Formation and survival of virus variants.

Bruce Phillips bap at MED.PITT.EDU
Wed Feb 7 14:43:16 EST 1996

	Because of the ongoing discussion concerning RNA replicases/
reverse transcriptase as the cause of RNA virus variation and in the spirit
of stimulating some interesting discussion, I am submitting the following
contribution.  Here goes:

                    STABILITY OF RNA VIRUSES
          A.  RNA replicases as well as reverse transcriptases have
error-prone rates of 10-4 X 10-5 per base per genome replication
event.  Thus, all RNA viruses, and NOT just Lentiviruses, have the
same potential for creating virus mutants. or variants.
          B.  For purposes of simplification, the only variants
considered here are those that can escape efficient antibody-
mediated neutralization because "escape" mutants: (i) have a
selective advantage for growth, (ii) are important in terms of
viral pathogenesis, particularly for HIV infections.
          C.  Most RNA viruses, e.g., measles virus, poliovirus
types 1-3, have a limited number of serotypes and those serotypes
breed true.  Even Rhinoviruses, with >112 serotypes, breed true. 
To the best of my knowledge, repeated passage of human rhinovirus
type 2 (HRV2) does not give rise to other serotypes. Thus, for
practically all RNA viruses, it is not possible to create in the
laboratory new serotype variants, e.g., with polioviruses, measles,
etc.  This apparently holds true for non-lentivirus retroviruses
such as HTLV-I and avian leukoviruses strains.
          D.  HIV, and other lentiviruses, are exceptional in that
they generate very large numbers of viable, serotypically unique
virus particles, thus constantly escaping host immune responses.
          A.  The potential for creating large numbers of virus
variants is approximately (within an order of magnitude) the same
          B.  Lentiviruses are UNIQUE in their capacity NOT BY
VIRTUE OF THEIR ENGINE (i.e., their reverse transcriptase) FOR
GENERATING VARIANTS but rather in their tolerance for enormous
serotype variation among progeny virions.  Thus, HIV possesses a
unique mechanism for tolerating structural variation in infectious
virus particles.
          C.  For nonenveloped viruses, the constraint on
structural variation lies in their capsid proteins, which must 
encapsidate viral genomes (virus assembly), provide a stable shell
for the genome,  allow release of the particle from the cell, and
function to permit attachment and entry into susceptible cells to
reinitiate virus replication.  A serious compromise in any of these
functions places the virion at a severe disadvantage in terms of
survival.  It follows that only changes in serotype determinants
consistent with the multiple functions of the capsid are permiss-
ible.  In virtually all cases, constraints in serotype determinants
severely limit capsid protein sequence variation.
          D.  For enveloped viruses, such as measles or influenza
viruses, constraints must exist for the number of forms of the
relevant surface proteins consistent with virus viability.  Within
a given viral subtype, e.g., all H1 strains, the flu HA protein
possesses a very finite number of altered forms consistent with
either (i) assembly of HA in the envelope, or (ii) ability of HA to
find cell receptor sialic acid moieties and mediate virus entry
into uninfected cells.  Since there is only one measles serotype,
it is fair to postulate that serotype determinants are rigidly
constrained such that serotype variants are not viable.
          E.  In contrast to other viruses, including other
retroviruses, the structure of the lentivirus surface proteins,
particularly the SU protein, gp120, allows for relatively large
variations in neutralization sites while still maintaining a
capacity to assembly properly and recognize cellular receptors to
facilitate virus entry into uninfected cells.  Therefore, it is the
unique ability of gp120 to tolerate amino acid sequence variation
and still perform its functions that lies at the heart of HIV
variability and not the error-proneness of its RT (which is
actually no different than the polymerases of other RNA viruses).
          F. As pointed out by a colleague, one additional source
allowing for variation in virus structure consistent with viral
viability is the extent to which that virus's cellular receptor
tolerates variation in its ligand.  In the same sense that viral
structural protein function must tolerate mutations in order to
produce viable virions, one can hypothesize that the cellular
receptor for a given virus can tolerate considerable variation in
a viral protein's conformation (e.g., gp120) and still allow for
virus attachment and penetration.

Let's hear it!

More information about the Virology mailing list