-A cells (Major et al., 1985). A pCGT7-ORF3 plasmid was transfected into SVG-A cells on tissue culture slides and at 24h posttransfection, cells were then processed for ICC using a primary mouse α-T7 monoclonal antibody (Novagen, catalog no. 69522-4) and the secondary FITC-conjugated goat α-mouse antibody (Licor, IRDye^® 800CW Goat anti-Mouse IgG Secondary Antibody). Samples were then stained with DAPI (ThermoFisher, 4′,6-Diamidino-2-Phenylindole, Dihydrochloride, catalog no. D1306), mounted using “ProLong^® Gold Antifade” mounting medium (Life Technologies, catalog no. P36934) and dried overnight. Slides were then examined under a fluorescence microscope (Leica, DMI-6000B, objective: HCX PL APD 40×/1.25 oil, employing LAS AF operating software) for visualization of the ORF3 protein. (B) Analysis of the subcellular distribution patterns ORF4 SVG-A cells. In parallel to the ICC experiments described for panel A, SVG-A cells were also transfected cells with a pCGT7-ORF4 plasmid and analyzed for the detection of ORF4 protein and promyelocytic leukemia protein (PML) using a combination of the primary (Millipore, anti-T7 polyclonal, catalog no. AB3790) and anti-PML monoclonal, Santa Cruz, catalog no sc-966) antibodies overnight. After washing the cells with 1xPBS buffer for 10 min intervals, cells were incubated with a combination of a (FITC)-conjugated-goat α-rabbit (Abcam, catalog no. Ab6717) and Rhodamine-conjugated goat α-mouse (Millipore, catalog no. AP124R) secondary antibodies. Slides were then incubated with DAPI, mounted and examined under a fluorescence microscope as described for panel A.

[MKMAPTKRKGERKDPVQVPKLLIRGGVEVLEVKTGVDSITE (1–41 aa region overlaps with VP1), NKVPRWNNFSKECHSAISSHEHRAQGVPR (a 42–70 aa unique region)].

2.5.2. ORF4 protein splicing pattern and its cellular distribution

The JCV late coding region also creates another interesting ORF during the splicing process, which can encode a 173 aa long protein (ORF4). ORF4 exhibits a complete identity to the C-terminal coding region of VP1 (aa 181–354) (Fig. 1A and E). Analysis of ORF4 expression by immunocytochemistry demonstrated that unlike ORF3, ORF4 showed a nuclear and punctate distribution pattern (Fig. 2B), suggesting that it could co-localize with specifically known proteins or protein groups in the nucleus, including Cajal bodies, nuclear stress bodies, polycomb-group proteins, speckles, paraspeckles, promyelocytic leukemia nuclear bodies (PML-NBs) and/or nucleoli (Cremer and Cremer, 2001; Misteli, 2007). In parallel studies, ORF5 expression was not detected by either Western blotting or immunocytochemistry assays, suggesting that ORF5 is perhaps an unstable protein, or it is not expressed in vitro cases (Saribas et al., 2023) (Fig. 1F).

2.5.3. ORF4 protein targets PML-NBs in the nucleus and reorganizes their distribution

Immunocytochemistry studies using specific antibodies directed against the nuclear macromolecular protein assemblies including PML-NBs, nucleoli, polycomb-group proteins, Cajal bodies, nuclear stress bodies, paraspeckles and speckles demonstrated that ORF4 targets PML-NBs and reorganizes their distribution pattern in the nucleus (Fig. 2B). Such an expression pattern is not cell-type specific but universal (Saribas et al., 2023). Other JCV proteins, including Sm t-Ag, ORF1, agnoprotein, ORF2, LT-Ag, and ORF3, were also analyzed to address whether any of these proteins show a similar distribution and reorganizational pattern with PML-NB as ORF4 does. These studies confirmed that none of those but only ORF4 protein targets PML-NBs and reorganizes their distribution pattern in the nucleus (Saribas et al., 2023). Similar immunocytochemistry studies were also performed using the putative BKV ORF4 protein. Such studies also showed the co-localization of BKV ORF4 protein with PML-NBs, resulting in their reorganization. However, unlike JCV ORF4, BKV ORF4 protein is not strictly confined to the nucleus but showed cytoplasmic distribution too. That is, it exhibited both a nuclear and cytoplasmic distribution pattern in cells (Saribas et al., 2023).

It is known that PML-NBs are composed of several transient and permanent members. For example, the PML protein itself is one of the major permanent components of the PML-NBs. hDaxx (a transcriptional co-repressor), ATRX (a chromatin remodeling protein containing an ATPase/helicase domain that belongs to the SWI/SNF family), and Sp100 (an interferon-inducible protein) are the other three permanent members of this protein complex. In order to be structurally part of the PML-NB complex, any protein (viral or cellular) must be sumoylated either by SUMO-1 or SUMO-2 enzymes or harbor a SUMO-interacting domain (SIM) in their own sequence (Bund et al., 2014; Meinecke et al., 2007; Zhong et al., 2000). Our immunocytochemistry studies also further revealed that, in addition to the PML protein itself, the ORF4 protein also co-localizes with two other permanent members of the PML-NBs (ATRX and hDaxx) and reorganizes their distribution pattern in the nucleus (Saribas et al., 2023) but not with a third permeant member, Sp100 (Saribas et al., 2023) This suggests the following possibilities: 1) Either ORF4 mediates the dispersal of Sp100 into the nucleoplasm or 2) it influences the degradation of this protein. By both cases, ORF4 could inhibit the function of PML-NBs. This observation is similar to a case observed for the human papillomavirus L2 protein (Florin et al., 2002). These findings further imply that ORF4 most likely has a defining negative effect on the assembly of the PML-NB components so that ORF4 inhibits the antiviral effects of PML-NBs and thus allowing a rapid spread of JCV infection in the body.

It is implicated that PML-NBs have regulatory roles in the induction of intrinsic and innate antiviral responses (Scherer and Stamminger, 2016). Thus, it is believed that PML-NBs play critical regulatory roles in type I and type II interferon (IFN) immune responses as being the co-activators of the IFN induction pathways (Regad and Chelbi-Alix, 2001; Regad et al., 2001; Ulbricht et al., 2012). However, in the case of polyomavirus infections, the structural composition of the PML-NBs is most likely altered by ORF4-type proteins, and thus, their function in IFN induction pathways is inhibited during the viral infection cycle. This allows the rapid spread of the viral infection and exacerbates the disease development.

2.5.4. ORF4 protein harbors its own nuclear localization signal

Immunocytochemistry studies showed that ORF4 protein primarily localizes to the nucleus of the ORF4 positive cells, suggesting that it harbors its own nuclear localization signal. Indeed, bioinformatic prediction studies revealed a putative nuclear localization signal for ORF4 located towards the middle portion of the protein (QLRKRRVKNP) (Fig. 3). This sequence has high sequence homology to those of BKV and SV40 (Fig. 3). Mutational analysis revealed that this NLS is a typical sequence for both JCV VP1 and ORF4 proteins and functions as an authentic NLS sequence for both proteins. This finding contradicts a common assumption that the NLS sequence of JCV VP1 is localized towards the N-terminus of the protein (Ishii et al., 1996; Shishido-Hara et al., 2000). In fact, the previously designated NLS sequence of JCV VP1 is quite divergent than those of BKV and SV40 (Fig. 3). It is important to note that since BKV and SV40 also have potential ORF4 coding sequences harboring potential NSL sequences similar to that of JCV ORF4 (Fig. 3). These NLS sequences may also function as authentic NLS sequences for their VP1 and ORF4 proteins in cases of BKV and SV40. However, this hypothesis needs to be experimentally demonstrated. The known small viral proteins produced by polyomaviruses are summarized in Table 2.

Fig. 3.

Clustal multiple sequence alignment for JCV, BKV and SV40 VP1 and ORF4 proteins. VP1 and thus ORF4 protein sequences of JCV, BKV and SV40 were aligned using the “Clustal Omega” program (https://www.ebi.ac.uk/Tools/msa/clustalo). Each VP1 protein contains two putative nuclear localization signals (NLS) designated as NLS-1 and NLS-2 as highlighted in yellow and green color respectively. ORF4 coding region is highlighted in dark gray. The putative ORF4 protein of these three viruses contains a predicted NLS sequence (NLS-2) of their own. This was predicted by using the following website: http://nls-mapper.iab.keio.ac.jp/cgi-bin/NLS_Mapper_form.cgi. The functionality of the NLS-2 of JCV ORF4 and VP1 was previously validated (Saribas et al., 2023).

Among polyomaviruses, JCV, BKV, and SV40 are the only viruses that produce agnoprotein from their late coding region. Agnoprotein is a highly basic small phosphoprotein with 71, 68, and 62 amino acid long structures, respectively (Khalili et al., 2005). Among these, JCV agnoprotein is the most studied one, concerning its structural and functional features and identification of its cellular targets (Coric et al., 2014, 2017; Saribas et al., 2012, 2013, 2018b; Saribas and Safak, 2020; Sariyer et al., 2006, 2008, 2011). We will further highlight its properties in the upcoming sections.

2.6. Agnoprotein of polyomaviruses

2.6.1. Cellular distribution of agnoprotein

Agnoprotein is a cytoplasmic protein, accumulating relatively high concentrations at the perinuclear area of the infected cells (Ellis and Koralnik, 2015; Gerits and Moens, 2012; Nomura et al., 1983; Rinaldo and Hirsch, 2007; Saribas et al., 2012, 2016). Its perinuclear localization strongly suggests that it interacts with the endoplasmic reticulum (ER). Indeed, immunocytochemistry studies by Suzuki et al. using antibodies directed against the ER markers including calreticulin and BiP (Suzuki et al., 2010) clearly demonstrated the co-localization of agnoprotein with ER in JCV-infected cells. Mutational analyses further demonstrated the importance of the basic residues, including Lys22, Lys23, and Arg24, in targeting agnoprotein to the ER (Suzuki et al., 2010). In addition, a low level of agnoprotein was also detected in the nucleus by immunocytochemistry studies (Saribas et al., 2012), which is highly consistent with the presence of a weak “bipartite nuclear localization signal” (BNLS) (Fig. 4B) (Dingwall and Laskey, 1986, 1991) within its structure which is located towards its N-terminal region (Saribas et al., 2012, 2019). It should be noted here that the distribution of agnoprotein to both the cytoplasmic and the nuclear compartments of the cells suggests diverse roles for this protein. Similarly, BKV agnoprotein was also investigated concerning its cellular distribution pattern using a green fluorescent protein (EGFP) fusion protein by Unterstab et al. (Unterstab et al., 2010). Interestingly, these studies demonstrated that agnoprotein also co-localizes with “lipid-droplets (LD)” in the infected cells. Further studies suggested that amino acids encompassing the region aa 20 through 42 mediate this localization (Unterstab et al., 2010).

Fig. 4.

Illustration of the primary structure, functional domains and 3D structure of agnoprotein. (A) Primary structure of agnoprotein. Agnoprotein is a small and highly basic protein containing positively charges residues (Arg and Lys). However, a region towards the central region of the protein contains a Leu/Ile/Phe-rich domain, which forms the major alpha helix of the protein. It contains only three Phe (Phe31, Phe35 and Phe39), all of which confined to this region. This protein contains only one Cys residue at position amino acid 40. (B) Illustration of the predicted and confirmed functional motifs of agnoprotein. The functionality of the phosphorylation sites (Ser7, Ser11 and Thr21) was previously demonstrated (Sariyer et al., 2006). Analysis of the MTS, NLS and NES, ER membrane retantion signal were also previously reported (Saribas and Safak, 2020; Saxena et al., 2021; Suzuki et al., 2012). (C) Three dimensional structure of agnoportein was previously resolved by magnetic nuclear resonance (NMR) (Coric et al., 2017). According to this structure, agnoprotein contains one major α-helix (Lys23-Phe39) and one minor α-helix (Leu6-Lys13) and the rest of the protein adopts intrinsically unstructured conformation. Note that the MTS and minor a-helix overlaps.

2.6.2. Stable dimer and oligomer formation by agnoprotein

Several viral and cellular proteins form dimers and oligomers essential for the homeostasis of biological systems. The biological activity of many proteins, including enzymes, ion channels, receptors, and transcription factors, is mediated through weak subunit interactions (Marianayagam et al., 2004). In some cases, however, the dimeric and oligomeric forms of some proteins were shown to have highly stable structures. Nitric oxide synthetase (NOS), a cellular enzyme, for example, forms such stable dimeric structures. These structures are sometimes resistant to various denaturing reagents and heat (Kolodziejski et al., 2003). Some of the HIV-1-encoded small regulatory proteins, such as vpr, rev, nef, vif and tat were previously reported to form oligomeric structures (Cullen, 1998a; Frankel et al., 1988a; Frankel et al., 1988b; Kwak et al.; Yang et al., 2001; Zhao et al., 1994) and those types of the structural forms are known to be critical for the execution of their functions. For example, it was reported that dimerization of vpr is required for its interaction with HIV gag protein and this interaction leads to the accumulation of gag protein at the plasma membrane (Fritz et al., 2010). Oligomerization of HIV-1 Rev is also required for its interaction with the Rev response element (RRE) (~350 nt long) to transport the HIV RNA from the nucleus to cytoplasm (Cullen, 1998a, 1998b, 2003a, 1998b; Pollard and Malim, 1998). During this process, oligomeric forms of Rev also interact with CRM-1, the human nuclear export receptor, to export Rev-bound RNA from the nucleus to the cytoplasm (Cullen, 2003b; Fornerod et al., 1997). A transcription factor from Ebola virus (EBOV), VP30, also depends on its homodimerization feature to function in viral transcription (Hartlieb et al., 2003). Some viral proteins require RNA molecules for their oligomerization. A prime example of such a oligomerization type is the Borna disease virus (BDV) N nucleoprotein. It requires the presence of a 5′-specific region of BDN RNA for its oligomerization (Hock et al., 2010). Polyomavirus LT-Ag is another interesting oligomeric viral protein in this regard, which binds to the origin of the viral DNA replication in a double hexamer manner to initiate viral DNA replication (Auborn et al., 1988; Lynch and Frisque, 1990; Simmons et al., 1990a, 1990b).

Our group previously reported the formation of stable dimeric and oligomeric structures by agnoprotein of JCV, BKV, and SV40 in vitro (Saribas et al., 2011). It turned out that agnoprotein was also found to form these stable structures in infected cells in vivo (Saribas et al., 2011), which are highly resistant to denaturing agents, including heat denaturation, sodium dodecyl sulfate (SDS) and 8M urea treatment (Saribas et al., 2011). These oligomeric structures formed by agnoprotein do not result from a covalent bond formation between the monomers. It is suggested that numerous strong ionic attractions and hydrophobic interactions are the main contributing factors for forming such stable structures (Saribas et al., 2011). Additional in vivo fluorescence resonance energy transfer (FRET) studies by Suzuki et al. also suggested the formation of oligomeric structures by agnoprotein (Suzuki et al., 2010). Later mapping studies demonstrated that the Leu/Ile/Phe-rich agnoprotein region is required for this protein’s dimer and oligomer formation (Saribas et al., 2011).

2.6.3. Structure of agnoprotein determined by nuclear magnetic resonance (NMR)

Agnoprotein plays an important role during the life cycle of three polyomavirus, including JCV, BKV, and SV40, because in the absence of its expression, none of these viruses can sustain a productive life cycle (Carswell and Alwine, 1986; Carswell et al., 1986; Dang et al., 2012b; Gerits and Moens, 2012; Hay et al., 1984; Hou-Jong et al., 1987; Merabova et al., 2008; Myhre et al., 2010; Ng et al., 1985; Okada et al., 2002; Resnick and Shenk, 1986; Saribas et al., 2013; Sariyer et al., 2011; Suzuki et al., 2010, 2012; Unterstab et al., 2010). It comprises a highly basic amino acid composition. The majority of the positively charged residues such as Lys (K) and Arg (R) are localized towards the N-terminus of the protein, which are highly conserved among the three agnoproteins of BKV, JCV, and SV40 (Saribas et al., 2016) (Fig. 4A). The amino acids located towards the middle portion of the protein are, however, mostly composed of hydrophobic residues ranging from aa 28 to 39 and this region is designated as the Leu/Ile/Phe-rich domain of the protein (Fig. 4A). Other two unique features of the agnoprotein amino acid composition are the following: 1) it contains only three Phe residues, all of which are confined to the Leu/Ile/Phe-rich domain and 2) it has only one cysteine residue (Cys40), localized right after the Leu/Ile/Phe-rich region (Fig. 4A).

Our group and others have attempted to resolve the 3D structure of agnoprotein by conventional X-ray crystallography studies. However, the difficulties associated with the protein purification studies using various expression systems and due to the ability of agnoprotein to form stable dimers and oligomers did not allow us to obtain satisfactory results (Saribas et al., 2011). However, employing NMR techniques and using synthetic peptides (encompassing amino acids Thr17 to Gln52), we and others resolved the first partial three-dimensional (3D) structure of agnoprotein in 2014 (Coric et al., 2014). In 2017, however, using the same technique and the synthetic peptide approach again, the 3D structure of the full-length agnoprotein was resolved (Coric et al., 2017) (Fig. 4C). These NMR studies revealed the formation of two alpha helices; one (major alpha helix) is present in the middle portion of the protein, encompassing amino acids Arg24-Phe39 and the second (minor alpha helix) is localized towards the N-terminus of the protein, encompassing amino acids Leu6-Lys13. The remaining regions of the protein exhibit an intrinsically disordered conformation, meaning that these regions do not have a mainly defined tertiary structure (Coric et al., 2014, 2017).

Further analysis of the major helical region by the “helical wheel illustration" revealed the formation of four distinct faces for this region: a hydrophobic face 1 (containing Ala25, Leu29, Leu32 and Leu36), a hydrophobic face II (containing Ile30, Leu33 and Leu37), a hydrophilic face (containing residues Lys23, Arg27, Glu34 and Asp38) and an aromatic face (containing Phe31, Phe35, Ile28 and Phe39) (Coric et al., 2014). As it will be discussed further in the upcoming sections, amino acid residues located at the hydrophilic face of agnoprotein are highly critical for the release of agnoprotein to the extracellular matrix. In addition, it is believed that the overall low hydrophobicity and high net charge result in the formation of the disordered regions of the proteins. Thus, the disordered regions provide considerable flexibility to agnoprotein to interact with its various targets within the cells to amplify and diversify its functions (Dyson and Wright, 2005; Fink, 2005). The presence of intrinsically disordered structures was also reported previously for various viral proteins. For instance, E4-ORF3, a small adenovirus protein, was shown to form multivalent functional matrices through its intrinsically disordered structural regions in the infected cells (Ou et al., 2012) and this interaction was suggested to inactivate the several tumor suppressor proteins such as TIM24, MRE11/R-ad50/NBS1 (MRN), promyelocytic leukemia factor and p53 (Ou et al., 2012). Thus, it is conceivable that the intrinsically unstructured regions of agnoprotein may function similarly in targeting its cellular targets and thereby significantly contribute to the overall outcome of the viral growth.

Over the years, our group and others demonstrated the multi-functionality of agnoprotein and showed that it targets multi-viral and cellular proteins in infected cells (Gerits and Moens, 2012; Johannessen et al., 2008; Saribas et al., 2016; Sariyer et al., 2006). This most likely happens through its various functional domains, including phosphorylation sites. In fact, agnoprotein has several phosphorylation sites for various kinases such as protein kinase C (PKC), cAMP-dependent kinase and casein kinase II (CKII) [amino acids 41 to 44 (TGED)] as indicated on Fig. 4 (Johannessen et al., 2008; Sariyer et al., 2006). In fact, our group previously demonstrated the phosphorylation of agnoprotein by PKC at Ser, Ser11 and Thr21 (Sariyer et al., 2006). Agnoprotein has additional functional domains for various predicted functions (Fig. 4B). For instance, both the N-terminal and C-terminal regions of the protein contain two “endoplasmic reticulum (ER) membrane retention” motifs. One is located at the amino acids 2 to 5 (VLRQ) and the other at the amino acids 67 to 70 (EPKA).

Studies by Suzuki et al. corroborate with the presence of the predicted functional domains of agnoprotein where the targeting of ER by agnoprotein was demonstrated (Suzuki et al., 2010). agnoprotein also contains a “weak bipartite nuclear localization signal” (NLS), located towards the N-terminus of the protein including amino acids 8–24 (RKASVKVSKTWSGTKKR). The functionality of this site was previously demonstrated by immunocytochemistry studies in that agnoprotein was detected in the nucleus in low amounts (Saribas et al., 2012; Saribas and Safak, 2020). Interestingly, the Leu/Ile/Phe-rich region, also known as the major alpha helix domain, appears to contain a strong nuclear export signal (NES), which displays high similarity to those of bovine papillomavirus E1 proteins and the HIV-1 Rev protein (Daugherty et al., 2010) (Fig. 4B). Our group also demonstrated that agnoprotein interacts with a cellular factor, exporting (Crm-1), perhaps through the residues located within the NES of agnoprotein, such as Leu33 and Glu34. Double mutations (L33D + E34L) within this region resulted in a phenotype where agnoprotein mostly localized to the nucleus in contrast to its main localization region, the cytoplasm, confirming the functionality of NES of agnoprotein (Saribas and Safak, 2020). Another predicted signature sequence of agnoprotein is located toward the middle portion of the protein (aa 47–50, DGKK), called amidation site. Such a modification site is, in general, found in the structure of hormones, neurotransmitters, and growth factor peptides and is essential for their biological activity. These peptides are first produced in inactive forms, but upon cleavage at certain sites, the C-terminus of the peptide is modified by amidation and secreted into the bloodstream (Walsh and Jefferis, 2006), where they then fulfill their activities. Moreover, agnoprotein contains another fascinating organelle targeting sequence in its far N-terminus called "Mitochondrial Targeting Sequence" (MTS) [(amino acids, 1–14 (MVLRQLSRKASVKV)]. This sequence has two main features to be designated as a MTS. One is that this region adopts a short α-helical structure, and the other is that it contains several positively charged residues (Arg3, Lys9, and Lys 13), all of which are located on the same surface of the alpha-helix (Fig. 4C). These two features are known to be characteristics of a mitochondrial targeting sequence.

2.6.4. Release of agnoprotein to the extracellular matrix from agnoprotein-positive cells

Viruses have evolved remarkable strategies to alter the host and the neighboring cell environment to complete their life cycle and propagate their progeny. To achieve such an important task, they secrete some of their own regulatory proteins to the extracellular space from the infected cells either to interact with the neighboring cell surface molecules or even enter those cells. Such a phenomenon most likely creates a more conducive environment for the initiation of a relatively more rapid replication cycle for the incoming virus. After release, some of those viral proteins would perform various functions, such as acting as complement inhibitors (Al-Mohanna et al., 2001; Anderson et al., 2002), cytokine inhibitors (Alcami et al., 1998; Liu et al., 2000), cytokine mimickers (Liu et al., 2004; Suzuki et al., 1995), or inflammatory cell inhibitors (Lucas et al., 1996) to evade either the host immune system or perhaps to establish a long-term chronic infection cycle in host. Both some DNA and RNA viruses display such an interesting behavior. Poxvirus, a double-stranded DNA virus, for example, releases a viral protein homologous to the IL18, named IL18BPs, which negatively regulates host IL18 activity and thus contributes to virulence by modulating the host immune response (Smith et al., 2000). Similarly, the human cytomegalovirus (hCMV) releases several viral proteins from the infected cells again to evade the immune system, including UL128, UL21, UL146, and UL147. Of those, the UL21 is known to bind to CC15/RANTES to block its interaction with cellular receptors (Luganini et al., 2016). Another example of a released viral protein is the adenovirus E4/49K, which binds to leukocyte common antigen (CD45) to alter leukocyte function (Windheim et al., 2013).

There are also various RNA viruses that release their regulatory protein from the infected cells. For example, a dimeric and oligomeric nonstructural regulatory protein of the West Nile virus, NS1, was reported to be released from infected cells (Muller and Young, 2013). Another prime example of an RNA viral protein released from infected cells is the HIV-1 viral protein, vpr. Vpr, like agnoprotein forms dimeric and oligomeric structures (Levy et al., 1994; Zhao et al., 1994). Nearly two decades ago, the detection of this protein in the serum of infected patients was first reported by Levy et al. (1994), suggesting the release of vpr from the HIV-infected cells in HIV patients. Vpr was demonstrated to be readily taken up by neurons and inhibits their axonal growth (Kitayama et al., 2008). It was also shown to target mitochondria to interact with a mitochondrial protein, named adenine nucleotide translocator (ANT) which is located on the inner mitochondrial membrane. This interaction caused alterations in the mitochondrial membrane potential (Δѱm), and thus plays a disruptive role in Ca²⁺ metabolism and apoptosis (Brenner and Kroemer, 2003; Mukerjee et al., 2011).

A decade ago, Otlu et al., reported the release of JCV agnoprotein from agnoprotein-positive cells (Otlu et al., 2014). Following this report, based on the 3D NMR structure (Coric et al., 2014), the mechanism of its release was investigated by mutational analyses (Saribas et al., 2018b). These studies showed that the major α-helix region of agnoprotein plays a critical role in its release (Saribas et al., 2018b) (Fig. 5A). As previously reported (Coric et al., 2014), the major α-helix region is made up of four distinct faces: two hydrophobic, a hydrophilic and an aromatic face. Certain amino acid residues on each face of the alpha helix were analyzed by mutagenesis for their contribution to the release of agnoprotein. Data showed that certain amino acids specifically located on the hydrophilic surface of the alpha helix are highly critical for agnoprotein release including those of two negatively (Glu34 and Asp38) and two positively (Lys22 and Lys23) charged residues (Saribas et al., 2018b). Interestingly, although Phe31 is an aromatic residue, which is also located on the hydrophilic face, it significantly contributes to the release process (Saribas et al., 2018b) (Fig. 5A). It is important to note that the overwhelming majority of the amino acid residues involved in agnoprotein release are conserved among JCV, BKV and SV40 agnoproteins, except Asp38, which is substituted with Glu in BKV and with Gln in SV40 agnoprotein. Such a high level of conservation suggests that they most likely play similar roles in the release of SV40 and BKV agnoproteins, too. Although the precise mechanism of agnoprotein release is currently unknown, the endoplasmic reticulum (ER) to Golgi pathway is most likely involved in this process because the use of a specific inhibitor (Brefeldin A) of protein trafficking from the ER to the Golgi was shown to inhibit the transport of the agnoprotein in tissue culture system (Otlu et al., 2014).

Fig. 5.

Presentation of the hydrophilic face of agnoprotein involved in its release and illustration of a possible connection between the JCV infection and initiation of the development of neurological disorders. (A) The hydrophilic surface of the major alpha-helix of agnoprotein (determined by NMR studies) (Coric et al., 2014) plays a major role in the release of agnoprotein from the agnoprotein-positive cells. Two negatively (Glu34 and Asp38), two positively (Lys22 and Lys23) charged amino acid residues and an aromatic amino acid Phe3, located on the hydrophilic surface were determined to be critical for contributing to the release of agnoprotein (Saribas et al., 2018b). (B) Illustration of a possible connection between the JCV infection and initiation of the development of neurological disorders. Agnoprotein released from the infected cells (astrocytes and oligodendrocytes) can be taken up by the neighboring cells (neurons). Then it can have toxic effects on neurons which can lead to neuronal death, perhaps by dysregulating the metabolic and Ca²⁺ handling pathways of mitochondria (Saxena et al., 2021); and targeting various proteins, protein complexes and organelles (Saribas and Safak, 2020).

3. Possible role of agnoprotein in the initiation of neurological disorders

The functional consequences of the released agnoprotein are currently unknown. However, the released agnoprotein most likely executes its functions by directly entering the cells and thus engaging its targets or initiating its effects on cells by interacting with the cell surface receptors. In this individual or a combination of both cases, agnoprotein influences the biology of the neighboring cells to support the growth of the incoming virus and ensure a successful infection cycle. These possibilities were partially addressed by Saribas et al., by treating the cells with a synthetic agnoprotein peptide in a tissue culture model system (Saribas et al., 2018b). However, these studies consistently revealed that agnoprotein tends to bind to an unidentified cell surface receptor rather than directly entering the cells (Saribas et al., 2018b). However, these studies also showed that some portion of agnoprotein in fact is internalized by the cells at low levels. In either case, it is unclear what the released agnoprotein’s function might be in neighboring cells. The following three possibilities are plausible: (i) In the first possibility, one can speculate that the cell surface-bound or internalized agnoprotein, directly or indirectly, modulates the neighboring host-cell gene expression so that the incoming virus grows more efficiently in the infected cells. (ii) In the second possibility, it is also likely that either internalized or the cell surface-bound agnoprotein could alter (promote or inhibit) the expression levels of various cytokines and chemokines and thereby contribute to the virus growth. (iii) In the third possibility, one could speculate a scenario where a constant and long-term exposure of the neighboring cells to the toxic effects of agnoprotein could lead to a diminished survival rate of the neighboring cells. Such toxicity by agnoprotein may have severe consequences in the central nervous system (CNS), such as cellular and/or neuronal loss (Fig. 5B), and thus perhaps lead to the initiation of neurological disorders, including Alzheimer’s disease. In fact, in support of such a scenario, it was previously demonstrated that agnoprotein negatively affects the secretion of a neuronal survival chemokine, CXCL5/LIX, from CG4 rat oligodendroglial cells (Merabova et al., 2008, 2012). Since these chemokines are related to the human CXCL5/ENA78 and CXCL6/GCP-2, it is conceivable to speculate that the released agnoprotein may have harmful effects on neuronal survival in CNS in long-term exposure cases and thus might lead to initiation of the neurological disorders.

4. Viral infections and Alzheimer’s disease development

Alzheimer’s disease (AD) is a debilitating neurodegenerative disease, characterized by memory deficits and cognitive decline (Ferrari and Sorbi, 2021; Knopman et al., 2021; Long and Holtzman, 2019). These deficits ultimately affect visuospatial orientation, speech, and motor functions of the patients. AD is the most common form of dementia of the aged population. Concerning the biochemical diagnosis, the extracellular deposition of amyloid beta (Aβ) plaques and intracellular accumulation of the Tau neurofibrillary tangles are the main hallmarks of the disease. The exact etiology of the disease has yet to be determined. There is still no effective cure for AD (Chen et al., 2017; Long and Holtzman, 2019). Thus, there is an unmet need to understand the disease’s possible root-causing etiology. Mounting evidence from the molecular and epidemiological studies suggests that specific infectious agents, including viruses, might be significant contributing factors to certain neurodegenerative diseases, including AD (Albaret et al., 2023; Levine et al., 2023; Readhead et al., 2018; Sipilä et al., 2021; Tejeda et al., 2023). Recent studies in this regard indicate that a human neurotropic virus, HSV-1, could play a role in the development of AD (Albaret et al., 2023; Itzhaki, 2014, 2017; Itzhaki et al., 1997, 1998; Jamieson et al., 1992; Mangold and Szpara, 2019; Marcocci et al., 2020; Protto et al., 2020; Readhead et al., 2018; Sait et al., 2021; Tejeda et al., 2023) due to its persistent neuronal infection in humans. Some investigators believe that the induction of both Aβ deposition and Tau tangles are part of natural innate immune responses to viral infections (Eimer et al., 2018; Kumar et al., 2016a, 2016b; Moir et al., 2018). However, it is possible that these aggregates gain toxic potential when dysregulated and thus significantly contribute to the progression of the disease (Rischel et al., 2022). It is also conceivable that other neurotropic viruses could be neurological disease risk factors. Thus, the concept of viral etiology should be investigated for involvement in the initiation of certain neurodegenerative diseases, including AD. The human neurotropic JCV is a potential candidate for such a role for several reasons, including (i) establishing latency in various organs (kidneys, specific immune cells and in the brains of individuals without PML), (ii) infecting various brain cells (oligodendrocytes, astrocytes and, and neurons), (iii) expressing a toxic protein, agnoprotein, with several interesting features such as being released from the infected cells and taken up by the bystander cells, through which JCV perhaps promotes its neurodegenerative activities other than PML.

Additionally, our recent work studying the effects of JCV agnoprotein on mitochondrion revealed that agnoprotein significantly dysregulates the metabolic and Ca²⁺ homeostatic functions of it (Saxena et al., 2021). These findings are consistent with the reports suggesting a link between the impairments in neuronal Ca²⁺ handling and Alzheimer disease development (Calvo-Rodriguez et al., 2020; Jadiya et al., 2019). Thus, it is plausible to speculate that a periodical but limited reactivation of JCV in the brain under certain conditions, such as stress, attenuated immune response, and immunosuppression, etc., could lead to cumulative damage to the brain cells and eventually may cause neurodegeneration (Fig. 5B). This possibility should be further investigated whether JCV infections contribute to the development of AD-like neurodegeneration. We propose that the agnoprotein-mediated toxic effects on the brain cells can be studied by employing in vitro tissue culture model systems and in vivo the brain-specific inducible agnoprotein knock-in mouse model systems.

5. Agnoprotein targets a wide range of viral and host cell proteins

The targeting of a number of viral and cellular proteins by JCV agnoprotein was previously reported by various laboratories. These targets include tubulin (Endo et al., 2003), JCV LT-Ag (Safak and Khalili, 2001), capsid protein, VP1 (Suzuki et al., 2012), p53 (Darbinyan et al., 2002), FEZ1 and HP1-α (Okada et al., 2005; Suzuki et al., 2005), JCV Sm t-Ag (Sariyer et al., 2008), adaptor protein complex 3 (AP-3) (Suzuki et al., 2013). Some of these interactions are implicated in various aspects of the viral life cycle, including in viral replication (Safak et al., 2001), virion formation (Sariyer et al., 2006; Suzuki et al., 2012), deregulation of cell cycle progression (Darbinyan et al., 2002) and functioning as a viroporin (Suzuki et al., 2010, 2013). Similarly, BKV agnoprotein has also been reported to contribute to a wide range aspects of the BKV life cycle, including, but not limited to, inhibition of viral DNA replication (Gerits et al., 2015), interfering with exocytosis (Johannessen et al., 2011), targeting the lipid droplets cells (Unterstab et al., 2010), evading innate immune components (Manzetti et al., 2020) and egress of virions from the infected cells (Panou et al., 2018).

In addition, our group further investigated cellular targets of agnoprotein by employing proteomic studies and thus revealing more exciting results. These proteomics studies showed that agnoprotein interacts with various cellular proteins containing “coiled-coil” motifs, and targets numerous protein networks, including but not limited to the ubiquitin degradation pathway, eukaryotic translation initiation factors complex, mitochondrial protein import complexes, tRNA synthetase complexes 26S proteasome complex subunits, nuclear import and export systems, nuclear pore complexes, and ER and Golgi vesicle trafficking pathway (Saribas and Safak, 2020).

6. Agnoprotein targets mitochondrion and modulates its functions

Viruses target a wide range of host components, including the hostprotein complexes, the network of signaling pathways, and organelles to create a favorable cellular environment to complete their replication cycle (Brito and Pinney, 2017; Franzosa and Xia, 2011). Mitochondria are one of the organelles targeted by viruses (Boya et al., 2004). Besides generating energy for the cells (Boyman et al., 2019), mitochondria are (Boyman et al., 2019) also involved in various cellular mechanisms, including initiation of apoptosis (Boya et al., 2003; Thomson, 2001), Ca²⁺ dynamics (Li et al., 2007) and antiviral immunity (Boya et al., 2003; Cloonan and Choi, 2012; Foy et al., 2005; Li et al., 2005, 2007; Thomson, 2001). Mitochondria harbor their own translational machinery and circular DNA genome; and encode 13 mitochondrial proteins including 2 rRNAs, 22 mitochondrial t-RNAs and respiratory chain complexes (I, III, IV and V) (Anand and Tikoo, 2013). The rest of the structural and regulatory proteins of mitochondria are nuclear DNA-encoded and imported (Anand and Tikoo, 2013). These imported proteins contain mitochondrial targeting sequences (MTS) located at either the internal or amino-terminal region of those proteins (Pfanner, 2000). The length of MTS varies depending on the protein type (10–80 aa long) and play an essential role in the import process. This MTS contains mostly positively charged amino acid residues, which are removed right after being imported into the mitochondrial matrix. The imported proteins are then distributed into the respective locations within the mitochondria depending on their functional properties (Pfanner, 2000; Wiedemann and Pfanner, 2017). Some viral proteins also have their own MTS, which mediates their mitochondria targeting. These viral proteins could then impact the various functions of mitochondria, including alteration in the mitochondrial stress due to the production of the reactive oxygen species (ROS), Ca²⁺ dynamics, mitochondrial membrane potential, and mitochondrial antiviral immunity (Anand and Tikoo, 2013; Boya et al., 2004).

As mentioned above, JCV agnoprotein contains several functional domains, including an MTS located at its N-terminus (MTS, aa 1–14, MVLRQLSRKASVKV), suggesting that it can mediate the import of agnoprotein into mitochondria and thus allows agnoprotein to alter mitochondrial functions. Our recent data indeed confirmed the agnoprotein targeting to the mitochondria (Saribas and Safak, 2020) and demonstrated that agnoprotein significantly but negatively affects mitochondrial respiration rates, mitochondrial membrane potential, and energy (ATP) production. We also observed a substantial increase in ROS production rates and alterations in Ca²⁺ influx parameters (Saribas and Safak, 2020). Functional consequences of such mitochondrial targeting by agnoprotein need to be further investigated to explore a possibility of whether there is any link between the initiation of some of the neurological disorders and long-term exposure of CNS to the toxic effects of agnoprotein.

7. Expression of microRNAs by polyomaviruses

Gene expression of microorganisms is regulated by various mechanisms, one of which is the use of small noncoding microRNAs (miRNAs) at the post-transcriptional level. This mechanism is widely employed by a variety of organisms, including protists, viruses, plants, and animals (He and Hannon, 2004). These types of RNA molecules are encoded in the introns of the cellular genes and most often regulate gene expression through binding to 3′-untranslated regions (UTRs) of the target messenger RNA (mRNA), causing their degradation, preventing them from being translated or preserving them for later translation (Iwakawa and Tomari, 2015; O’Brien et al., 2018). Long precursors miRNAs are produced in the nucleus by RNA polymerase II and III and cleaved into short precursor miRNAs (~70 nt) by Drosha and DGCR8 complex (Denli et al., 2004) and subsequently exported to the cytoplasm by exportin 5 complex, where they are further processed for maturation by Dicer, an RNAse III endonuclease, to produce a mature miRNA duplex (~20 nt) (Okada et al., 2009; Zhang et al., 2004). The matured miRNA then involves in intracellular regulation of the cellular genes or they are exported from the cells either through exosomes or apoptotic bodies to execute the same task extracellularly (Makarova et al., 2016; Turchinovich and Cho, 2014; Turchinovich et al., 2011; Zhang et al., 2015).

Like many other viruses, polyomaviruses also encode miRNAs, the first of which was reported by Sullivan et al., in 2005 for SV40 (SVmiRNA) (Sullivan et al., 2005). Following this report, miRNA expression was also reported several other polyomaviruses, including BKPyV (BKV), (Broekema and Imperiale, 2013), MCPyV, (Lee et al., 2011; Seo et al., 2009), MPyV (Sullivan et al., 2009) and racoon polyomavirus (RacPyV) (Chen et al., 2015; Dela Cruz et al., 2013). SV40 miRNA (SVmiRNA) shows complete complementarity to the viral early RNA sequences and thus suppresses LT-Ag expression by promoting their degradation. This miRNA also reduces the susceptibility of SV40-infected cells to the host cytotoxic T cells (Sullivan et al., 2005). Thus, SV40 miRNA helps SV40 to evade the host immune response and therefore most likely enhances successful completion of the viral life cycle. Similar to the SV40 case, JCV and BKV early coding regions generate similar types of miRNAs with similar functions as described for SV40 miRNA. That is, JCV and BKV miRNA downregulate the expression of the early coding region by degrading the early pre-mRNA species. Additionally, both JCV and BKV miRNAs were shown to promote the evasion of the innate and adaptive immune systems. A prime example of such an evasion is targeting a stress-induced ligand ULBP3 RNA expression by degradation. This ligand protein is recognized by natural killer cell receptor NKG2D to kill the infected cells. Thus, in the absence of the NKG2D receptor, the natural killer cells cannot effectively eliminate the infected cells and this results in immune evasion by JCV and BKV miRNA. (Bauman et al., 2011). Like JCV, BKV, and SV40, MCPyV also encodes two mature miRNAs from its early coding region, designated MCV-miR-M1-5p and MCV-mir-M1-3p (Lee et al., 2011; Seo et al., 2009). The MCV-miR-M1-5p is entirely complementary to the coding sequences of LT-Ag of MCPyV. Thus, it reduces the expression of LT-Ag and, therefore, limits viral replication (Seo et al., 2009). In another study, MCV-miR-M1-5p was found to target cellular Sp100 mRNA. Sp100 protein is implicated in the induction of the innate immune responses against dsDNA viruses, including MCPyV infection. This suggests that MCPyV uses its miRNA to regulate both the host and its early gene expression to evade the host immune system (Akhbari et al., 2018). In 2009, Sullivan et al., also reported the expression of two mature miRNA for mouse polyomavirus MPyV, (Sullivan et al., 2009). Like the miRNA encoded by the other polyomaviruses mentioned above, MPyV miRNA also targets the viral early transcripts for degradation. Although they are not required for the infection of the cultured cells (Sullivan et al., 2009), they instead play a role in promoting viruria during the acute phase of infection (Burke et al., 2018). Lastly, the expression of another polyomavirus miRNA was detected in the RacPyV-associated neuroglial brain tumors. In contrast to other polyomavirus miRNAs, RacPyV miRNA is abundantly expressed in tumor tissue (Brostoff et al., 2014) and is also expected to inhibit the early LT-Ag gene expression. However, it was found that LT-Ag mRNA levels were relatively high in those tumors, suggesting an unknown function for RacPyV miRNA (Brostoff et al., 2014; Chen et al., 2015; Dela Cruz et al., 2013).

8. Concluding remarks

In this review, we have summarized recent advances in the discovery and functional analysis of primarily small viral proteins expressed as a result of the cis- and trans-splicing events that take place within early and late coding region transcripts of polyomaviruses as well as the regulatory miRNAs that were encoded from the same regions. Recently, analysis of the alternatively spliced RNA products of these viruses demonstrated that they encode novel, sometimes unpredictable viral proteins, some of which were proven to play important regulatory roles in the viral replication cycle and cell transformation. Detailed functional analysis of these proteins is needed to understand their fundamental roles in viral replication cycle and cell transformation. It is now clear that several human polyomaviruses can cause various human diseases. For example, polyomavirus-associated nephropathy (PNV) is caused by BKV, PML is induced by JCV, Merkel cell carcinoma [a benign hyper-proliferative skin cancer] is associated with Merkel cell carcinoma virus (MCPyV) and Trichodysplasia Spinulosa (TS) is caused by TSPyV. The human polyomavirus-6 (HPyV6) has been implicated in the induction of a chronic inflammatory disorder called Kimura Disease (Nguyen et al., 2017; Rascovan et al., 2016; Schowalter et al., 2010; Schrama et al., 2014). The human polyomavirus 7 (HPyV7) was detected in pruritic rash cases in transplant recipients (Ho et al., 2015; Schowalter et al., 2010). Therefore, further investigation of the functional roles of the novel proteins expressed by these polyomaviruses is pivotal to our understanding of the progression of the diseases caused by these viruses. Some of these viruses are neurotropic, such as JCV. Thus, it is also important to explore whether there is link between expression of some of these viral proteins and induction of any of the neurological disorders in the infected individuals in long-term exposure cases. For instance, a small toxic regulatory protein expressed by two human polyomaviruses, JCV and BKV, called agnoprotein is a multifunctional viral protein that forms highly stable dimers and oligomers (Saribas et al., 2011); and targets a wide range of proteins, protein complexes and organelles in agnoprotein-positive cells (Saribas and Safak, 2020). It is also released from the infected cells and taken up by the bystander cells (Otlu et al., 2014; Saribas et al., 2018b) and alters mitochondrial functions (Saxena et al., 2021). Based on all these interesting and noteworthy functions, one can speculate that agnoprotein may have considerable negative effects on the brain cells of latently infected individuals including neurons, astrocytes, oligodendrocytes, and microglia in long-term exposure cases. It is also known that agnoprotein is constitutively expressed in the JCV-infected individuals in low amounts; but this low amount of expression can cause considerable harm to the CNS of those individuals and thus may contribute to the onset of age-related neurological disorders, including Alzheimer’s disease (Jackson, 2018), as is the widely accepted case for the association of this disease with HSV infection. Exploring the answers to such relevant questions may open new avenues for us to better understand age-related neurological disorders.