ISOLATION 101: THE BASICS
By Paul Philpott
Reappraising AIDS June, July, Aug. 1997
Isolating a virus
To isolate a virus, scientists take a heterogeneous sample (fluid from
a patient or a culture) and add it to a graduated density gel, which they
spin in a centrifuge. The contents of the sample settle into separate piles,
or bands , at different depths according to their characteristic
densities. These bands are called density-purified samples .
Because all microbiological entities have characteristic densities,
scientists can obtain density-purified samples that contain only certain
viruses, and no other material. There is only one way to confirm this:
a photograph with an electron microscope that contains nothing but
identical virus-looking objects.
If the micrograph reveals contaminating entities, that means the sample
contained some material that had the same density as the virus-looking
objects. In that case, scientists would have to add additional steps to
the isolation process, ones that purify based on other characteristics--like
size, or electrical affinity--until they could produce a sample that contained
only the virus-looking objects. However, this usually is not necessary.
Density purification typically produces true isolates of virus-looking
objects.
If the density, appearance, and size of these objects match those of
a previously characterized virus, scientists can label the sample a
virus isolate. If not, scientists must subject the sample (actually,
a fresh sample, since electron microscopy destroys whatever it photographs)
to a battery of tests to prove that the virus-looking objects are viruses.
Proving an isolate consists of viruses
Looking like a virus is just one feature of a virus. To be a virus, virus-looking
objects must behave like a virus, and their constituents must relate to
each other in special ways. Scientists demonstrate these criteria by adding
an isolate of virus-looking objects to a culture of suitable cells. If
the isolate consists of viruses, they will infect the cells and multiply
to numbers much greater than those present in the original isolate.
Scientists confirm this by attempting to re-isolate the virus-looking
objects from the culture after enough time has passed for substantial viral
replication to have taken place. The new isolate should form at the same
density as the original, and contain objects that look the same as those
in the original sample. But the new isolate should consist of a much thicker
band, indicating a larger number of viruses.
Scientists also have to examine the constituent molecules of the isolate.
Among other things, they have to confirm that the DNA or RNA codes for
all the proteins.
This being the case, scientists declare that the objects are indeed
viruses, and that these viruses are characterized by a certain size, shape,
and appearance, and consisting of a particular number of proteins and genetic
molecules of certain molecular weights or base pair lengths.
Isolating viral constituents
To isolate the contents of a virus, scientists must dismantle the viruses
into their constituent molecular parts. They do this by adding a special
detergent, SDS, to a viral isolate. The isolate will then consist of the
individual molecules that compose the viruses. These molecules include
proteins that decorate and line the outer membrane envelope, the globs
that form the hollow inner core, and the contents of the inner core: enzymes
and DNA or RNA.
Next scientists separate these molecular species from each other, using
electrophoresis ,, whereby an electric field pulls the molecules
through a gel so that they band according to their weights (instead
of densities). Some of the bands will contain proteins, and others contain
genetic material, either RNA or DNA.
Scientists call an electrophoresed sample a Western blot if they
are considering the bands that contain proteins, a Southern blot
if they are considering the bands that contain DNA, and a Northern blot
if they are considering the bands that contain RNA. (The unusual names
are a salute to E. M. Southern, the scientist who devised this process.)
Protein bands and their constituent molecules are named according to
the weight (in daltons) of the molecules. The prefix "p" stands for protein
, and "gp" stands for glyco-protein (glyco meaning that the
protein has some sugar molecules stuck to it). RNA and DNA bands are named
according to the number of nucleic acids or base pairs (in kilobases) that
make up the constituent RNA or DNA molecules.
Proving a virus causes a disease
If a virus is hardy and abundant enough to cause a disease, scientists
should have no trouble isolating it from the cell-free fluids of affected
tissue. This is exactly what scientists must do to convict a virus of causing
a disease. They select a group of people who have the disease, and try
isolating the virus from the patients' plasma (cell-free blood),
or other fluids, depending on the disease. If they fail to isolate the
virus from some of the patients, then they must absolve that virus of responsibility
for any progressive disease in those individuals, and conclude those people
are sick for some other reason.
But what about patients who do present isolatable amounts of
the virus? Is that virus responsible for their conditions? Or is the virus
an innocent bystander? After all, most viruses cause no disease.
Only microbiological experimentation can establish the culpability or
innocence of a virus isolated from the fluid of diseased tissue. To do
this, scientists prepare cultures of healthy, uninfected cells of the type
damaged or destroyed in the patient. They add viral isolates and watch
to see if this effects the culture cells in ways that can explain the disease.
Understanding viral tests
Although isolation is the only direct evidence of a virus, cost and
time considerations make it impractical for clinicians. Among other things,
for example, it requires confirmation by an electron microscope.
Viral tests, on the other hand, are much simpler. Most require
clinicians to just add patient fluids (usually plasma, depending on the
virus in question) to the tests and look for reactions to take place.
Scientists construct these tests using components abstracted from
viral isolates. Some of the proteins from viral isolates, for example,
will react with antibodies secreted into plasma by the immune systems of
patients infected by the virus. Antibody tests consist of those proteins.
Genetic tests consist of probes made from the DNA or RNA contained in viral
isolates. The probes react with viral RNA or DNA in patient fluids.
When constructed and validated properly, and used under the proper
circumstances, viral tests can be nearly as accurate and reliable as viral
isolation itself. The need for proper test validation and result interpretation
stems from the fact that the reactions upon which they depend (antibody-antigen
interactions, and genetic probing) are not perfectly specific. Antibodies
against one viral protein can react with a similar protein from other microbes,
or even some non-microbial proteins. Similar proteins mean similar gene
sequences, so genetic tests are less than perfectly specific as well.
Furthermore, even when tests react with their intended viral entities,
this doesn't necessarily mean the patient has an active infection, the
only sort that can cause disease. Antibodies, for example, can circulate
for years—even a lifetime—after the host immune system has suppressed a
viral infection to permanent and harmless latency, or even eliminated it
entirely. Viral genetic material can also persist in the plasma and other
fluids during viral latency.
Therefore, viral tests cannot absolutely and unambiguously identify
an actively infected person. Only isolation of objects that possess the
appearance and density of the virus in question can do that. So viral tests
must not only be constructed from viral isolates, they must also be validated
against their ability to predict patients from whom scientists can obtain
viral isolates.
Validation studies tend to show that positive test results are
highly accurate for patients who express the symptoms that the virus has
been proven to cause. On the other hand, positive results are usually very
inaccurate for people who have no symptoms. In other words, the virus can
be isolated from some very large fraction of positive testing people who
express the associated symptoms, but only from a small fraction of positive
testing people who express no symptoms. Thus positive tests in healthy
people usually don't indicate active infections.
Antigens and antibodies
One measure of the immune system's response to a substantial viral infection
is the production by B-cells of proteins called antibodies. Antibodies
latch onto and neutralize other proteins.
Proteins that elicit an immune response are called antigens. Viral
antigens tend to be those proteins that compose the inner core, and those
that decorate or line the outer membrane envelope. These are the proteins
that the immune system "sees," whereas proteins inside the core—the viral
enzymes—are shielded from immune surveillance. The immune system does not
respond to non-protein molecules, like RNA and DNA.
Western blot antibody tests and ELISAs
Scientists construct Western blot antibody tests by transferring to
paper some of the protein bands from a Western blot gel. These bands will
react when exposed to fluid that contains antibodies against the proteins
in the bands.
Another test called the ELISA consists of viral isolates for which
the constituent molecules have been broken apart from each other, but have
not been separated from each other by electrophoresis. ("ELISA" stands
for Enzyme-Linked Immuno-Sorbent Assay, which describes how positive reactions
are demonstrated chemically.) This makes ELISAs easier and cheaper to make
than Western blot tests.
But ELISAs are not as accurate. People test ELISA-positive if
their plasma contains antibodies against just one of the viral proteins,
whereas Western blot tests consist of the proteins separated into different
bands, so clinicians can see exactly which proteins react with a person's
plasma.
ELISAs are usually used as screening tests. Since people who test
ELISA-negative have antibodies against none of the viral proteins, negative
ELISAs just as accurately identify uninfected people as do Western blot
tests showing no reactive bands.
But positive ELISAs are not as good as positive Western blots
at identifying people with active infections. This is because there is
no such thing as a specific antibody. Antibodies against a certain viral
protein may react also with proteins of another virus, or even non-viral
proteins. So positivity for antibodies against viral proteins is not unambiguous
evidence that a person has been exposed before to a particular virus.
However, people positive for antibodies against all the antigens
of a particular virus are much more likely to have been exposed to that
virus than someone positive for antibodies against only one or a few of
the antigens. Yet each receives the same positive ELISA. Only a Western
blot can distinguish these people. Proper validation studies show higher
accuracies (fraction of positive subjects from which viral isolates can
be obtained) for positive Western blot tests than for positive ELISAs.
But since ELISAs are cheaper, Western blot tests are usually reserved for
people who first test ELISA-positive.
Northern and Southern blot tests
Southern and Northern blots from viral isolates represent pure samples
of viral DNA or RNA. Scientists use the material in these samples to produce
viral tests that react with viral RNA or DNA in patient fluids. To do this,
they construct small DNA or RNA molecules, called probes, that complement
segments of the viral DNA or RNA. To test patients, clinicians make Northern
or Southern blots from patient fluid (usually plasma, depending on the
virus in question) that has been treated so that any constituent viruses
will be broken apart, exposing the genetic material inside.
If there is lots of virus in the plasma, a distinct band will
appear in the gel at the location characteristic of the genetic material
of the virus in question. Adding the probes will confirm that such a band
consists of viral DNA or RNA. If only a small amount of virus exists in
the plasma, the genetic material settling at the characteristic location
in the gel will become detectable only after the probes are added.
Antigen tests
During the early course of a substantial viral infection, the plasma
contains lots of virus, and consequently lots of viral antigens, but very
few antibodies against these antigens. This is because the immune response
has not yet caught up with the viral activity.
Sometimes there is not even enough antibody to cause a reaction
with ELISA or Western blot antibody tests, which contain the antigens.
So scientists have developed tests that contain antibodies against viral
antigens. These tests react with patient plasma that contains viral proteins.
Antigen tests, then, are the inverse of antibody tests.
Viral load tests
Patients rarely ever have enough "HIV RNA" to yield a detectable signal
on Northern blots of fresh patient serum. Thus the necessity to invent
"viral load" testing, which employs the polymerase chain reaction, PCR.
PCR generates millions of RNA or DNA copies out of an original indetectably
few molecules.
AIDS reappraisers consider these tests invalid. The concentrations
of HIV RNA these assays usually indicate—hundreds of thousands per ml of
plasma—would easily show up on Northern blot tests. But it doesn't show
up at all.
Active vs. inactive viral infections
Cells with inactive, or dormant, infections have inside them viral DNA
molecules, called proviruses, that are asleep. Sleeping proviruses produce
no virus, and thus can cause no disease, since viral replication is what
destroys or damages cells in the course of a viral disease.
Viral DNA goes to sleep when the host immune system gains the
upper hand. Among the anti-viral molecules secreted by immune cells are
substances that put viral DNA to sleep. When immunity is suppressed, the
plasma levels of these substances diminish, and sleeping viral DNA awakens
to start producing new viruses, which show up in the plasma.
Since cell cultures contain no immune systems, they contain none
of these anti-viral substances. That makes them ideal nurseries for viruses.
When cultures are made from cells containing dormant proviruses, the proviruses
have their ideal circumstance to spring back to life and generate a maximum
amount of new virus. Some proviruses awaken from dormancy only when stimulated
by agents that promote viral activity. These sorts of viruses make very
poor candidates for disease causation, for obvious reasons.
The culture skinny
Viruses replicate in two sorts of cells, in vivo (those inside living
organisms, such as people), and in vitro (those maintained in laboratory
culture dishes). Virus isolation from human plasma demonstrates in vivo
viral activity, and isolation from culture fluids demonstrates in vitro
viral activity.
However, isolation from the fluids of a culture composed of donor
cells can not demonstrate that the donor harbors an active infection. It
only demonstrates that the donor cells contain proviruses that are active
under culture conditions. Transferring cells from a living organism to
a lab dish can permit sleeping proviruses to awaken. Only examination of
uncultured tissue fluids can diagnose viral disease.
RNA and DNA viruses
Viruses carry in their core only one sort of genetic material, either
RNA or DNA molecules. These molecules are called proviruses when they reside
inside a host cell, outside the viral core. Proviruses direct the production
of all viral components, even replication of themselves, in the manufacture
of new viruses.
Except for retroviruses, viruses that carry RNA are always active,
but viruses that carry DNA can be active or inactive.
This is because RNA constantly produces proteins when it is in
contact with amino acids (protein building blocks) and ribosomes (enzymes
that translate RNA molecules into corresponding protein molecules). In
the viral core, viral RNA has no contact with amino acids or ribosomes.
But inside a host cell, viral RNA has all the material it needs to produce
new viral proteins.
DNA, on the other hand, can not be directly translated into proteins.
First it must be transcribed into RNA by an enzyme called transcriptase.
But DNA has the ability to regulate its own transcription. So DNA can be
active or inactive, whereas RNA can only be active.
Enzymes called reverse transcriptase will reverse transcribe retroviral
RNA into corresponding DNA molecules. Consequently, retroviruses share
with DNA viruses the ability to be either active or inactive.
Virus counting
How many viruses do infected people have circulating in their blood?
There is only one way to answer this question definitively, and it of course
involves preparing an isolate from patient plasma, and counting the viruses
in the isolate.
Scientists start by obtaining from the patient a fluid sample,
which they serial dilute,. Serial dilution results in one undiluted sample,
and several others of equal volume diluted by varying amounts. From each
sample scientists prepare a viral isolate, which they view on a standard
grid using an electron microscope. If the patient has a high viral concentration,
the undiluted sample will contain too many to count, since the viruses
will be stacked on top of each other, and overlap.
The viruses in one of the diluted isolates will be spread out
enough so they can be counted accurately against the grid, which represents
some fraction of the area occupied by the sample. By multiplying factors
that account for the gridding and diluting of the sample, counting the
number of viruses in the grid will yield the number of viruses present
in the undiluted sample. Dividing this number by the pre-diluted volume
yields the viral concentration (in particles per milliliter) in the patient's
plasma.
This of course is too expensive and complicated for the clinical
setting. So scientists can calibrate some of the viral tests to approximate
viral concentrations. For example, the thickness and staining intensity
of Northern and Southern blot bands are directly proportional to viral
concentration. So is the staining intensity of antigen tests. So by examining
these test results for patients who have had their viral concentrations
established, scientists can derive numbers that convert band thickness
or staining intensity into viral concentration.
Tissue Culture Infectious Doses (TCID)
One of the ways that clinicians can characterize a viral infection is
to approximate the plasma concentrations of Tissue Culture Infectious Doses,
or TCIDs. One TCID is the minimum amount of virus required to produce viral
activity in a standard culture of stock laboratory cells. Scientist determine
viral activity by obtaining viral isolates, or by observing phenomena previously
shown in isolation studies to be viral, such as the appearance of certain
proteins.
To determine TCID concentrations, scientists take plasma from
a patient and serial dilute it. Serial dilution involves producing from
an original sample a sequence of samples, each of the same volume, but
each one ten times more diluted than the previous. Thus going down the
line from the original sample to the last, each will contain in relation
to the previous one, a tenth of the plasma—and a tenth of the viruses—contained
in the original sample.
Each sample is added to a separate standard culture. If the undiluted
sample causes no replication, then neither will any of the diluted samples,
and the plasma contains no TCIDs. If the undiluted sample does cause replication,
then it contains at least one TCID. If the first diluted sample causes
replication, then the undiluted sample contained at least ten TCIDs. If
the second diluted sample produces no replication, then the original sample
contained at least 10, but fewer than 100, TCIDs. This range can be narrowed
down by now using a dilution factor smaller than ten.
Because the usual dilution factor is ten, this process is called
titration, and the term TCID can be substituted with the term titer (or
titre). In the above example, scientists would say that the original sample
had 10 TCIDs, or a viral titer of ten. By dividing this figure by the volume
of the original sample, they can calculate the plasma concentration for
the patient, in TCIDs per milliliter.
Validation studies can correlate TCID concentrations with actual
viral concentration. However, this usually is not done, because TCID values
provide more important information than viral concentrations. Remember,
only infectious viruses can cause disease, and only if they are present
at concentrations great enough to cause a productive infection. So it is
more important to know the concentration of TCIDs than the concentration
of actual viruses. *