Full text of DOI:10.1086/315886

1
515
CONCISE COMMUNICATION
The Amino Acid Sequence of the PKR-eIF2a Phosphorylation Homology
Domain of Hepatitis C Virus Envelope 2 Protein and Response to Interferon-a
1
,a
3
3
1
Alexandra Cochrane, Alexander Orr, Megan L. Shaw,
Department of Infectious Diseases, Brownlee Centre,
2
2
2,3
and Department of Gastroenterology, Gartnavel General Hospital,
and Division of Virology, Institute of Biomedical and Life Sciences,
Peter R. Mills, and Elizabeth A. B. McCruden
3
University of Glasgow, Glasgow, Scotland
A region of the hepatitis C virus (HCV) envelope 2 protein, the protein kinase, PKR and
early initiation factor 2a phosphorylation homology domain (PePHD), may be important in
interferon (IFN)–a resistance. The PePHD was amplified by polymerase chain reaction and
sequenced, and the amino acid sequence derived from pretreatment serum of 14 genotype
3
–infected patients with a range of responses to IFN-a therapy. Only 1 patient had a PePHD
variant. IFN-resistant PePHD variants present at low titers in pretreatment serum should be
selected by therapy; therefore, the PePHD amino acid sequence was also obtained from serum
collected during or after treatment in 5 patients with breakthrough or relapse of HCV RNA
positivity. No difference was found between the pre- and posttreatment PePHD sequences.
Thus, it appears that pretreatment sequencing of the PePHD would not enable clinicians to
predict the treatment response. There was no evidence that IFN therapy exerts selection
pressure in this region.
Hepatitis C virus (HCV) is a major cause of chronic liver
genotype 1 than in genotype 2 or 3. Constructs containing the
E2 gene of HCV genotype 1 can block the action of PKR in
yeast and human cell lines, and this is abrogated by mutation
of the PePHD, to resemble that of HCV genotypes 2 and 3. It
was suggested that the PePHD-PKR interaction could explain
the relative IFN resistance of genotype 1, compared with that
of other genotypes.
Although clinical data confirm that HCV genotype 3 is more
IFN sensitive than genotype 1 [3], a significant proportion of
genotype 3–infected patients do not have a sustained response
[1]. We hypothesized that within genotype 3, IFN-resistantstrains
may carry a PePHD variant similar to that of HCV genotype 1.
If such a variant were identified in pretreatment samples, it would
provide a powerful prognostic tool for clinicians. To investigate
this, we sequenced the PePHD from 14 genotype 3–infected pa-
tients with a range of treatment outcomes. Since HCV exists as
a quasi species [4], failure to demonstrate a PePHD variant in
pretreatment samples could simply reflect its presence as a mi-
nority. If so, the resistant variant would be selected by treatment
and would be present at higher titers in samples collected during
or after treatment. To capitalize on this selection pressure, we
also sequenced the PePHD in serum samples of 5 patients col-
lected during or after treatment.
disease; 80% of those infected become long-term carriers, with
attendant risks of cirrhosis and hepatocellular carcinoma. The
mainstay of treatment is interferon (IFN)–a, which, when com-
bined with ribavirin, results in sustained viral clearance in 38%–
3% of patients treated [1]. This poor response rate, combined
with the high cost of IFN, has fueled research into the mech-
anism of HCV IFN resistance.
IFN-a is a cellular protein that acts by inducing the tran-
scription of several antiviral genes. One such gene expresses the
protein kinase PKR (double-stranded RNA–activated protein
kinase), which phosphorylates the translation initiation factor
eIF2a, thereby blocking protein synthesis. Recently, a 12-aa
sequence was identified within the HCV envelope 2 (E2) protein
4
(
aa 659–671 of the HCV polyprotein), which shows homology
to both the PKR and eIF2a phosphorylation sites. The se-
quence was called the PKR-eIF2a phosphorylation homology
domain (PePHD) [2]. The degree of homology is greater in
Received 30 May 2000; revised 31 July 2000; electronically published 9
October 2000.
The study was approved by the ethics committee of the West Glasgow
University Hospital National Health Service Trust, and informed consent
was obtained from all patients.
Financial support: Scottish Office (K/MRS/50/C2492); Wellcome Trust
(
046745/Z/96).
a
Present affiliation: Laboratory of Clinical and Molecular Virology, Uni-
versity of Edinburgh, Edinburgh, Scotland.
Patients and Methods
Reprints or correspondence: Dr. E. A. B. McCruden, Institute of Virology,
Church St., Glasgow G11 5JR, Scotland (e.mccruden@bio.gla.ac.uk).
Patients. Eighty patients infected with HCV were offered IFN-
a monotherapy if liver biopsy showed changes typical of chronic
infection. Exclusion criteria included age 170 years, other cause
for chronic hepatitis, decompensated cirrhosis, human immuno-
The Journal of Infectious Diseases 2000;182:1515–8
᭧
2000 by the Infectious Diseases Society of America. All rights reserved.
0022-1899/2000/18205-0030$02.00

1
516
Cochrane et al.
JID 2000;182 (November)
deficiency virus infection, significant comorbidity, injection drug
use within the last year, or contraindication to IFN-a. Each patient
was prescribed IFN-a–2a (Roferon; Roche Products, Lewes, UK)
at 6 million units (MU) three times weekly for 12 weeks. Patients
negative for HCV after 12 weeks were continued on 3 MU thrice
weekly for another 36 weeks (total dose, 540 MU over 48 weeks).
Serum samples were collected before treatment and after 2, 4, 8,
Table 1. Time of withdrawal (during or after treatment) of serum
in relation to interferon (IFN)–a therapy.
Time of posttreatment
Patient
Response to IFN-a
failure sample
4
5
6
8
9
Breakthrough
Breakthrough
Breakthrough
Relapse
16 Weeks after end of treatment
While receiving IFN-a (week 36)
While receiving IFN-a (week 20)
6 Weeks after end of treatment
18 Weeks after end of treatment
1
2, 48, and 72 weeks and were stored at Ϫ70ЊC.
Relapse
Only 3 of 36 patients with genotype 1 infections had sustained
responses to treatment, whereas 8 of 35 with genotype 3 had sus-
tained responses. We studied serum samples from 16 genotype
NOTE. Breakthrough was defined as hepatitis C virus (HCV) RNA negative
at 12 weeks and positive by 48 weeks; relapse was defined as HCV RNA negative
at 48 weeks and positive by 72 weeks.
3
–infected patients with sufficient remaining sera. Definition of
outcome was as follows: sustained responder, HCV RNA negative
at 48 and 72 weeks; relapser, HCV RNA negative at 48 weeks and
positive by 72 weeks; breakthrough, HCV RNA negative at 12
weeks and positive by 48 weeks; and nonresponder, HCV RNA
positive at 12 weeks. Detection of HCV RNA was by in-house
nested reverse transcription–polymerase chain reaction (RT-PCR)
0
0
YTGYAACATCTAGGGG-3), 2503–2522, and ACSQ3 (5-AYG-
0
AGGAYRACGAACTCCCA-3) of HCV genotype 3a [7]. Se-
quences were translated and were aligned by use of TRANSLATE,
PILEUP, and PRETTY programs in the GCG package (Genetics
Computer Group, Madison, WI) [8].
[5] or by Amplicor (Roche Diagnostics, Lewes, UK). Among pa-
tients with sustained responses and in those who did not respond,
only pretreatment serum samples were studied. Among 5 patients
who experienced a relapse or breakthrough, a serum sample taken
during or after treatment was also analyzed (table 1). For simplicity,
these serum samples are referred to as posttreatment samples.
Amplification of HCV E2 RNA by RT-PCR. To minimize the
introduction of sequence error during processing, serum samples
were separated and were frozen within 4 h of sampling, and pre-
cautions to avoid contamination were adopted throughout PCR
experiments [6]. Complementary DNA was synthesized using Su-
perscript II reverse transcriptase (Gibco BRL, Paisley, UK), ac-
cording to the manufacturer’s instructions. The initial primer used
Results
Treatment outcome was assigned according to the results of
viral RNA testing by a PCR method sensitive to 2000 RNA
copies/mL [5]. A satisfactory nucleotide sequence equivalent to
aa 638–704 was obtained from 14 of the 16 patients identified.
The amplification system used (Advantage plus) contains a
0
0
polymerase that provides 3r5 proofreading, and the Taq Start
antibody allows for automatic “hot start” PCR. These measures
minimize the production of inaccurate sequences and reduce
nonspecific amplification products. The response to therapy
and the derived amino acid sequences for each patient are
aligned in figure 1. The PePHD in genotype 3 is from aa
665–677. The Genbank accession numbers for the nucleotide
sequences are AF289520–AF289538. Only 1 patient (a non-
responder) had a PePHD variant, a single amino acid change
toward the genotype 1b sequence. Therefore, there was no con-
sistent difference between sustained virological responders and
nonresponders, and there was no difference in the PePHD se-
quence between pretreatment and posttreatment samples. The
0
0
was MS2 (5-TGAYAAGGTAAAGAAGCCG-3). Two samples
were successfully reverse transcribed (patients 2 and 14), but owing
0
to a high failure rate, a new primer was designed: AC5 (5-CRA-
0
TCCARTGCTTATACCA-3). After RT, 1 U of ribonuclease H
(RNase H; Roche Molecular Biochemicals, Lewes, UK) was added
on ice and then incubated at 37ЊC for 20 min. The 2 samples reverse
transcribed with MS2 were amplified using Taq polymerase (Gibco
0
BRL) with MS2 and JM3E2TPAHCV (5-CCGGGAATTCTT-
0
0
GGATCCCACACATATACCACCGG-3), followed by MS4 (5-
0
0
GCAGGCATGGGCGTGAA-3) and MSQ7 (5-GTCTGCGGC-
CCTGTGTACTG-3). The rest were amplified with a PCR kit
3
0 amino acids downstream of the PePHD were identical in all
0
study samples; the upstream sequence showed some variation
but did not correlate with treatment response or with time of
sample collection.
(Advantage plus cDNA) for first-round PCR and then with Taq
polymerase with Taq Start antibody (both from Clontech, Basing-
stoke, UK) for the second round. The following primers were used:
0
0
AC6 (5-CAGCTGYAAGCCCATCAC-3) and AC5 (see above),
0
0
followed by AC7 (5-ACCRTAYTGCTGGCACTA-3) and AC8
5-GCMACRCACACGCGTGCGTC-3). (Full details of reac-
0
0
Discussion
(
tions, including thermal cycling conditions, are available by
request.)
Gel purification and silica column extraction. We subjected 95
mL of PCR product to electrophoresis on an 0.8% agarose TAE
gel. DNA was extracted from gel slices with silica gel columns
In this study, only 1 of 14 patients infected with HCV ge-
notype 3 had a PePHD variant. The pretreatment and post-
treatment failure PePHD sequences were identical in all patients
tested. This study was primarily designed to assess the value
of the PePHD sequence as a prognostic marker for clinical use.
For this reason, the HCV RNA was extracted from serum
samples, a source readily available to clinicians, and attention
was paid to the reliability of patient data and the accuracy of
the sequencing. We are confident that patient compliance was
(Recovery DNA purification kit II; Hybaid, Ashford, UK).
Nucleotide sequencing and analysis. Sequencing was performed
with an ABI prism 377 automated sequencer (Perkin-Elmer, Bea-
consfield, UK). Sense and antisense primers used were located,
0
respectively, at nucleotide positions 2038–2061, ACSQ4 (5-CCC-

JID 2000;182 (November)
E2 Protein of HCV and IFN
1517
Figure 1. Alignment of derived amino acid (aa) sequences of part of envelope 2 (E2) protein for 14 hepatitis C virus (HCV) genotype 3–infected
patients. Pre- and posttreatment failure sequences are shown for 5 patients. Protein kinase, PKR and early initiation factor 2a phosphorylation
homology domain (aa 665–667), denoted in boldface, is compared with published sequences for genotypes 1b [9] and 3a [7]. Numbering of aas
is as for genotype 3a. Dashes indicate aa identical to consensus sequence. BT, breakthrough; NR, nonresponder as defined in Patients and Methods;
Pre, pretreatment sample; Post, after or during treatment sample (exact timing given in table 1); REL, relapser; SR, sustained responder.
good, since treatment was supervised by a single specialist nurse
whose records agree with those of the hospital pharmacy.
In the second part of the study, we probed for evidence of
PePHD involvement in IFN resistance by looking for selection
of PePHD variants during therapy. To maximize the likelihood
of finding a difference between the pretreatment and posttreat-
ment sequences, we elected to study patients who had HCV
disease breakthrough or relapse. In these patients, the initial
virological response to therapy suggests that the dominant pre-
treatment species is IFN sensitive. Subsequent virological re-
bound may represent replication of an IFN-resistant variant
that was previously present as a minority. The posttreatment
serum samples were drawn after an average of 40 weeks of IFN
therapy. Given the short turnover time of HCV (half-life,
representative of the dominant species, comes from previousstud-
ies in our laboratory, in which change in the dominant species,
demonstrated by single-stranded conformation polymorphism,
corresponded with change in the consensus sequence [5].
In a study of 7 genotype 3–infected patients, there was no
correlation between PePHD sequence and response to IFN [11].
Of interest, 1 “responder” had a PePHD variant that resembled
genotype 1. Assuming that this patient had a sustained re-
sponse, this shows that the variant alone is insufficient to confer
IFN resistance. Unfortunately, the term “responder” was not
defined, nor were the times of serum sample withdrawal related
to IFN therapy. The PePHD was sequenced from hepatocytes
rather than from serum samples. This difference is unlikely to
be important because, although there is often a difference in
quasi-species complexity between HCV isolates from these 2
tissues, the consensus sequence is usually identical [12]. Another
study [13] of the PePHD in 33 genotype 3a–infected patients
treated with IFN alone or in combination with ribavirin showed
that the majority who achieved a sustained response had no
amino acid differences from the consensus. Hydrophobic mu-
tations occurred only in 4 of 16 sustained responders, but 1 of
these differences (Q668L) was found in a nonresponder in the
current study. This and the small numbers of differences seen
overall suggest a random finding.
100–182 min) [10], this should have allowed time for selection
of IFN-resistant variants. It is possible that these variants may
be less “fit” than other strains and therefore could rapidly lose
dominance after withdrawal of the IFN selection pressure.
Thus, the lack of PePHD change in patients 5 and 6 is partic-
ularly important, since the posttreatment failure sample was
drawn while IFN therapy was ongoing.
We derived PePHD sequences for each serum sample from a
single PCR product. Therefore, our sequence represents a con-
sensus of all the variants that make up the quasi species in the
serum samples. Support for the use of a consensus sequence, as
The low frequency of PePHD variants, combined with the

1
518
Cochrane et al.
JID 2000;182 (November)
2
3
. Taylor DR, Shi ST, Romano PR, Barber GN, Lai MMC. Inhibition of the
interferon inducible protein kinase PKR by HCV E2 protein. Science
lack of correlation with response, indicates that pretreatment
sequencing of the PePHD is unlikely to give any useful infor-
mation to clinicians. IFN-a resistance in any individual patient
is probably the result of a combination of host and viral factors.
Our results do not preclude a role for the PePHD-PKR inter-
action, but the lack of selection of a PePHD variant during
IFN-a therapy suggests that it was not the primary resistance
mechanism in the 5 patients tested. Similar studies of the NS5a
IFN sensitivity–determining region (ISDR) have produced con-
flicting results, and there is some evidence that a region down-
stream of the original ISDR may be important [14]. We se-
quenced the E2 gene 50 aa upstream and 30 aa downstream
of the PePHD without discovering any correlation with IFN-
a response.
1999;285:107–10.
. Davis GL, Lau JY. Factors predictive of beneficial response to therapy of
hepatitis C. Hepatology 1997;26:122S–7S.
4. Martell M, Esteban J, Quer J, et al. Hepatitis C virus (HCV) circulates as a
population of different but closely related genomes: quasi-species nature
of HCV genome distribution. J Virol 1992;66:3225–9.
5
6
7
. McKechnie VM, McCruden EAB, Mills PR. The NS5a gene of hepatitis C
virus in patients treated with interferon. J Med Virol 2000;60:367–78.
. Kwok S, Higuchi R. Avoiding false positives with PCR. Nature 1989;339:
237–8.
. Sakamoto M, Akahane Y, Tsuda F, Tanaka T, Woodfield DG, Okamoto H.
Entire nucleotide sequence and characterization of a hepatitis C virus of
genotype V/3a. J Gen Virol 1994;75:1761–8.
8
. Wisconsin package. Version 10.1. Madison, WI: Genetics Computer Group,
2000.
9
. Kato N, Hijikata M, Ootsuyama Y, et al. Molecular cloning of the human
hepatitis C virus genome from Japanese patients with non-A, non-B hep-
atitis. Proc Natl Acad Sci USA 1990;87:9524–8.
Acknowledgments
Automated nucleotide sequencing was performed by Lesley Taylor
Division of Virology, IBLS, University of Glasglow). We thank Eliz-
10. Ramratnam B, Bonhoeffer S, Binley J, et al. Rapid production and clearance
of HIV-1 and HCV assessed by large volume plasma apheresis. Lancet
(
1999;354:1782–5.
abeth Spence (Gartnavel General Hospital, Glasglow), who supervised
the treatment of the patients, Carol-Anne Deans (Division of Virology,
IBLS, University of Glasglow) for management of the specimen da-
tabase, and D. Kennedy (Gartnavel General Hospital, Glasglow) for
granting study leave for experimental work.
1
1
1. Abid K, Quadri R, Negro F. Hepatitis C virus, the E2 envelope protein, and
a-interferon resistance. Science 2000;287:1555.
2. Cabot B, Esteban J, Martell M, et al. Structure of replicating HCV quasi
species in the liver may not be reflected by analysis of circulating HCV
virions. J Virol 1997;71:1732–4.
1
3. Sarrazin C, Kornetzky I, Ruster B, et al. Mutations within the E2 and NS5A
protein in patients infected with hepatitis C virus type 3a and correlation
with treatment response. Hepatology 2000;31:1360–50.
References
1
. McHutchison JG, Gordon SC, Schiff ER, et al. Interferon a-2b alone or in
combination with ribavirin as initial treatment for chronic hepatitis C. N
Engl J Med 1998;339:1485–92.
14. Sarrazin C, Berg T, Lee JH. Mutations in the protein kinase–binding domain
of the NS5A protein in patients infected with hepatitis C virus type 1a
are associated with treatment response. J Infect Dis 2000;181:432–41.