2. Chargaff Difference Analysis

Transcribed duplex DNA has an mRNA-synonymous strand and an mRNA-template strand. If transcription is to the right of the site of RNA polymerase initiation (promoter), the "top" strand (i.e. the sequence recorded in GenBank) is the mRNA-synonymous strand. Szybalski and coworkers (1966) showed that mRNA-synonymous strands (and hence the corresponding mRNAs) have purine-rich clusters. Combining this observation with Chargaff's second parity rule, it follows for mRNA-synonymous strands that purines in the clusters might be balanced by an equal number of dispersed pyrimidines, or that there might be small deviations from the second rule ("Chargaff differences") in favour of purines, as is indeed found.

    Chargaff differences are simply the differences between the numbers of the classical Watson-Crick pairing bases in a nucleic acid segment ("AT-skew", "GC-skew"). The sign of the differences depends on the direction of subtraction, which in some previous work was determined alphabetically. For some purposes Chargaff differences are best expressed as positive or negative base excesses, which may be combined to provide an index of the degree of purine-loading, with purine excesses scoring positive and pyrimidine excesses scoring negative (Lao & Forsdyke, 2000). If the ORFs (open reading frames) in a sequence are known, purine-loading indices may be calculated either directly from ORF base composition, or from codon-usage tables (Nakamura et al., 1999). The indices are then 1000[(A-T)/N] for the W bases (A and T) and 1000[(G-C)/N] for the S bases (G and C), where N is the total number of bases (i.e. N = W + S). These two values are then summed to obtain a value for the overall purine-loading index (bases per kb).

    Chargaff differences (absolute or %) may be calculated as A-T [ or as (A-T)/W], and as G-C [or as (G-C)/S]. Here, A, T, G and C can be the frequency of each base in 1 kb sequence windows. This approach makes no assumption about the disposition of ORFs, and can be applied to uncharted DNA. When an ORF is located, values for windows whose centres overlap the ORF can be averaged to obtain an approximate value for that ORF. For the importance of 1 kb window sizes and other details see Dang et al., (1998) and Bell and Forsdyke (1999).

    For AT-rich genomes, purine-loading is usually with respect to adenine, whereas for GC-rich genomes, purine-loading is usually with respect to guanine. In thermophilic bacteria, whatever the (G+C)%, purine-loading involves both purines (Lao & Forsdyke, 2000). An organism in which one or both Chargaff differences reflect the purine-loading of a significant excess of mRNAs is held to comply with "Szybalski's transcription direction rule" (Bell & Forsdyke, 1999).

FIG. 1. Distribution of purine-loading among species. The purine-loading of coding regions was calculated from codon usage tables for all species (2958) represented in the August 1999 release of the GenBank database by more than three genes and more than 2500 bases. The purine-loading index (bases/kb) for a particular species was calculated as the sum of 1000[(G-C)/N] and 1000[(A-T)/N], where G, C, A, and T correspond to the numbers of individual bases, and N corresponds to the total number of bases, in the codon usage table. This measure of the purine-loading of RNAs disregards 5' and 3' non-coding sequences, including poly(A) tails.

3. Extremes of Positive and Negative Purine-Loading

Indexes of purine-loading calculated from codon usage tables, although not taking into account 5' and 3' non-coding sequences, demonstrate the universality of the purine-loading phenomenon. Fig. 1 shows how purine-loading is distributed across all species for which sequences of more than 3 ORFs and 2500 bases were available. The value for all human genes (excluding mitochondria) is 42 bases/kb, meaning that, on average, there are 42 more purines than pyrimidines for every kilobase of coding sequence. The shoulder with negative purine-loading values (pyrimidine-loading) corresponds mainly to mitochondrial genes, which are disproportionately abundant in GenBank.

Distribution of purine-loading indices among 494 viruses of eukaryotes, excluding plants. Viruses at the extremes of the distribution are listed in the figure. For example, Brazilian caprine lentivirus (325) and Hepatitis D virus (321) fall in the same interval, and provide a frequency value of 2 at the extreme right of the distribution. Positive indices were observed for murine polyoma virus (32), RSV (54), Vaccinia virus (64), SFV-1 (73), and SV40 (103).

     Fig. 2 shows a subset of these data (494 species), which corresponds to all eukaryotic viruses, with the exclusion of plant viruses. Most viruses have positive purine-loading indices, which are often greater than the average for human ORFs (vertical dashed line). There is a tendency for members of certain viral groups to be either extremely pyrimidine-loaded or extremely purine-loaded. Among retroviruses, at one extreme (highly pyrimidine-loaded) are Human T-cell Leukaemia (Lymphotropic) Virus-1 (HTLV-1) and some similar retroviral species, whereas at the other extreme are Human Immunodeficiency Virus-1 (HIV-1) and some similar retroviral species (highly purine-loaded).

    While the value for the agent of classical acute infectious hepatitis (Hepatitis A virus; 37) is close to the human average (42), that for the agent of classical serum hepatitis (Hepatitis B virus), which produces a chronic infection and requires reverse transcriptase for replication, is highly negative (-127). Hepatitis virus C, which usually produces chronic infections is negative (-27). Hepatitis virus D, which requires coinfection by Hepatitis B virus in order to be packaged, is very positive (+321). Hepatitis virus E, which causes acute infections in humans and emerges periodically from an unknown source (and hence may be chronic in that source), is very negative (-140).

    Members of the Herpes virus group show less extreme Chargaff differences. The shoulder in Fig. 2 from -50 to -20 includes Herpes simplex virus-1 (HSV-1), Epstein-Barr virus (EBV) and human cytomegalovirus (HCMV), –all pyrimidine-loaded. Many other herpes-related viruses show overall purine-loading. The main ascending limb in the range + 10 to + 40 includes Varicella-Zoster virus (VZV), human herpesvirus 6 (HHV6), Herpes saimiri (HVS) and Ictalurid herpesvirus 1. Human herpesvirus 8 (HHV8; Kaposi sarcoma-associated virus) has a value of -1.

    Several species with negative purine-loading indices produce chronic infections (see Section 8) and have high (G+C)% in ORFs. In general, there is an inverse relationship between species (G+C)% and purine-loading index. Thus, a linear regression plot of average (G+C) percentages calculated from codon usage tables for each viral species used in Fig. 2, relative to the corresponding purine-loading indices, has a downward slope which is significantly different from zero (P<0.0001; data not shown). A more detailed analysis was made of the retroviral and herpes groups.

4. Extreme Purine-Loading Indices in Retroviruses

In many genomes transcription directions vary so that, while total base composition of the "top" strand of DNA reflects Chargaff's second parity rule (e.g. G = C), Chargaff differences for leftward transcribing ORFs (e. g. C>G) tend to cancel out the differences for rightward transcribing ORFs (e. g. G>C). Thus, with genome-sized sequence windows, compliance with Szybalski's transcription-direction rule, assessed as Chargaff differences, is not usually evident. However, retroviral genomes are transcribed entirely in one direction (considered by convention to be to the right), and mere examination of total base composition can illustrate some major points.

    Table 1 compares the "top" strands of four retroviruses whose (G+C) percentages vary from 39.2% (Simian foamy virus-1; SFV-1; Kupiec et al., 1991), to 54.4% (Rous sarcoma virus; RSV; Schwartz et al., 1983). Three of the genomes obey Szybalski's transcription direction rule for rightward transcription (purines>pyrimidines). For AT-rich genomes (SFV-1, HIV-1; Ratner et al., 1985), A>T and G>C with the W bases providing the largest deviation from the parity rule. Similarly, for the GC-rich RSV genome, A>T and G>C, with the S bases providing the largest deviation.

    Calculations for different parts of the sequences (1 kb windows) show that these purine-rich patterns are sustained throughout, particularly in the case of the base pairs with the largest deviations from the parity rule (data not shown). A similar compliance with Szybalski's rule has been noted in the case of some other AT-rich viruses (SV40, polyoma, vaccinia; Smithies et al., 1981; Bell & Forsdyke, 1999), again with the W bases providing the largest deviation. In contrast, HTLV-1 (Malik et al., 1988) is GC-rich, yet C>G. There is only a weak tendency for compliance with Szybalski's rule with respect to the W bases. The C-rich pattern is sustained throughout the genome, affecting all ORFs.

TABLE 2  Distribution of pyrimidine- and purine-loading among leftward- and rightward-transcribing ORFs of various herpes-related viruses^a

Direction of

transcription

To left

To right

W bases

S bases

W bases

S bases

Virus^b

C+G (%)

HVS

35.0

T>A

C>G

A>T

C>G

22:15 (P=0.020)

28:9 (P=0.00016)

35:5 (P<0.00003)

28:12 (P=0.0035)

VZV

46.1

T>A

C>G

A>T

C>G

19:15 (P=0.15)

27:7 (P=0.001)

24:13 (P=0.02)

25:12 (P=0.04)

EBV

60.1

T>A

G>C

T>A

C>G

23:21 (P=0.24)

30:14 (P=0.00007)

49:37 (P=0.08)

64:22 (P<0.00001)

HSV

68.3

A>T

G>C

T>A

C>G

20:17 (P=0.15)

29:8 (P=0.0007)

24:17 (P=0.037)

32:9 (P=0.00019)

5. Extreme Purine-Loading Indices in Herpes Viruses

Table 2 shows that, like HTLV-1, GC-rich members of the herpes virus family (EBV, HSV-1) do not follow Szybalski's transcription direction rule (i.e. the majority of mRNAs are pyrimidine-loaded). This applies strongly to the S bases (e. g. C>G when transcription is to the right, and G>C when transcription is to the left), but less so to the W bases.

    For most ORFs of AT-rich members of the Herpes virus family (HVS, VZV), A>T when transcription is to the right (i.e. they follow the transcription-direction rule), but the S bases do not follow the rule (C>G when transcription is to the right). When transcription is to the left, both the W and the S bases follow Szybalski's rule (pyrimidines>purines), but this is most evident in the case of the S bases. On balance, like AT-rich retroviruses (Table 1), the AT-rich herpes viruses seem to follow the rule. 

    From Table 2 it is evident that an individual ORF may be enriched for one of the W bases and one of the S bases, with four possible combinations (GA, GT, CA, CT). Table 3 shows that the ORFs of herpes viruses are distributed over all four groups, but with significant biases (bold number quartets with subscripts designating group). Rightward-transcribed ORFs which do not follow Szybalski's rule with respect to both the W and the S bases would be in the CT group. This group dominates in the cases of the two viruses of highest (G+C)% (EBV and HSV-1). Rightward-transcribed ORFs which follow Szybalski's rule with respect to both the W and the S bases would be in the GA group. This group is poorly represented in EBV and HSV-1. For EBV there are 37 rightward ORFs in the CT group, and only 10 in the GA group. Since assignment of functions to ORFs is not yet complete, whether the group biases relate to function remains for future study. That there can be a relationship is suggested by an ORF in the GA group, which encodes the EBNA-1 latency protein.

FIG. 3. Chargaff difference analysis of a section of the Epstein-Barr virus genome. ORFs are shown as open boxes with arrows indicating transcription direction. A, T, G, and C are the number of bases counted in 1 kb windows which were moved along the sequence in 0.1 kb steps. Each data point corresponds to the middle of a window. Chargaff differences (%) are expressed either as G - C (red squares), or as A - T (yellow circles). The major ORF encoding Epstein-Barr nuclear antigen-1 (EBNA-1) is labelled.

6. Gene Encoding the Major Latency Transcript Obeys the Rule

Fig. 3 shows Chargaff difference analysis of the section of the GC-rich Epstein-Barr virus (EBV) genome from which a major latency-associated transcript (encoding EBNA-1 protein) is derived. Whereas, like the majority of EBV genes, most neighbouring genes are pyrimidine-loaded (C>G when transcription is to the right; G>C when transcription is to the left), the rightward-transcribed gene encoding EBNA-1 protein follows Szybalski's rule (G>C; A>T), and very dramatically so.

    The EBNA-1-encoding gene is exceptional. It is the only viral gene expressed in the most fundamental type of EBV latency (the "EBNA-1 only program"; Thorley-Lawson et al., 1996). Among the other EBV latency-associated genes, those encoding EBNAs 2-6 purine-load their mRNAs only with respect to the W bases, and much less dramatically than the gene encoding EBNA-1; those encoding LMP1 and LMP2 pyrimidine-load their mRNAs with respect to both the W bases and the S bases (data not shown).

7. Simple Sequence Repeats Reinforce Compliance with Szybalski's Rule

It appears that, unlike most other EBV genes, the rightward-transcribed gene encoding EBNA-1 has been under pressure to accepted mutations which increase the purine content of the top (mRNA-synonymous) strand. If this were not possible without disrupting protein functional domains, the gene might have locally increased its content of purine-rich codons in inter-domain regions. Indeed, the EBNA-1 gene has a long "simple sequence" region (Karlin et al., 1988) containing exclusively either glycine codons (GGN), or alanine codons (GCN). Table 4 shows that choices of third bases (N) in these codons are almost exclusively purines (Karlin et al., 1990). Although the EBNA-1 gene (BKRF1) without the simple sequence is already slightly purine-loaded, the additional purines in the simple sequence greatly increase Chargaff difference values in favour of Szybalski's rule (Table 5).

TABLE 4 Codon usages^a of genes containing long simple sequences in Herpes simplex-related viruses

Virus

Gene

Codons^b

Complete

protein

Less the

simple

sequence

Simple

sequence

alone

Human

average^c

(%)

EBV

BKRF1
(EBNA-1)

Gly

GGG

63

11

52

24

GGA

144

43

101

25

GGT

25

24

1

16

GGC

19

19

0

34

Ala

GCG

4

3

1

10

GCA

85

2

83

23

GCT

6

6

0

26

GCC

8

8

0

40

HVS

ORF 48

Gly

GGG

34

2

32

24

GGA

55

7

48

25

GGT

1

1

0

16

GGC

6

6

0

34

Glu

GAG

92

10

82

58

GAA

75

29

46

42

ORF 73

Gly

GGG

2

2

0

24

GGA

15

7

8

25

GGT

1

1

0

16

GGC

1

1

0

34

Glu

GAG

2

1

1

58

GAA

128

3

125

42

Arg

AGG

2

2

0

20

AGA

17

11

6

20

CGG

1

1

0

21

CGA

1

1

0

11

CGT

4

3

1

8

CGC

1

1

0

19

Ala

GCG

2

2

0

10

GCA

8

6

2

23

GCT

43

6

37^d

26

GCC

0

0

0

40

VZV

ORF 11

Gly

GGG

6

6

0

24

GGA

28

16

12

25

GGT

11

11

0

16

GGC

5

5

0

34

Glu

GAG

58

14

44

58

GAA

30

30

0

42

Asp

GAT

20

19

1

47

GAC

35

16

19

53

Ala

GCG

31

11

20

10

GCA

19

19

0

23

GCT

13

13

0

26

GCC

13

13

0

40

TABLE 5 Contribution of long simple sequences to Chargaff differences of ORFs of Herpes simplex-related viruses

Chargaff difference formula^a

A - T

G - C

Virus

Gene

CpG

island

(local)^b

ORF length in codons

Compl-
ete
ORF

Less the
simple
sequ-
ence

Simple
sequ
ence
alone

Compl-
ete
ORF

Less
the simple
sequ-
ence

Simple
sequ-
ence
alone

Comp-
lete
ORF

Simple
sequ
-ence

EBV

BKRF1

-

642

238

A>T (141.2)

A>T (46.2)

A>T (95.0)

G>C (270.0)

G>C (86.2)

G>C (187.4)

HVS

ORF 48

-

797

301

A>T (135.9)

A>T (31.8)

A>T (104.1)

G>C (225.0)

G>C (25.9)

G>C (199.1)

ORF 73

-

407

183

A>T (270.3)

A>T (73.7)

A>T (196.6)

G>C (110.5)

C>G (-11.5)

G>C (122.0)

VZV

ORF 11

+

819

102

A>T (11.0)

T>A (-18.7)

A>T (29.7)

G>C (57.8)

G>C (5.3)

G>C (52.5)

HHV6

LJ1

+

321

117

T>A (-140.2)

T>A (-77.9)

T>A (-62.3)

G>C (145.4)

C>G (-36.3)

G>C (181.7)

    The ORF encoding the latency-associated nuclear antigen (LANA) of HHV8 also contains a long simple sequence repeat (Rainbow et al., 1997). Fig. 4 shows Chargaff differences in the region of the ORF. Being leftward-transcribed, it obeys Szybalski's rule in having an excess of pyrimidines in the top strand (C>G; T>A), whereas most neighbouring genes, whatever their transcription direction, disobey the rule. Thus, there is purine-loading of LANA mRNA which is reflected in codon choice (data not shown).

    Human herpes virus 6 (GenBank accession number X83413) has a 117 codon simple sequence in ORF LJ1, which displays purine-loading with respect to the S bases, and has a CpG island; such association with CpG islands (Table 5) is an expected feature of the promoter regions of latency-associated genes with hypomethylated CpGs (Honess et al., 1989; Tao et al., 1998).

    Intriguingly, there is a long region of repetitive DNA in an ORF of unknown function (ORF 50) in Ictalurid Herpesvirus 1 (channel catfish virus; GenBank M75136); here again, codon-choice (with respect to glycine, valine and alanine) suggests purine-loading (data not shown).

FIG. 4. Chargaff difference analysis of a section of the genome of Kaposi's sarcoma associated herpesvirus (HHV8). ORFs are numbered. Those with the prefix "K" may be unique to this virus. The major latency-associated nuclear antigen (LANA) is encoded by ORF 73. Note that, since transcription is to the left, the upper (template) strand of ORF 73 is pyrimidine-loaded, so that the lower (mRNA-synonymous) strand, and hence the corresponding mRNA, would be purine-loaded. Other details are as in Figure 3.

8. Disruption of Host Traffic

Viruses can have acute or chronic (persistent) patterns of host infection (Villarreal et al., 2000). Certain acutely lytic viruses (e.g. Hepatitis A, Vaccinia) purine-load their RNAs in compliance with Szybalski's rule, whereas viruses causing chronic (sometimes sub-clinical) infections tend to pyrimidine-load their RNAs (e.g. Hepatitis B; Bell & Forsdyke, 1999). Prolonged and profound clinical latency is a characteristic of some viruses that pyrimidine-load (Tables 1-3; Figure 3). In contrast to individuals latently infected with HIV-1, most individuals infected with HTLV-1 remain asymptomatic and live normal lives. Cytotoxic T cells appear able to target only peptides from the Tax protein (Gould & Bangham, 1998). Furthermore, HTLV-1 is likely to transfer between individuals when integrated in host DNA within intact cells. Virions alone show low infectivity. Similarly, Herpes simplex-related viruses permanently infect many individuals in their host species, who are often asymptomatic (Baer et al., 1984; Davison & Scott, 1986; McGeoch et al., 1988; Albrecht et al., 1992).

    Thus, certain persistent GC-rich viruses appear to risk interactions with host RNAs, which would be initiated through complementary base pairing between loops. In vivo, C-rich loops of virus RNAs would interact with G-rich loops of host RNAs, (just as in vitro poly(rC) interacts rapidly at low temperatures with mRNA-synonymous DNA strands; Szybalski et al., 1966). When in a functionally latent state most virus mRNAs would not be transcribed; thus the risk would be minimized. When triggered to move from the latent state to one of rapid productive cytolysis, the viruses would transcribe RNAs which, when released from the nucleus, might suddenly flood the cytosol with RNAs "driving on the wrong side of the road". The multiplicity of distracting loop-loop RNA interactions might slow host cell "traffic" and impair defence responses, including those triggered by any dsRNA which was formed. This "surprise" strategy might be of adaptive value to the virus.

9. The Role of Simple-Sequence Repeats in Latency-Associated Gene Products

In the functionally latent state most viral mRNAs would not be expressed, and so would not be available to interact with host mRNAs. However, although HTLV-1 has no latency-specific transcripts, most herpesviruses do. Remarkably, often herpesvirus transcripts include purine-biased simple sequence elements (see Section 7).

    Like those of other members of the herpes virus family, the EBV genome is very compact with little intergenic DNA; this suggests an evolutionary selection pressure to eliminate non-functional sequences. The long simple sequence (encoding Gly-Ala repeats) in the EBNA-1 gene has been explained by Karlin et al. (1988, 1990; Karlin 1995) as an adaptation operative at the protein level. However, the Gly-Ala region can be removed experimentally without affecting the known functions of EBNA-1 (Yates & Camiolo, 1988; Summers et al. 1997).

    The paradox appeared to be resolved by evidence that the Gly-Ala region functions in cis at the protein level to inhibit antigen processing for MHC presentation (Levitskaya et al., 1995, 1997; Mukherjee et al., 1998). However, in order to express the protein these authors used a vector which first had to express the corresponding mRNA; this was then translated into the protein. Their evidence is consistent with the Gly-Ala region being simply a device for purine-loading a foreign mRNA ("non-self") to make it appear like host mRNA ("self"); this might subvert intracellular self/not-self discrimination (Forsdyke, 1994; 1995a,b; 1999).

    If selection had been acting at the protein level to conserve the Gly-Ala region there should not have been such extreme codon bias (Table 4). On the other hand, the bias might have arisen by the amplification of an initially small Gly-Ala-coding segment, which just happened to have the purine bias. However, we note that in several other members of the Herpes virus family there are similar purine-biases in long (>100 amino acids) simple sequence-encoding regions within genes likely to be latency-associated (see Section 7). These biases might also be a consequence of a selection pressure for purine-loading of mRNA to assist maintenance of latency. For example, the LANA protein of HHV8 stabilizes latency by preventing p53-mediated apoptosis (Friborg et al., 1999).

10. Messenger RNAs as "Antibodies"

So why is EBNA-1 mRNA purine-loaded? Distracted by the messenger role of mRNA molecules, we may fail to note that the diverse spectrum of cell mRNA species, like the diverse spectrum of antibodies in serum, constitutes a repertoire of specificities with the potential to react with complementary sections of non-self RNA "antigens".

    Just as interactions between antibody and foreign antigen provoke extracellular inflammatory responses to the antigen, so interactions between host RNA and foreign RNA might provoke intracellular responses to foreign RNA, which could include gene silencing. If EBNA-1 mRNA ("sense") in latent EBV-infected cells were not purine-loaded to avoid "kissing" interactions, then it is possible that a self RNA species would have a sufficient degree of complementarity ("antisense") to progress beyond kissing interactions. Molecules of dsRNA of a length sufficient to alert host defence systems might then be formed (Susuki et al., 1999). The alarm might serve to increase MHC protein expression since only newly synthesized MHC proteins bind peptides efficiently for presentation to T cells (Townsend et al., 1990). Thus, through its purine-loading of EBNA-1 mRNA ("stealth" strategy; Cristillo et al., 1996), EBV would fail to provoke gene silencing or increase MHC expression. This would impede the MHC-dependent cytotoxic T cell response (Callan et al., 1998), and so assist maintenance of the latent state.

    There are reports of natural antisense RNAs derived from overlapping genes with different transcriptional orientations (Vanhee-Brossollet & Vaquero, 1998). In the light of the present thesis, such transcripts should not normally coexist in the same cell or intracellular compartment, or should be special cases for which there are adaptations to prevent the inadvertent firing, or response to, dsRNA alarms.

11. Charge cluster domains decrease immunogenicity of other domains?

The use of simple sequence to purine-load mRNAs means that, at the protein level, the simple sequences often contain runs of charged amino acids (e. g. Glu, Asp). Karlin (1995) refers to such regions as "hyper-charge runs", and notes that "for most of the hypercharge runs [in proteins] there is considerable variation in codon usage, which suggests an important function for these charge runs" (our parentheses and italics). However, our studies show that the variation is restricted to purine-rich codons (Table 4), which is more consistent with selection acting at the nucleic acid level.

    Many of the codons characteristic of the triplet expansion diseases, some of which generate charge runs, are also purine-rich (Green & Wang, 1994; Hancock & Santibanez-Koref, 1998). We suggest that charge runs themselves may not have an important function with respect to the function of the end-product (protein) of the ORF in which they locate (although they may affect protein solubility). When attempting to relate a protein's sequence to its biological function, the possibility that the major selective pressure has been at the nucleic acid level must be considered (Ball, 1973; Rocha et al., 1999; Lao & Forsdyke, 2000). As the result, a protein of less than optimum function may be synthesized, or the protein sequence may have to counter-adapt to improve function in the face of a primary selection pressure operating at the level of the corresponding nucleic acid (Forsdyke, 1996; Forsdyke & Mortimer, 2001).

    For example, to counter a tendency of its protein product to provoke autoimmune attack by cytotoxic T cells, there would be a selection pressure for a gene to purine-load its mRNAs, thus generating long charge-rich alpha-helices which might be irrelevant to the function of the protein itself (Dohlman et al., 1993). In this respect, we note the prevalence of charge clusters in antigens implicated in various autoimmune diseases. The clusters do not coincide with major autoantigenic epitopes (Brendel et al., 1991). This suggests that charge cluster domains may not be the primary cause of the diseases, as has been supposed, but may have evolved in response to the disease-provoking characteristics of other domains (i.e. the domains corresponding to autoantigenic epitopes).

    Unlike EBV, HSV-1 does not show CpG suppression (indicating no general methylation of CpGs), and there are no long simple sequence regions. However, HSV-1 would be expected to have pyrimidine-rich loops in most mRNAs (Tables 2, 3). Intriguingly, the main HSV-1 RNAs transcribed during latency correspond to the "antisense" strand (Croen et al., 1987; Stevens et al., 1987). We predict that any parts of these latency-associated transcripts which persist in the cell would be relatively purine-rich (Goldenberg et al., 1997; Arthur et al., 1998). A similar prediction applies to certain antisense transcripts found in EBV latency (Karran et al., 1992; Brooks et al., 1993).

FIG. 5. Fold analysis of a segment of the Epstein-Barr virus genome containing the EBNA-1-encoding gene (labelled box with horizontal arrow indicating transcription to the right). The region of the Gly-Ala repeat is marked by shading and two vertical dashed lines. FONS, values for the folding of the natural sequence. High negative FONS values (e.g. –40 kcal/mol) correspond to high folding potential (stem-loop potential; Forsdyke, 1998). FORS-M, values for the base composition-dependent component of the FONS values. FORS-D (upper plot with standard errors of the mean), values for the base order-dependent component of the FONS values (such that FONS = FORS-M + FORS-D).

12. Low Stem-Loop Potential of the Gly-Ala Repeat-Encoding Region

Simple sequence elements in proteins, as found in the trinucleotide expansion diseases (e.g. polyglutamine tracts encoded by poly(CAG)), sometimes cause intracellular protein aggregation. It is possible that such aggregates are responsible for the underlying pathology (Hancock & Santibanez-Koref, 1998; Forsdyke, 2000). However, trinucleotide repeats are often capable of adopting stem-loop conformations, which at the RNA level can activate PKR. This may contribute to the disease mechanism (Tian et al., 2000). In the case of the gene encoding EBNA-1, the region of the repeat has very low stem-loop potential, as revealed by a sustained low negative value for the folding of the natural sequence (FONS value for the mRNA-synonymous strand; Fig. 5). This is contributed to both by the base composition of the repeat (FORS-M value), and by its base order (FORS-D value; Forsdyke, 1998). In the corresponding mRNA, most parts encoding known functions of the protein may adopt compact secondary structures, whereas the part encoding the simple sequence repeat and the region on its immediate 3' side may have a structure more available for intermolecular interactions.

13. A Role for Non-Genic DNA?

The threshold for binding to PKR is approximately 15 trinucleotide repeats (45 bases), which corresponds to a dsRNA segment of approximately 22 bases (Tian et al. 2000). There are 4²² possible combinations in the universe of 22 base sequences, of which half (4²²/2) complements the other half. A virus encoding 10 mRNAs of average length 1021 bases, would have 10000 (approx. 4⁷) "windows" of 22 bases, any one of which could potentially act as an RNA "antigenic determinant" in the host cell. Assuming 10000 different host mRNA species in a cell, there would be 10000 x 1000 (approx. 4¹²) potential complementary RNA "windows" in host mRNAs. With a much higher mutation rate, a virus species might appear capable of adapting to ensure that its 4⁷ specificities did not complement the host's 4¹² specificities. Various factors militate against this.

    First, a high degree of polymorphism among host transcripts (Sunyaev et al. 2000) would make it likely that what a member of a virus species "learned" (through mutation) on one member of its host species, would not be applicable to the next member of the host species which it encountered (Forsdyke, 1991; Forsdyke, 2000, 2001). Second, due to a low level of read-through transcription (failure of transcription termination) of host mRNAs, a low level of transcription of extragenic DNA occurs (Heximer et al., 1998). Thus, the maximum potential repertoire of "RNA antibodies" would be limited only by genome size (approx. 4¹⁶ potential specificities in humans). Indeed, one function of the promoters of repetitive DNA elements (e.g. human Alu elements) might be to provide such read-through transcription, as has been observed (Manley & Colozzo, 1982; Feuchter et al., 1992). It would be of adaptive advantage to the host to activate such promoters under conditions of cell stress (heat shock or viral infection); again, this is observed (transcription by RNA polymerase III; Jang & Latchman, 1989; Liu et al., 1995).

    Thus, non-genic "junk" DNA (Dang et al. 1998) can be viewed in much the same way as we view the diverse genes encoding the variable regions of immunoglobulin antibodies. Just as B-cells capable of synthesizing a unique anti-self antibody would be eliminated during somatic time to prevent self-reactivity, so junk DNA would be screened over evolutionary time (by positive selection of individuals in which favourable mutations had been collected together by recombination) to decrease the probability of two complementary "self" transcripts interacting to form dsRNA segments of more than 21 bases. High polymorphism of non-genic DNA (Beck et al. 1996; Nickerson et al. 1998) would make it difficult for viruses to anticipate the RNA "antibody" repertoire of future hosts (Forsdyke, 1999, 2000, 2001).

   We thank J. Gerlach for assistance with computer configuration, J. T. Smith for statistical advice, and G. McPherson for assistance with the Silicon Graphics Computer maintained by Base4 BioInformatics Inc., Mississauga. The National Research Council of Canada, Academic Press, Cold Spring Harbor Laboratory Press, and Elsevier Publishing Corporation gave permission for the display of full-text versions of some of the cited references at our internet site.

REFERENCES

ALBRECHT, J. C., NICHOLAS, J., BILLER, D., CAMERON, K. R., BIESINGER, B., NEWMAN, C., WITTMANN, S., CRAXTON, M. A., COLEMAN, H., FLECKENSTEIN, B. & HONESS, R. W. (1992). Primary structure of the Herpes saimiri genome. J. Virol. 66, 5047-5058.

ARTHUR, J. L., EVERETT, R., BRIERLEY, I. & EFSTATHIOU, S. (1998). Disruption of the 5' and 3' splice sites flanking the major latency-associated transcripts of herpes simplex virus type 1: evidence for alternative splicing in lytic and latent infections. J. Gen. Virol. 79, 107-116.

BAER, R. et al., (1984). DNA sequence and expression of the B95-8 Epstein-Barr virus genome. Nature 310, 207-211.

BALL, L. A. (1973). Secondary structure and coding potential of the coat protein gene of bacteriophage MS2. Nature New Biol. 242, 44-45.

BECK, S., ABDULLA, S., ALDERTON, R. P., GLYNNE, R. J., GUT, I. G., HOSKING, L. K., JACKSON, A., KELLY, A., NEWELL, W. R., SANSEAU, P., RADLEY, E., THORPE, K. L. & TROWSDALE, J. (1996). Evolutionary dynamics of non-coding sequences within the class II region of the human MHC. J. Mol. Biol. 255, 1-13.

BELL, S. J., CHOW, Y. C., HO, J. Y. K. & FORSDYKE, D. R. (1998). Correlation of CHI orientation with transcription indicates a fundamental relationship between recombination and transcription. Gene 216, 285-292.

BELL, S. J. & FORSDYKE, D. R. (1999). Deviations from Chargaff's second parity rule correlate with direction of transcription. J. theor. Biol. 197, 63-76.

BOSSI, L. & ROTH, J. R. (1980). The influence of codon context on genetic code translation. Nature 286, 123-127.

BRENDEL, V., DOHLMAN, J., BLAISDELL, B. E. & KARLIN, S. (1991). Very long charge runs in systemic lupus erythematosus-associated autoantigens. Proc. Natl. Acad. Sci. USA 88, 1536-1540.

BROOKS, L. A., LEAR, A. L., YOUNG, L. S. & RICKINSON, A. B. (1993). Transcripts from the Epstein-Barr virus BamH1 A fragment are detectable in all three forms of virus latency. J. Virol. 67, 3182-3190.

BULL, J. J., JACOBSON, A., BADGETT, M. R. & MOLINEUX, I. J. (1998). Viral escape from antisense RNA. Mol. Microbiol. 28: 835-846.

CALLAN, M. F. C., TAN, L., ANNELS, N., OGG, G. S., WILSON, J. D. K., O'CALLAGHAN, C. A., STEVEN, N., MCMICHAEL, A. J. & RICKINSON, A. B. (1998). Direct visualization of antigen-specific CD8⁺ T cells during the primary immune response to Epstein-Barr virus in vivo. J. Exp. Med. 187, 1395-1402.

CROEN, K. D., OSTROVE, J. M., DRAGOVIC, L. J., SMIALEK, J. E. & STRAUS, S. E. (1987). Latent Herpes simplex virus in human trigeminal ganglia. Detection of an immediate early gene "anti-sense" transcript by in situ hybridization. New Eng. J. Med. 317, 1427-1432.

CRISTILLO, A. D. (1998). Characterization of G₀ /G₁ switch genes in cultured T lymphocytes. Ph.D. Thesis. Queen's University, Kingston, Ontario.

CRISTILLO, A. D., HEXIMER, S. P. & FORSDYKE, D. R. (1996). A "stealth" approach to inhibition of lymphocyte activation by oligonucleotide complementary to the putative G₀/G₁ switch regulatory gene G0S30/EGR1/NGFI-A. DNA Cell Biol. 15, 561-570.

CRISTILLO, A. D., LILLICRAP, T. P. & FORSDYKE, D. R. (1998). Purine-loading of EBNA-1 mRNA avoids sense-antisense "collisions". FASEB. J. 12, A1453. Abstract #828. Click Here [Added note Feb 2008: In 1997 Tim Lillicrap assisted these studies, which were mainly performed by Tony Cristillo. The manuscript was rejected by several journals before finding a home. However, the delay was useful in that relevant studies of James Mortimer and Isabelle Barrette could be added.]

DANG, K. D., DUTT, P. B. & FORSDYKE, D. R. (1998). Chargaff differences correlate with transcription direction in the bithorax complex of Drosophila. Biochem. Cell Biol. 76, 129-137.

DAVISON, A. J. & SCOTT, J. E. (1986). The complete DNA sequence of Varicella-Zoster virus. J. Gen. Virol. 67, 1759-1815.

DOHLMAN, J. G., LUPAS, A. & CARSON, M. (1993). Long charge-rich alpha-helices in systemic autoantigens. Biochem. Biophys. Res. Comm. 195, 686-696.

EGUCHI, Y., ITOH, T. & TOMIZAWA, J. (1991). Antisense RNA. Annu. Rev. Biochem. 60, 631-652.

EHRENFELD, E. & HUNT, T. (1971). Double-stranded poliovirus RNA inhibits initiation of protein synthesis by reticulocyte lysates. Proc. Natl. Acad. Sci. USA 68, 1075-1078.

ELIA, A., LAING, K. G., SCHOFIELD, A., TILLERAY, V. J. & CLEMENS, M. J. (1996). Regulation of the double-stranded RNA-dependent protein kinase PKR by RNAs encoded by a repeated sequence of the Epstein-Barr virus genome. Nucleic Acids Res. 24, 4471-4478.

FEUCHTER, A. E., FREEMAN, J. D. & MAGER, D. L. (1992). Strategy for detecting cellular transcripts promoted by human endogenous long terminal repeats: identification of a novel gene (CDC4L) with homology to yeast CDC4. Genomics 13, 1237-1246.

FIRE, A. (1999). RNA-triggered gene silencing. Trends Genet. 15, 358-363.

FORSDYKE, D. R. (1991). Early evolution of MHC polymorphism. J. theor. Biol. 150, 451-456.

FORSDYKE, D. R. (1994). Relationship of X chromosome dosage compensation to intracellular self/not-self discrimination: a resolution of Muller's paradox? J. theor. Biol. 167, 7-12.

FORSDYKE, D. R. (1995a). Entropy-driven protein self-aggregation as the basis for self/not-self discrimination in the crowded cytosol. J. Biol. Sys. 3, 273-287.

FORSDYKE, D. R. (1995b). Fine tuning of intracellular protein concentrations, a collective protein function involved in aneuploid lethality, sex determination and speciation? J. theor. Biol. 172, 335-345.

FORSDYKE, D. R. (1996). Different biological species "broadcast" their DNAs at different (G+C)% "wavelengths. J. theor. Biol. 178, 405-417.

FORSDYKE, D. R. (1998). An alternative way of thinking about stem-loops in DNA. A case study of the human G0S2 gene. J. theor. Biol. 192, 489-504.

FORSDYKE, D. R. 1999. Heat shock proteins as mediators of "danger" signals: implications of the slow evolutionary fine-tuning of sequences for the antigenicity of cancer cells. Cell Stress Chaperones 4: 205-210.

FORSDYKE, D. R. (2000). Double-stranded RNA and/or heat-shock as initiators of chaperone mode switches in diseases associated with protein aggregation. Cell Stress Chaperones 5, 375-376.

FORSDYKE, D. R. (2001). The Origin of Species, Revisited. Montreal: McGill-Queens University Press.

FORSDYKE, D. R. & MORTIMER, J. R. (2000). Chargaff's legacy. Gene 261, 127-137.

FRIBORG, J., KONG, W-P., HOTTINGER, M. O. & NABEL, G. J. (1999). P53 inhibition by the LANA protein of KSHV protects against cell death. Nature 402, 889-894.

GOLDENBERG, D., MADOR, N., BALL, M. J., PANET, A. & STEINER, I. (1997). The abundant latency-associated transcripts of Herpes simplex virus type 1 are bound to polyribosomes in cultured neuronal cells and during latent infection in mouse trigeminal ganglia. J. Virol. 71, 2897-2904.

GOULD, K. G. & BANGHAM, C. R. M. (1998). Virus variation, escape from cytotoxic T lymphocytes and human retroviral persistence. Sem. Cell Devel. Biol. 9, 321- 328.

GREEN, H. & WANG, N. (1994). Codon reiteration and the evolution of proteins. Proc. Natl. Acad. Sci. USA 91, 4298-4302.

HAMILTON, A. J. & BAULECOMBE, D. C. (1999). Role of a species of small   antisense RNA in post-transcriptional gene silencing in plants. Science 286, 950-951.

HANCOCK, J. M. & SANTIBANEZ-KOREF, M. F. (1998). Trinucleotide expansion diseases in the context of micro- and mini-satellite evolution. EMBO. J. 17, 5521-5524.

HEXIMER, S. P., CRISTILLO, A. D., RUSSELL, L. & FORSDYKE, D. R. (1998). Expression and processing of G₀/G₁ Switch Gene 24 (G0S24/TIS11/TTP/NUP475) RNA in cultured human blood mononuclear cells. DNA Cell Biol. 17, 249-263.

HONESS, R. W., GOMPELS, U. A., BARRELL, B. G., CRAXTON, M., CAMERON, K. R., STADEN, R., CHANG, Y.-N. & HAYWARD, G. S. (1989). Deviations of expected frequencies of CpG dinucleotides in Herpesvirus DNAs may be diagnostic of differences in the states of their latent genomes. J. Gen. Virol. 70, 837-855.

HUNTER, T., HUNT, T., JACKSON, R. J. & ROBERTSON, H. D. (1975). The characteristics of inhibition of protein synthesis by double-stranded ribonucleic acid in reticulocyte lysates. J. Biol. Chem. 250, 409-417.

IZANT, J. G. & WEINTRAUB, H. (1984). Inhibition of thymidine kinase gene expression by anti-sense RNA: a molecular approach to genetic analysis. Cell 36, 1007-1015.

JANG, K. L. & LATCHMAN, D. S. (1989). HSV infection induces increased transcription of Alu repeated sequences by RNA polymerase III. FEBS. Lett. 258, 255-258.

KARLIN, S. (1995). Statistical significance of sequence patterns in proteins. Curr. Opin. Struct. Biol. 5, 360-371.

KARLIN, S., BLAISDELL, B. E., MOCARSKI, E. S. & BRENDEL, V. (1988). A method to identify distinctive charge configurations in protein sequences, with application to human herpesvirus polypeptides. J. Mol. Biol. 205, 165-177.

KARLIN, S., BLAISDELL, B. E. & SCHACHTEL, G. A. (1990). Contrasts in codon usage of latent versus productive genes of Epstein-Barr virus: data and hypothesis. J. Virol. 64, 4264-4273.

KARRAN, L., GAO, Y., SMITH, P. R. & GRIFFIN, B. E. (1992). Expression of a family of complementary-strand transcripts in Epstein-Barr virus-infected cells. Proc. Natl. Acad. Sci. USA 89, 8058-8062.

KUMAR, M. & CARMICHAEL, G. G. (1998). Antisense RNA: function and fate of duplex RNA in cells of higher eukaryotes. Microbiol. Mol. Biol. Rev. 62, 1415-1434.

KUPIEC, J. J., KAY, A., HAYAT, M., RAVIER, R., PERIES, J. & GALIBERT, F. (1991). Sequence analysis of the simian foamy virus type 1 genome. Gene 101, 185-194.

LAO, P. J. & FORSDYKE, D. R. (2000). Thermophilic bacteria strictly obey Szybalski's transcription direction rule and politely purine-load RNAs with both adenine and guanine. Genome Res. 10, 228-236.

LEVITSKAYA, J., CORAM, M., LEVITSKY, V., IMREH, S., STEIGERWALD-MULLEN, P. M., KLEIN, G., KURILLA, M. G. & MASUCCI, M. G. (1995). Inhibition of antigen processing by the internal repeat region of the Epstein-Barr virus nuclear antigen-1. Nature 375, 685-688.

LEVITSKAYA, J., SHARIPO, A., LEONCHIKS, A., CIECHANOVER, A. & MASUCCI, M. G. (1997). Inhibition of ubiquitin/proteasome-dependent protein degradation by the Gly-Ala repeat domain of the Epstein-Barr virus nuclear antigen 1. Proc. Natl. Acad. Sci. USA 94, 12616-12621.

LIU, W-M., CHU, W-M, CHOUDARY, P. V. & SCHMID, C. W. (1995). Cell stress and translational inhibitors transiently increase the abundance of mammalian SINE transcripts. Nucleic Acids Res. 23, 1758-1765.

MALIK, K. T., EVEN, J. & KARPAS, A. (1988). Molecular cloning and complete nucleotide sequence of an adult T cell leukaemia virus/Human T cell leukaemia virus type I isolate of Caribbean origin. J. Gen. Virol. 69, 1695-1710.

MANLEY, J. L. & COLOZZO, M. T. (1982). Synthesis in vitro of an exceptionally long RNA transcript promoted by an AluI sequence. Nature 300, 376-379.

MARCUS, P. (1983). Interferon induction by viruses: one molecule of dsRNA as the threshold for induction. Interferon 5, 115-180.

MCGEOCH, D. J., DALRYMPLE, M. A., DAVISON, A. J., DOLAN, A., FRAME, M. C., MCNAB, D., PERRY, L. J., SCOTT, J. E. & TAYLOR, P. (1988). The complete DNA sequence of the long unique region in the genome of Herpes simplex virus type 1. J. Gen. Virol. 69, 1531-1574.

MELTON, D. A. (1985). Injected antisense RNAs specifically block messenger RNA translation in vivo. Proc. Natl. Acad. Sci. USA 82, 144-148.

MEYER, R. K. & KRUEGER, D. D. (1994). Minitab Computer Supplement. Macmillan College Publishing, New York.

MITTELSTEN SCHEID, O. 1999. New tool for Swiss army knife. Nature 397, 25.

MUKHERJEE, S., TRIVEDI, P., DORFMAN, D. M., KLEIN, G. & TOWNSEND, A. (1998). Murine cytotoxic T lymphocytes recognize an epitope in an EBNA-1 fragment, but fail to lyse EBNA-1-expressing mouse cells. J. Exp. Med. 187, 445-450.

NAKAMURA, Y., GOJOBORI, T. & IKEMURA, T. (1999). Codon usage tabulated from the international DNA sequence databases: its status 1999. Nucleic Acids Res. 27, 292.

NICKERSON, D. A., TAYLOR, S. L., WEISS, K. M., CLARK. A. G., HUTCHINSON, R. G., STENGARD, J., SALOMAA, V., VARTIAINEN, E., BOERWINKLE, E. & SING, C. (1998). DNA sequence diversity in a 9.7-kb region of the human lipoprotein lipase gene. Nature Genet. 19, 233-240.

RAINBOW, L., PLATT, G. M., SIMPSON, G. R., SARID, R., GAO, S-J., STOIBER, H., HERRINGTON, C. S., MOORE, P. S. & SCHULZ, T. F. (1997). The 222- to 234-kilodalton latent nuclear protein (LNA) of Kaposi's sarcoma-associated herpesvirus (human herpes virus 8) is encoded by orf73 and is a component of the latency-associated nuclear antigen. J. Virol. 71, 5915-5921.

RATNER, L. et al. (1985). Complete nucleotide sequence of the AIDS virus. Nature 313, 277-284.

ROBERTSON, H. D. & MATHEWS, M. B. (1996). The regulation of the protein kinase PKR by RNA. Biochimie 78, 909-914.

ROCHA, E. P. C., DANCHIN, A. & VIARI, A. (1999). Universal replication biases in bacteria. Mol. Microbiol. 32, 11-16.

SCHWARTZ, D. E., TIZARD, R. & GILBERT, W. (1983). Nucleotide sequence of Rous sarcoma virus. Cell 32, 853-869.

SHARP, P. (1999). RNAi and double-strand RNA. Genes Devel. 13, 139-141.

SMITHIES, O., ENGELS, W. R., DEVEREUX, J. R., SLIGHTOM, J. L. & SHEN, S. (1981). Base substitutions, length differences and DNA strand asymmetries in the human Gl and Al fetal globin gene region. Cell 26, 345-353.

STEVENS, J. G., WAGNER, E. K., DEVI-RAO, G. B., COOK, M. L. & FELDMAN, L. T. (1987). RNA complementary to a herpesvirus a gene mRNA is prominent in latently infected neurons. Science 235, 1056-1059.

SUMMERS, H., FLEMING, A. & FRAPPIER, L. (1997). Requirements for Epstein-Barr nuclear antigen 1 (EBNA-1)-induced permanganate sensitivity of the Epstein-Barr latent origin of DNA replication. J. Biol. Chem. 272, 26434-26440.

SUNYAEV, S. R., LATHE, W. C., RAMENSKY, V. E. & BORK, P. (2000). SNP frequencies in human genes: an excess of rare alleles and differing modes of selection. Trends Genet. 16, 335-337.

SUZUKI, K., MORI, A., ISHII, K. J., SAITO, J., SINGER, D. S., KLINMAN, D. M., KRAUSE, P. R. & KOHN, L. D. (1999). Activation of target-tissue immune-recognition molecules by double-strand polynucleotides. Proc. Natl. Acad. Sci. USA 96: 2285-2290.

SZYBALSKI, W, KUBINSKI, H, & SHELDRICK, P. (1966). Pyrimidine clusters on the transcribing strands of DNA and their possible role in the initiation of RNA synthesis. Cold Spring Harbor Symp. Quant. Biol. 31, 123-127.

TAO, Q., ROBERTSON, K. D., MANNS, A., HILDESHEIM, A. & AMBINDER, R. F. (1998). The Epstein-Barr virus major latent promoter Qp is constitutively active, hypomethylated, and methylation sensitive. J. Virol. 72, 7075-7083.

THORLEY-LAWSON, D. A., MIYASHITA, E. M. & KAHN, G. (1996). Epstein-Barr virus and the B cell: That's all it takes. Trends Microbiol. 4, 204- 207.

TIAN, B., WHITE, R. J., XIA, T., WELLE, S., TURNER, D. H., MATHEWS, M. B. & THORNTON, C. A. (2000). Expanded CUG repeat RNAs form hairpins that activate the double-stranded RNA-dependent protein kinase PKR. RNA 6, 79-87.

TOWNSEND, A., ELLIOTT, T., CERUNDULO, V., FOSTER, L., BARBER, B. & TSE, A. (1990). Assembly of MHC class 1 molecules analysed in vitro. Cell 62, 285-295.

VANHEE-BROSSOLLET, C. & VAQUERO, C. (1998). Do natural antisense transcripts make sense in eukaryotes? Gene 211, 1-9.

VILLARREAL, L. P., DEFILIPPIS, V. R. & GOTTLIEB, K. A. (2000). Acute and persistent viral life strategies and their relationship to emerging diseases. Virology 272, 1-6.

YATES, J. L. & CAMIOLO, S. M. (1988). Dissection of DNA replication and enhancer activation functions of Epstein-Barr virus nuclear antigen 1. Cancer Cells 6, 197-205.

End Note 2001 (not in original paper): A key observation in this paper is that the EBNA-1 protein is longer than necessary because of the need to contain a low complexity GLY-ALA repeat region between functional domains. This allows purine-loading at the nucleic acid level. Essentially the same observation was reported by Pizzi and Frontali in Genome Research (2001) 11, 218-229. They noted:

End Note May 2002 on large scale non-genic transcription

Supporting Section 13 Kapranov et al (Science 296, 916-919) report large scale, not necessarily genic, transcriptional activity by human chromosomes 21 and 22.

End Note Jan 2006 on autoimmunity

Supporting Section 11 Carl, Temple and Cohen (2005; Arthritis Research & Therapy 7, R1360-74) note that "most nuclear systemic autoantigens are extremely disordered proteins." The disordered regions have a distinctive composition of amino acids, many of which correspond to purine-rich codons, but the disordered regions themselves are not particularly antigenic. If such regions arose as products of purine-loaded insertions into parts of genes corresponding to inter-domain regions of protein, then they would be expected to be disordered relative to the functional regions of the protein.

End Note June 2006 on trinucleotide expansions

In Section 12 the possibility that the underlying pathology in various trinucleotide expansion diseases reflected a gain-of-function (purine-loading) mutation at the RNA level. When this idea was presented at the 1998 EMBO Workship on Trinucleotide Expansion Diseases (see below), a member of the audience pointed out that, although many trinucleotide expansions would result in an increased purine-content of the corresponding mRNA, some expansions involve pyrimidines (e.g. repeats of CTG). To explain this the ad hoc postulate was made that in such cases there would exist an antisense transcript (e.g. repeats of CAG). Recently evidence for this has been found in two independent cases. 

1. Cho, D. H. et al. (2005) Mol. Cell 20, 483-489. Antisense transcription and heterochromatin at the DM1 CTG repeats are constrained by CTCF.

2. Moseley, M. L. et al. (2006) Nature Genet. 38, 758-769. Bidirectional expression of CUG and CAG expansion transcripts and intranuclear polyglutamine inclusions in spinocerebellar ataxia type 8.

End Note Feb 2008 on Samuel Karlin

Papers of Samual Karlin (1924-2007) and his "biometric" associates are cited frequently above. The New York Times published his obituary (see Globe & Mail Toronto 25 Feb 2008). Having made many contributions to pure mathematics, in the 1960s he turned to population genetics in the mode of Ronald Fisher and Sewall Wright. There were productive interactions with one of Fisher's students, Walter Bodmer, who had visited Stanford to work with Joshua Lederberg. At age 65 (circa 1989) Karlin contributed to the development of Stephen Altschul's "Blast" method for aligning sequences ("basic local alignment search tool").

End Note July 2008. Support for our hypothesis

Our observation of the extreme codon bias in favour of purines in the Glycine-Alanine repeat domain showed that the domain exerted a major influence at the nucleic acid level. The work (Cristillo, Lillicrap & Forsdyke 1997) was stonewalled by anonymous reviewers before eventual acceptance (see above note in reference list). A decade later our work was supported by Tellam et al. (2008) who showed that decreasing the purine-loading without changing the encoded protein enhanced the expression of fragments of the EBNA1 protein complexed with MHC protein at the cell surface - thus triggering a T-cell response.

Zalumbide et al. (2007) had attributed the immunosuppressive power of the repeat sequence in the LANA-1 encoding gene of KSHV to function at the protein level. But I pointed out (End Note, Nov 2010) that their actual data implied RNA-level function in this respect. Kwun et al. (2007) had noted the similarities in the EBNA-1 and LANA-1 repeats at the mRNA level, indicating a common evolutionary origin, but still maintained that function would be at the protein level. Tellam et al. (2012) have now formally repeated "Nature's experiment" and changed the reading frame of the EBNA-1 repeat so that the protein sequence of the repeat was drastically altered, while the mRNA sequence remained essentially unchanged (still heavily purine-loaded). Since this did not change the host cell's ability to decrease immune recognition, they "conclude that the cis-inhibitory effect of the internal repeat sequences of gammaherpesviruses operates at the nucleotide level and is unlikely to be mediated through the direct action of the GAr [Gly and Ala rich] polypeptide." While not excluding "a more general immune evasive strategy", Tellam et al. (2012) propose that "an unusual RNA secondary structure within the repeat region may interfere with translation of the EBNA1 mRNA by inhibiting ribosome transit through the purine-rich sequence, thereby leading to a reduction in the levels of EBNA1 such that the infected cell evades the normal host immune surveillance mechanisms."

Tellam JT, Lekieffre L, Zhong J, Lynn DJ, Khanna R (2012) Messenger RNA sequence rather than protein sequence determines the level of self-synthesis and antigen presentation of the EBV-encoded antigen, EBNA-1. PLOS Pathogens 8, issue 12, e1003112.

The interpretation of Calis et al. (2010; see note above) was questioned in an accompanying paper (Lavasseur & Pontarotti 2010). There were other studies (e.g. Khrustalev & Barkovsky 2011), that also focused on GC% (rather than the inversely related AG%). Using various programs that reveal the antigenic parts of various alphaherpesviruse proteins (epitopes) that stimulate B cells to make antibodies, there was a correlation of high GC% of a gene (i.e. low AG%) with the immunogenicity of its protein product. Thus, high AG% (low GC%) correlates with low immunogenicity. Of course, within the host cell a virus potentially displays both its nucleic acids and its proteins. Although it may be parts of the proteins that are eventually displayed to the host's extracellular immune system, we argue (above) that it is viral mRNA that, by purine-loading, avoids triggering the intracellular defences that first lead to peptide presentation by MHC complexes.

The elegant Yewdell and Fahraeus papers have again been brought into question by work from the Rock laboratory (see 2012 above). Their conclusions (Colbert et al. 2013) are now less restrained: "For the constructs we have analysed, mature functional proteins, rather than defective ribosomal products, are the predominant source of MHC class I presented peptides."

And the story continues: "The proposal that MHC class I molecules preferentially present peptides that come from errors in protein biosynthesis or failures of successful folding (DRiPs), is an attractive idea, and one that has influenced general conceptions of how the immune system monitors for non-native gene products. However, on close examination, the data that actually support this proposal are still limited, often indirect, and based on questionable assumptions. ...  Additionally, recent data indicate that for many antigens, most of the MHC class I-presented peptides probably originate from functional mature molecules, and not defective, newly synthesized proteins. The immune system thus appears to monitor the peptides that are generated through proteasomal breakdown of both short-lived and stable cellular proteins. However, further research is necessary to quantify the relative importance of the newly synthesized and mature proteins in generating the bulk of MHC class I-presented peptides and whether DRiP-like mechanisms may function in certain cases."

Tellam, Khanna and their coworkers report that the number and arrangement of G-residues in the purine-loaded loops of the Gly-Ala repeat regions of EBNA-1 mRNA are such as to facilitate G-quadruplex formation, thus slowing down mRNA translation, which is likely to decrease the concentration of the EBNA-1 protein: "We identified putative G-quadruplex sequences within the ORF of EBV-encoded EBNA1 mRNA and confirmed that these motifs form intramolecular parallel G-quadruplex structures in the context of long transcripts. Remarkably, similar G-quadruplex motifs were also mapped within the conserved purine-rich repetitive sequences of viral ORFs encoding numerous other gammaherpesviral maintenance proteins, suggesting that this class of virus may exploit RNA G-quadruplex structure to downregulate their maintenance protein expression levels to escape immune recognition. ... In the present study, we define the role of RNA G-quadruplex motifs within purine-rich ORFs and demonstrate their ability to impede ribosomal activity during translational elongation. We have shown in this study that G-quadruplex structures can alter the association of ribosomes with mRNAs by inducing premature termination and/or ribosome stalling and dissociation events."

    These observations are consistent with the idea that the intracellular self/not-self discrimination process that precedes MHC-peptide presentation avoids presentation of "self" by decreasing both (i) dsRNA (achieved by R-loading self mRNAs), and (ii) specific aggregation (achieved by synthesising self protein at a rate insufficient to allow specific protein aggregates). EBV adopts this mechanism for the key EBNA-1 mRNA, the G-quadruplexes of which slow protein synthesis. Thus, to the hypothesis that failure to form dsRNA would prevent a "receptive state" in MHC for peptide acceptance, should be added effects at the translational level that might impede peptide formation.

That prior signalling at the nucleic acid level may be necessary for the emergence of pathology in the polyglutamine expansion diseases (see section 12 of paper) is suggested by Gusella and the Huntington's Disease Consortion. For detailed discussion see Chapter 14 of Evolutionary Bioinformatics (Forsdyke 2016).

The LANA protein's acidic domains that can "read" the basic non-acetylated domains of other proteins, may include p53 among their targets (1). Viral long-term persistence (latency) may depend on this mechanism: "LANA's acidic domain reader is critical for viral latency," and thus "this LANA sequence may exert its effects on viral persistence through p53." Other proteins that, like p53, have one basic amino acid domain, may be targeted by one or more of the LANA protein's three dispersed acidic amino acid domains. However, "the basis for LANA's extended acidic domain reader sequence is unclear," so it could act "as a multivalent receptor for one or more unacetylated partners."

This mechanism could explain the domain's ability to "inhibit cis MHC class I peptide presentation," which would protect latent virus from host immune attack (1). Here the authors invoke "central repeat domain effects on protein processing," so interfering with the generation of LANA peptides that would complex with host MHC proteins (2,3). However, the LANA protein domain's acidic amino acids may be merely passive "placeholders" that exists by default because the domain, in fact, operates at the level of the mRNA that encodes the protein (4-7).

Evidence for this hypothesis was obtained by experimentally changing the reading frame (by addition or subtraction of a single base). Thus, the protein sequence of the domain was drastically altered while the mRNA sequence remained essentially unchanged. Since this did not affect viral ability to decrease host immune recognition, it was concluded "that the cis-inhibitory effect of the internal repeat sequences of gamma-herpesviruses operates at the nucleotide level and is unlikely to be mediated through the direct action of the - polypeptide" (6). While the latter considered the EBNA1 protein of Epstein-Barr virus (EBV), 'Nature' herself has carried out the analogous experiment with the evolutionarily related KSHV. Thus, citing Zaldumbide et al. (8), it was noted that: "While EBNA1 and LANA1 have no amino acid similarity, the gene sequences of the two viral repeat sequences are highly homologous to each other but are frameshifted for translation" (2).

Since the acidic amino acid-encoding domains are "extended" throughout the length of the LANA mRNA (1,7), some general function is indicated. This would be consistent with the hypothesis that the corresponding codons - rich in both purines (A and G) - "purine-load" the mRNA to decrease the probability of chance interactions with host mRNAs that are generally purine-rich. Since purines pair poorly with purines, this would decrease the probability of generating double-stranded RNA regions to activate intracellular alarms that can influence mRNA translation (4). Thus, decreasing the purine-loading, without changing the encoded protein, enhances cell surface expression of peptides complexed with MHC protein - so triggering a T-cell response (4-7). Evolutionarily, this would be the primary role of the acidic amino acid-encoding domains. Naming the LANA sequence "reader" presumes it actually "reads" rather than acts as a default space filler - the actual role being at the mRNA level. But other functions, perhaps involving p53, could have arisen secondarily.

For more details please see endnotes at: http://wayback.archive-it.org/7641/20161129204729/http://www.queensu.ca/academia/forsdyke/EBV.htm