Sean D. Pitman M.D.
© December, 2004
Are genes within creatures such as bacteria capable of evolving unique and novel functions via mutations and natural selection? The short answer is - Yes. However, such evolution, although capable of yielding many startling new abilities, is limited to the lowest levels of functional complexity (like antibiotic resistance).
The de novo evolution of antibiotic resistance is based on the specificity of antibiotic interactions with various protein sequences within a bacterium. Because of the specificity of such interactions, a very high ratio of mutations are able to interfere with or completely disrupt these specific interactions - and antibiotic resistance is the result. In other words, there is a relatively high likelihood that a particular mutation in a target sequence will result in antibiotic resistance. This high likelihood translates into a very rapid evolutionary process. In real life, this is exactly what happens. The evolution of antibiotic resistance of any previously susceptible bacterial colony to just about any antibiotic is usually realized within a very short period of sustained antibiotic exposure.
This paper begins with the historical discovery of antibiotics and the rapid arrival of resistant strains of bacteria to these antibiotics. Also, the specific mechanisms of antibiotic action as well as resistance mechanisms are discussed - to include limits beyond which the powers of evolution seem to fail.
The euphoria was short lived however. Shortly after the general usage of antibiotics began the microscopic world revealed its genius for becoming resistant to antibiotics. It truly seemed like a fulfillment of Darwin's prophetic vision and the firm establishment of his Theory of Evolution as unshakable. Today, Darwin's Theory of Evolution seems more validated than ever by these little creatures. The early advances made by antibiotic and antiviral medications seem to be almost completely overcome by the continued evolution of antibiotic and antiviral resistance. So called "Superbugs" are springing up everywhere that are resistant to every antibiotic or antiviral currently known to man. Evolution seems not only to be a wonderful, but also a terrible reality. But is it all really as it seems?
Bacteria become resistant to antibiotics by a very simple method of natural selection. When a large number of bacteria are presented for the first time with an antibiotic, most, if not all of them, die off. If all of them die, then obviously no resistance is gained for that particular bacterial colony or group. The problem is that sometimes one or two or even a few bacteria survive the initial exposure. This is because they were previously resistant before exposure to the antibiotic. Of course, after they survive the initial exposure they reproduce themselves and make a new colony of bacteria. Now, every bacterium in that colony is a clone of the original resistant bacterium and so all of them are resistant to that particular antibiotic to the same degree. But how did the first one or two resistant bacteria survive to pass on their evolved resistance to their offspring?
There are three main targets that
Bacterial protein synthesis
Bacterial nucleic acid replication and repair
Cell wall biosynthesis enzymes and substrates
So, intuitively there are also three basic mechanisms of bacterial antibiotic resistance:
Alteration of the antibiotic target
Restriction of antibiotic access to the target
Direct inactivation of the antibiotic
Obviously then, if a bacterium can achieve any of these three blocks to the activity of an antibiotic, it is resistant to that antibiotic. If this ability is due to a genetic alteration, then this alteration will be passed on to each and everyone of its offspring since bacteria reproduce in a clonal fashion. In short, this is evolution in action.
(Back to Top)
This paper begins with a general discussion of antibiotic resistance and how it is gained. Following sections go into more specific detail regarding various types of antibiotic resistance.
Obviously, bacterial antibiotic resistance is achieved in many ways to include spontaneous genetic mutations that result in new traits or functions that were not present in previous generations of that bacterial group or species. New beneficial functions did in fact "evolve" spontaneously through mutation and natural selection in such cases. Clearly then, the process of evolution is actually demonstrated by these little creatures - or is it? Some argue that these changes are not "evolutionary" since they do not really increase the information content of the DNA within that bacterium. Consider the following comments by Dr. Carl Wieland (An Australian Medical Doctor):
This misconception [about
antibiotic resistance and evolution] may be partly due to the fact that even
many science graduates believe that the mechanism of antibiotic resistance
involves the acquisition of new DNA information by accidental mutations... But
resistance does not normally arise like this.
Loss of control over an enzyme's production can engender antibiotic resistance. Take for instance penicillin resistance in Staphylococcus bacteria. This requires the bacterium to have DNA information coding for production of a complicated enzyme (penicillinase), which specifically destroys penicillin. It is extremely unlikely that such complex information could arise in a single mutation step, and in fact it does not. Mutation can cause the loss of control of its production, so much greater amounts are produced, and a bacterium producing large quantities of penicillinase will survive when placed in a solution containing penicillin, whereas those producing lesser amounts will not. The information for producing this complicated chemical was, however, already present 1
Is Dr. Wieland correct in his statement here? Actually, as described in more detail below, Wieland is at least partly incorrect. Point mutations to the penicillin binding proteins (PBPs) can and do result in penicillin resistance even in the absence of the penicillinase gene. It may be true that penicillinase production is the most significant mechanism of penicillin resistance in gram-positive bacteria, but clearly this is not the only mechanism. The â€œevolutionâ€ of new resistance mechanisms via the mutation of existing genes does in fact occur. However, Wieland is correct in saying that some mutations can result in an increased production of a pre-existing penicillinase enzyme from a pre-existing penicillinase gene. He is also correct in stating that the information for producing the penicillinase enzyme does not spontaneously evolve in those bacteria that are able to produce it, but was already present either through vertical transmission from the previous generation or via horizontal transmission through the action of plasmids or other methods of DNA transfer between bacteria.
Also, although no one has ever observed the de novo evolution of a penicillinase enzyme, evolutionary scientists present evidence for the original evolution of penicillinase from existing bacterial genes - but this still remains hypothetical until such proposed evolutionary pathways can be demonstrated in real time. So far, not even a single hypothesized step in the pathway of penicillinase evolution has ever been demonstrated to actually evolve in any bacterium (see appendix).
Certainly there are many similarities, especially when one considers the three-dimensional structures and active sites between homologous sequences that are supposed to be the most likely precursor proteins to penicillinase (appendix), but there might be a few functionally significant neutral gaps between them. How are these neutral gaps overcome? Ultimately, if the crossing of such evolutionary paths between these two enzymes is truly an easy process, as many claim it is, then why is beta-lactamase evolution not being demonstrated in the lab? If it is so easy, then why is it so difficult to demonstrate?
Similar evolutionary limits have been found with other single protein enzymes - like lactase evolution experiments showing "limited evolutionary potential" in mutant forms of K12 E. coli bacteria. It seems like the average distance of selectable lactase sequences in sequence space is quite significant given the lack of lactase evolution over the course of many tens of thousands of generations in a lactose rich environment. The same limitations seem to be present when it comes other functions that exist at a similar level of functional complexity (i.e., having a similar minimum sequence size and specificity of amino acid residue arrangement for a minimal degree of selectable function).
Of course, some beta-lactamases, such as the class C beta-lactams, maintain some DD-peptidase activity. This might be a very good clue to the existence of a real evolutionary pathway, but not necessarily. Neutral gaps can still exist even if the other functions from proposed enzymes of origin exist in the "target" enzyme. The problem for the class C beta-lactams is that their DD-peptidase function is not great enough to produce a selective advantage as far as DD-peptidase function is concerned. Several specific mutations are required before a level of selectively advantageous DD-peptidase function can be realized. One or more of these specific mutations are most certainly neutral as far as their functional selective advantage is concerned. Also, the reverse is not necessarily true. Functional DD-peptidases do not have beta-lactamase activity. In the proposed evolutionary pathway from a DD-peptidase to a beta-lactamase, each and every step needs to increase beta-lactamase activity in a selectable way or there will be a neutral or even detrimental block in the pathway.
A neutral change is a change in the genetic sequence (genotype) that cannot be distinguished by natural selection from different genetic sequences that have that same function (phenotype). The problem is that with every additional neutral mutation that is required along a path toward new function, the average time required to traverse this path increases exponentially. This neutral gap problem seems to be the most likely source of "limited evolutionary potential" when it comes to evolving novel functions - like single protein enzymes (i.e., penicillinase).
So yes, it is statistically possible but
improbable that penicillinase evolution is responsible for anything as far as "de
novo" penicillin resistance within a newly resistant population.
Penicillinase, when detected in a bacterial population, was most likely already
there before the selection pressures of penicillin antibiotics were applied to
that population. In other words, penicillinase most
likely existed in the genomes of bacteria long before Alexander Fleming came on
The same can probably be said of many plant,
insect, rodent, and other "weed and pest" resistance to the chemicals used to
kill them off.
Consider the following quote from the geneticist Francisco Ayala:
The genetic variants required for resistance to the most diverse kinds of pesticides were apparently present in every one of the populations exposed to these man-made compounds.3
As additional support for this statement,
consider that bacteria recovered from historical isolation have been found to be
resistant to modern antibiotics. In 1988 bacteria were recovered from the colons
(intestines) of Arctic explorers who froze where they died in 1845. Many decades
later, these explorers where found and various studies where done on their
bodies. Bacteria from their intestines were actually grown and subjected to
various modern antibiotic medications. Many of the bacterial colonies grown were
found to be resistant to many modern antibiotics, proving that this resistance
did not evolve over just the past 60 years or so since the antibiotic age began,
but where already present before humans started using antibiotics to fight
bacterial infections. 4
These methods of resistance
are not limited to bacteria, but extend to the plant and animal kingdoms as
well. Rats, for example, have been shone to be able to develop a resistance to
the poison warfarin. In the 1950's, England was using warfarin to kill off its
rat population. In a few years, the rats had become resistant to warfarin. But
how did they do it? Warfarin kills rats by inhibiting an enzyme involved
in the metabolism of vitamin K. Since Vitamin K is vital for life, rats who lost
the ability to synthesize vitamin K in their cells died. Warfarin-resistant rats
were found to have a mutated form of the enzyme that warfarin inhibited. The
different shape and structure of the enzyme prevented warfarin from binding and
blocking Vitamin K synthesis. These mutant rats were still able to
synthesize enough vitamin K to live.2
Does this sound familiar? Many times bacteria achieve antibiotic resistance via the same kinds of mutations - though sometimes the mutated enzymes are not as efficient as the non-mutated or "wild-type" forms. For example, the mutated enzyme in rats that blocks the activity of warfarin is 10 times less efficient at synthesizing a given amount of vitamin K. Obviously then, if the warfarin was removed from the environment, those rats with this less efficient enzyme would not be as "fit" as those with the original wild-type form of this enzyme. Without warfarin in the environment, these rats would be quickly replaced by rats with the original enzyme unless the mutant rats already evolved "compensatory mutations" - which can, over time, overcome the lessened effects of the mutant enzyme so that it can also achieve "wild-type" levels of activity in many cases.48
The same thing happens with bacterial colonies when a particular antibiotic is removed from their environment. Sometimes they revert back to the original wild-type sequence and again become "sensitive" to the original antibiotic. In fact, due to this phenomenon, many hospitals are considering the use of cyclic antibiotic formularies where antibiotics would be rotated after a period of time. One type of antibiotic would be used for a while, and then replaced by a different type of antibiotic just as the bacterial population was gaining resistance. After a while, the second antibiotic would be replaced by a third type of antibiotic - and so on. Then after a bit more time, the original antibiotic could be used again will full effect and the cycle would start over.
There is just one little problem with this idea. The problem is that many of the disadvantages sustained by the initial evolution against many types of antibiotics can be compensated for by "compensatory mutations".48 In other words, bacteria that take on certain disadvantages in order to gain a particularly vital mutational advantage can often compensate for the disadvantages at a later time with additional mutations that re-enhance the function that was suppressed by the initial detrimental aspects of the original mutation(s).
Still, such novel functional changes are only found at the lowest levels of functional complexity where no more than a few hundred amino acid residue "parts" or "characters" are required to work together in a fairly specified arrangement. Any novel function requiring more than a few hundred fairly specified amino acids working together at the same time has never been observed to evolve in real time. Obviously there must be some blockade that slows evolutionary progress down in an exponential manner with increasing functional complexity.
Herbicides also work the same way. Consider the following quote from geneticist Maceirj Giertych (Ph.D., D.Sc.):
Much evolutionary publicity is attached to forms that develop resistance to man-made chemicals. Usually they are variants that exist in nature but were selected out by the chemical reagent. In one instance, it was demonstrated that a single nucleotide substitution in the genome was responsible for resistance to a weed-specific herbicide. The herbicide is 'custom-made' for attachment and deactivation of a vital protein specific for the weed plant. A single change in the genetic code for this protein, in the sector used for defining the herbicide attachment, deprives the herbicide of attachability and therefore of its herbicidal properties. Such a change has no selective value except in the context of the man-made herbicide. 5
Viruses also evolve using similar techniques. Various mutations in the viral genome result in blocks to antiviral binding sites - just like the blocks to antibiotic binding sights. These blocks may result in a change to the viral protein coat. Such changes may prevent antibody recognition. It all depends upon the specific action and target of the antiviral, but the principle is the same. A minimal change to the target gives significant resistance to the host. This phenomenon is made possible because of the high level of specificity between the antibody / antibiotic / antiviral etc. and their target sequences.51
But, what about the argument that evolution only
takes away information or that all mutations are harmful in one way or another?
Some argue that resistant bacteria are less "fit" than their "wild-type"
counterparts and that their pathogenicity as well as their metabolic and
reproduction rates are lower than before?
Consider the following quote from Wieland:
So-called 'supergerms' in
hospitals are not 'super' at all. What has happened is that the use of
antibiotics in modern hospitals has meant that the only ones surviving
those, which have all the resistance factors. If a person gets a serious
infection with one of these resistant types, the infection is not therefore more
aggressive than if it was a non-resistance form of the same bug; it is simply
that doctors are powerless to treat it. In fact, it is generally a weaker form
of the pathogen. 1
Is Wieland correct in this statement? Well yes, he is generally correct - sort of. Most resistant strains of infectious bacteria are not "more virulent" outside of their ability to resist certain antibiotics. Often they are even less metabolically active and less able to cause disease than their wild-type counterparts, at least for a while, but not always - and not even if they were hindered initially since they may be able to compensate for initial losses of function with compensatory mutations.48 Besides, this whole argument says nothing about an organism's ability to evolve new functions. Just because new functions evolved in a weaker organism or even if the new functions caused the organism to be "weak" does not negate the fact that a new function evolved that was not there before.
On top of this, many new functions, that are known to be the result of spontaneous genetic mutations, are not harmful in the least, but are entirely helpful. For example, consider again Barry Hall's work with lactase evolution in E. coli. The evolution of the lactase function in these colonies of E. coli came at no detrimental cost as compared to the mutant wild-type strains of K12 E. coli used by Hall.9
There are many other such examples in literature (such as the well known example of nylonase evolution). However, what is especially interesting to me is that the evolution of new functions generally involves just one or maybe two point mutations in functional sequences that are relatively short. Functions of greater complexity, requiring anything more than a few hundred fairly specified amino acid residues working together at the same time, simply do not evolve. Specifically, multi-protein systems, like bacterial systems of motility (i.e., the flagellum, with more than 20 different required parts and well over 5,000 fairly specified residues at minimum), where all the uniquely interdependent protein parts work together at the same time in a specified orientation with each other, just do not evolve in real life. Not even a single proposed step in flagellar evolution scenarios has ever been shown to evolve in real life - not one! Other examples of such cellular functions include DNA transcription and mRNA translation, phagocytosis, endocytosis and pinocytosis (cell "eating" and "drinking"), and many more.
But what about diseases such as sickle-cell
anemia and various forms of cancer? Are these really examples of the
evolution of new functions?
I would say that they are, but does this evolution explain the theory of
common descent when one moves up the ladder of functional complexity?
Consider the following comments from Dr. Felix Konotey-Ahulu, an
authority on sickle-cell anemia:
Sixth-graders I have lectured on genetic counseling invariably pop some questions such as: 'Is it true that the sickle-cell phenomenon has established Darwinian evolution as fact?' Behind the question, of course, lies the assumption that observing selection/adaptation involving a mutation (an inherited random change or defect) somehow implies that the more complicated forms seen today arose from simpler forms traced ultimately to one-cell organisms.7
So why, if new functions can and do evolve,
can't Darwinian evolution explain the origin of the huge variety of life forms
that we see in the natural world?
It seems rather intuitive that if small functionally unique changes can
occur that these changes should be able to add up over time to produce larger
and still larger changes - until the diversity and complexity that we see around
us comes into being. Where then is the limitation to
Over and over again we see that rapid evolution of new function is always the result of one or rarely two point mutations to previously existing genetic elements. It is obvious then that the evolution of new functions or traits does happen via such mutations combined with the power of natural selection. This cannot be denied nor should we find it surprising. It is statistically likely, given that a gene is but one point mutation away from new function, that such a gene will chance upon such a beneficial target sequence that is no more than an arm's distance away in a reasonable amount of time. The problem for evolution comes when more than a single point mutation is needed to achieve a given function - such as appears to be the case with the evolution of the penicillinase and other such relatively simple single-protein enzymatic-type functions.
Even the evolution of the beta-galactosidase
function in the E. coli mutants studied by Hall required the prior
existence of the evolved beta-galactosidase gene (ebg).
Colonies of E. coli that had the original lacZ as well as ebgA
genes removed, never evolved b-galactosidase function despite very large
colonies with high mutation rates living in a lactose enriched environment for
tens of thousands of generations. It appears that
without these two genes, no other genes or genetic elements in E. coli
are close enough to the minimum part requirement of beta-galactosidases to
achieve beta-galactosidase activity with just one or two point mutations.
The ratio of what will work compared with what will not work is relatively
small. This small ratio between non-junk and junk creates a neutral gap
between various non-junk functions at a certain level of functional complexity.
For all cellular functions there is a minimum part requirement consisting of amino acids in specific sequences. If changed beyond this minimum requirement, all function is lost. All genes and all proteins are in fact, "irreducibly complex." Despite the fact that many genes and proteins are quite flexible in their sequencing, all of them have a limit beyond which all beneficial function is lost. These limits may overlap with other genes and proteins, in which case, evolution or change between two functional genes or proteins is possible in a relatively rapid manner. However, as the level of functional complexity increases, the average neutral gaps between potentially beneficial proteins also increase. With the increase in neutral gaps comes a decrease in functional overlap between various sequences. At this point, multiple neutral mutations are required before a new beneficial function can be realized. These multiple mutations are invisible to the powers of natural selection. That is why such changes are called "neutral". They are neutral with respect to functional change and this makes them neutral with respect to any selective advantage that nature might provide.
So, the traversing of such a gap requires a truly random walk. And, as we all know, it is much faster to go from point A to point B by following a straight line. Walking along a random curvy path will take a whole lot longer. This is what happens with evolution when the pathway is neutral with regard to any sequentially selective advantages. The evolution of new functions at such levels requires exponentially greater amounts of time.
Clearly then, these neutral gaps
present insurmountable blockades to the evolution of new functions beyond the
lowest levels of functional complexity - even for such large populations and
such rapid generation turnovers as are realized in bacterial colonies (and we
are talking trillions upon trillions of years for the crossing of neutral gaps
averaging no more than a couple dozen
residue changes wide). Since such
isolated functions of higher and higher levels of complexity do in fact exist in
the natural world, it seems extremely difficult for the theory of evolution or
any other purely naturalistic theory based on mindless naturalistic processes to
explain their existence outside of deliberate design.
(Back to Top)
Specific Mechanisms of Antibiotic
Alteration of the Antibiotic Target
This is probably the most common mechanism of
When an antibiotic binds to its target, it limits the target's ability to
perform its normal function. So, if a mutation
occurred that blocked that antibiotic's ability to bind to the target, the
antibiotic would loose its ability to hinder the function of that target.
The actions of many types of antibiotics are successfully prevented by such
Bacterial Protein Synthesis
Aminoglycosides and macrolides are antibiotics
that target bacterial protein synthesis.
The great differences between prokaryotic and eukaryotic (bacteria and
humans) systems of protein synthesis allows for their great specificity in the
targeting of bacteria without harming the human host.
In order to understand this process more clearly we must first review the
process of protein formation from the coded information found in a bacterium's
DNA (deoxyribonucleic acid).
Of course the information in DNA is stored in coded form by the linear sequence of the chemical "letters" labeled with the names adenine (A), thymine (T), guanine (G), and cytosine (C). These chemical letters are defined according to three-letter sequences or "words" called "codons." Each one of these codons are defined by the "genetic code" of the bacterium as representing one of twenty amino acids (amino acids are the basic building blocks of proteins).10
This system works very much like the Morse Code.
Each symbol in the Morse Code represents one of the twenty-six letters in the
In just the same way, codons represent the "letters" or amino acids
in the protein "alphabet." However, DNA does not get
"translated" into proteins directly.
DNA is used as a template or a "master copy".
There is another molecule called, "messenger RNA" (ribonucleic acid),
that is used as a "working copy" of the DNA.
The cell in fact "transcribes" the desired portion of DNA into mRNA and
then the mRNA is "translated" into appropriate amino acid sequence or protein
molecule. RNA molecules are similar to DNA molecules
but do have some important differences. Thymine (T)
in DNA is replaced by uracil (U) in RNA.10
RNA is also a single stranded molecule while DNA is double stranded. In any case, it is the mRNA strand that is translated into an
amino acid / protein sequence.
The process of protein translation is quite complicated, but basically it involves the recognition of another molecule called, "transfer RNA" (tRNA). tRNA has a two-dimensional shape that resembles a cross (note illustration). One of the three loops that make up the arms of the cross is called the "anticodon loop." This loop contains a three-letter sequence that matches or compliments a specific codon on the mRNA molecule. For example, in the figure of mRNA (pictured above) "Codon 1" reads, "GCU." The matching anticodon of tRNA would read, "CGA." This matching of codon and anticodon is specifically linked to a particular amino acid. Also, each tRNA has a "3'-end" that is called, "the site of amino acid attachment." The association of a particular amino acid with a particular tRNA and its anticodon is extremely specific. Therefore, a particular anticodon is always associated with the same amino acid. 12
However, in order to bring the codon and anticodon together in such a way where amino acid linking can be made in sequential order, different molecules of RNA called "ribosomes" (rRNA) are needed. Ribosomes consist of two main parts or "subunits." These subunits come together around the mRNA like two sides of a hotdog bun surrounding the mRNA hotdog. In certain bacteria, such as E. coli, the ribosomal subunits come together in a specific location on the mRNA chain called the "Shine-Delgarno sequence" that reads, "AGGAGG." This particular sequence is located just in front of another sequence known as the "start" sequence or "AUG" sequence. Without the Shine-Delgarno sequence, the ribosomes would not know where to begin the process of protein sequencing. Obviously if the sequencing started in the middle of the mRNA molecule, one half of a desired protein would be made. So, it is vital that the process be able to recognize were to start sequencing. After the "beginning" of the mRNA molecule is recognized, the beginning message is recognized by a special "initiator tRNA." The initiator tRNA always has a methionine amino acid attached to it that will form the "N-terminal" amino acid in the growing peptide chain. This methionine molecule may be removed however before the functional protein has been completed. The process so far described is called, "initiation." The next required step is called, "elongation."
Elongation of the peptide chain involves the addition of amino acids to the carboxyl end of the growing polypeptide chain. During elongation the ribosomal complex moves from the 5'-end to the 3'-end of the mRNA molecule that is being translated. Elongation is made possible by the fact that the ribosomal complex has two binding sites for tRNA molecules called the "A" and the "P" sites that involve both ribosomal subunits. The A and P sites cover two neighboring codons on the mRNA chain. During translation the A site binds an incoming tRNA that matches the mRNA codon that is currently occupying the A site. This codon specifies the next amino acid to be added to the growing polypeptide chain. The P site is occupied by a tRNA molecule that matches the codon that was at the A site just one step before. In other words, as the ribosome complex moves along the mRNA chain, the codons come in one side of the ribosome complex and leave out the other side. The incoming codons pass sequentially through the A site and then the P site before leaving the ribosomal complex. As the complex moves from one codon to the next, the growing amino acid chain on the tRNA occupying the P site is attached to the N-terminal end of the single amino acid occupying the A site. After this is done, the complex moves forward to the next codon on the mRNA. As the complex moves forward, the tRNA molecule that was occupying the P site is bumped out if its place by the tRNA that was occupying the A site. Now the former A site tRNA occupant is the new P site occupant, still with the growing polypeptide chain attached to it. Then, the next tRNA molecule that matches the next codon arrives and the whole process repeats itself over and over again until the entire polypeptide is formed.12
Many more molecules and interactions are needed and four of the steps require energy in the form of energy molecules called adenosine triphosphate (ATP) and guanosine triphosphate (GTP). In fact, the addition of one amino acid to a growing peptide chain requires the use of 2 ATPs and 2 GTPs.12 Even from this short and highly simplified overview of polypeptide synthesis, it is clear that it is a very complex, detail dependent, and highly specific process. It is similar to very finely tuned assembly line. Obviously then, if any one step is blocked the entire process would fail and the host cell would quickly die.
Many antibiotics work by blocking one step or
another in the process of protein production.
Aminoglycosides, such as kenamycin and streptomycin, interfere with
protein production by binding to the smaller of the two ribosomal subunits (30S
subunit). This is significant because within
the ribosomal structure translational accuracy requires that the anti-codon of
tRNA and the codon of mRNA be specifically recognized.
This recognition occurs with the 30S ribosomal subunit.
Even slight changes in this region of the 30S ribosomal subunit interfere with
the fidelity of this process of recognition. The
requirement for such high fidelity in this step creates a weak point in the
Aminoglycosides have been extremely successful
in exploiting this weakness. For example, the
aminoglycoside "streptomycin" specifically attaches to a region around
the " site" of the 30S ribosomal subunit (specifically the 16S sub-subunit)
along with some other proteins within the 30S subunit and stabilizes a
conformational state in which the initial binding of "non-cognate" or
non-similar tRNAs are favored. This results in coding errors and, of course, the
production of nonfunctional proteins. Even today
aminoglycosides remain useful for the treatment of infections caused by aerobic
Gram-negative bacteria. Also, both streptomycin and
kanamycin are important second-line chemotherapeutics in the treatment of
multi-drug resistant tuberculosis.13
It is evident then that any change that affects the binding of aminoglycoside antibiotics to this 30S ribosomal subunit would block their adverse effects on protein synthesis, thus giving the bacterium "resistance" to this type of antibiotic. It is not too surprising then that streptomycin resistance involves a single point mutation in the gene (rrs locus) that codes for the 16S sub-subunit of the ribosome complex. Another point mutation in a gene (rpsL gene) that codes for the protein portion (S12 ribosomal protein) of this subunit can also result in streptomycin resistance.14 Both of these mutations cause a decreased binding affinity of streptomycin with its normal binding site on the 16S sub-subunit.
Obviously then, those bacteria with fewer copies of the ribosomal genes would achieve resistance much faster than those having multiple copies of these genes. This is exactly what happens. For example, M. tuberculosis only has a single copy of the rRNA genes and yet it can go from a streptomycin resistant rate of 2% to over 80% within three or four months time (if streptomycin is the sole antibiotic of use).15 However, E. coli, although much faster growing than M. tuberculosis, are far less likely to develop resistance because they harbor multiple copies of the rRNA genes. Multiple copies of genes and a rapid growth rate means that multiple independent mutations are needed to prevent streptomycin susceptibility. When two or more independent point mutations are required for resistance, this takes much more time to achieve than a single mutational event that instantly results in streptomycin resistance.
Macrolide antibiotics, such as erythromycin, are similar to aminoglycosides in that their target activity involves the ribosomal complex. However, instead of attacking the 30S subunit, they attack the larger 50S subunit (specifically at the peptidyl transferase center in the 23S sub-subunit). Since erythromycin is produced naturally by bacteria such as Streptomyces erythreus, it is no surprise then that a specific gene also exists within various bacteria that inhibits erythromycin function. This gene is known as the "erythromycin resistance methylase gene" or "Erm" gene. This gene either mono- or di-methylates "adenine 2058" in the peptidyl transferase loop of the rRNA.17 This methylation blocks the erythromycin affinity and restores the fidelity of the ribosome. Erythromycin resistance that involves the Erm gene does not occur through point mutation, but through an isolation of those bacteria that previously had the Erm gene present in their gene pool. However, even without the Erm gene, all is not lost for bacterial colonies under macrolide attack. Macrolide resistance can be also achieved by a point mutation that involves that replacement of an adenine molecule at position 2058 with either guanosine (G), cytosine (C), or uracil (U). The replacement of adenine at two other locations (G2057A and A2059G) is also known to result in macrolide resistance. 16
So, macrolide resistance can be achieved either
through the previous existence of the Erm gene or through a spontaneous point
mutation to three different sites within the 23S sub-subunit of the ribosomal
complex. As would seem intuitive, the Erm gene and
its variants are found in rapidly growing bacteria, such as Staphylococcus
and Streptococcus, while spontaneous mutations of the 2058 site and the
two other related sites usually occurs in species that contain a small number of
23S rRNA genes (such as Mycobacterium). As
with aminoglycoside resistance, rapidly growing bacteria require higher levels
of protein production and thus need more rapid protection of most of their rRNA
genes and gene products. Random point mutations take
too much time to build up in a population with such needs.
However, slower growing bacteria with few copies of rRNA genes can achieve full
resistance with just one point mutation.18
(Back to Top)
(Back to Top)
Bacterial Nucleic Acid
Replication and Repair
Like protein synthesis, DNA replication and
repair is extraordinarily complicated.
However, presented here are a few basic steps that will give a general
idea of the processes involved, how antibiotics can interfere with these
processes, and how bacteria can evolve resistance to the action of these
antibiotics. (Back to Top)
(Back to Top)
The first step in DNA
synthesis or replication involves the separation of the two strands in the DNA
double helix. Only when separated can each of the
individual strands begin the replication process.
The reason for this is because the replicating enzymes (polymerases) use only
single-stranded DNA as a template.
In prokaryotes, such as bacteria, DNA replication can only begin at a single
discrete site called the, "origin of replication."
As the two strands of DNA are unwound and are separated from each other, they
form a "V" shape where the active replication or copying of the two strands of
DNA occurs. This area is called the, "replication
fork." This fork moves along the DNA molecule as
synthesis occurs in both directions away from the origin of replication.12
A special group of proteins is responsible for maintaining the separation of the DNA strands and for unwinding the double helix ahead of the advancing replication fork. These include the helix-destabilizing (HD) proteins (also called single-stranded DNA binding proteins or SSBs) and DNA helix-unwinding proteins (also called helicases). HD proteins bind to single-stranded DNA in a cooperative manner. That is, the binding of one molecule makes it easier for additional molecules of HD protein to bind. These proteins serve a dual function in that they not only keep the two strands of DNA separated during replication, but also protect the single-stranded DNA from cleavage by nucleases that break apart unprotected segments of single-stranded DNA. The DNA helicases bind to single-stranded DNA near the replication fork and then move into the neighboring double-stranded region, forcing the strands apart. These enzymes require energy in the form of ATP. Once the strands have been separated by the helicases they are then stabilized by the HD proteins.12
The act of separating the strands of DNA causes
a "supercoiling" tension downstream from the replication fork. Unchecked,
this tension would quickly increased and eventually prevent further progression
of the replication fork. However, another group of enzymes, the topoisomerases introduce "swivel" points along the double helix to release the "supercoiling" tension caused by the uncoiling of the helical DNA.
This supercoiling phenomenon can be demonstrated by tying two ends of a
rope together, twisting them to form a double-helix, and then pulling them apart
in the center. The twists in the rope on either side
of the separation will twist more and more until they greatly limit further
separation of the two ropes in the middle. DNA
topoisomerases come in different types. Type I topoisomerases, also
called DNA swivelases, reversibly cuts a single
strand of the double helix thereby allowing the cut strand to swivel around the
uncut strand to relieve the built up tension on the two strands.
After the swiveling has released the tension, the swivelase enzymes can reseal
the cut. Therefore, swivelases have both nuclease (strand-cutting)
as well as ligase (strand-resealing) activities. They do not require ATP but rather store energy from the
phosphodiester bond that they cleaved and use the stored energy to reseal the
strand. Type II topoisomerases, also called DNA
gyrases, bind tightly to both strands of the DNA and make transient breaks in
both strands of the DNA helix. The enzyme then
causes a second stretch of the DNA double-helix to pass through the break and
finally reseals the break. The result is a "negative
supertwist" that allows easier unwinding of the DNA double-helix.12
Once the strands are separated, both of the original or parental strands of DNA serve as a template for the synthesis of compliment strands of DNA. The synthesis of each of these new strands of DNA is the result of DNA polymerases. There is a limitation to how these polymerases read however. Just as the English language is read only from left to right, DNA polymerases can only "read" DNA strands from the 3' to the 5' direction (This is can get confusing, but remember that the polymerases read DNA in the 3' to 5' direction, but form DNA in the 5' to 3' direction). Therefore, beginning with one parental double-helix, the two growing nucleotide chains must grow in opposite directions - one in the 5' to 3' direction toward the replication fork, and the other one in the 5' to 3' direction away from the replication fork. This fairly complicated ability is accomplished by a slightly different mechanism for each strand. The strand that is being copied in the direction of the advancing replication fork is called the "leading strand." Since the copying for this strand is in the same direction of the advancing replication fork, the synthesis of DNA is not interrupted, but occurs continuously. However, the other strand, the "lagging strand," is not so lucky since it must grow away from the direction of the replication fork. It is synthesized discontinuously, with small fragments of DNA being copied near the replication fork. These short stretches of discontinuous DNA, termed "Okazaki fragments," are eventually joined to become a single, continuous strand. However, these are not the only problems that must be overcome by the DNA polymerases.12
Before DNA polymerases can copy DNA, they must be initiated by an RNA polymerase or "primer" called primase. Interestingly enough they cannot initiate DNA synthesis by themselves. The primase primer synthesizes short stretches of RNA (approximately ten nucleotides in length) that are complementary to the DNA template. As seen in the figure, these short RNA sequences are constantly being synthesized on the lagging strand, but relatively few are required for the leading strand. But even before primase can bind to DNA a "prepriming complex" must bind the single stranded DNA first and displace some of the single-stranded DNA binding proteins. Only after the prepriming complex binds can the primase bind. The prepriming complex plus the RNA primase is called the "primosome." 12
Once the primosome is formed on the parent strand of DNA, DNA chain elongation can begin via the catalytic activity of DNA polymerase III. DNA polymerase III first recognizes the 3' hydroxyl group of the RNA primer as the acceptor of the first deoxyribonucleotide. Of course, from this starting point DNA chain synthesis proceeds in the 5' to 3' direction "antiparallel" to the parental strand, which is read in the 3' to the 5' direction. All four deoxyribonuceoside triphosphates (dATP, dTTP, dCTP, and dGTP) must be present in sufficient quantities for DNA elongation to occur. If one of the four is in short supply, DNA synthesis will fail.
Consider also the importance of the 3' hydroxyl
group on the deoxyribose ring. This 3' hydroxyl
group forms a phosphodiester bond with the next nucleotide at its 5' phosphate
group. If a nucleotide with an absent 3' hydroxyl
group is added to a growing DNA chain, further elongation cannot continue.
Certain chemotherapeutic (cytosine arabinoside-araC) and antiviral (araA) agents
take advantage of this fact to slow the division of rapidly growing cancer cells
Of course, it gets even more complicated.
In order to prevent as many copying errors as possible from making it to the
next generation, DNA polymerase III also has a "proofreading" function that
works in the 3' to 5' direction (3' to 5' exonuclease).
As each nucleotide is added to the chain, DNA polymerase III checks to make sure
that the added nucleotide is, in fact, correctly matched to its complementary
base on the template and edits its mistakes. For example, if the template base is adenine and the
enzyme mistakenly inserts a cytosine instead of a thymine in the new chain, DNA
polymerase III hydrolytically removes the misplaced cytosine and replaces it
with thymine. This exonuclease activity is specific
in that it only replaces mismatched nucleotides but never correctly paired
DNA polymerase III continues to synthesize DNA
until it is blocked by a stretch of RNA primer. When this occurs, the RNA is excised and the gap if
filled by a separate DNA polymerase by the name of DNA polymerase I.
In addition to all of the abilities of DNA polymerase III, DNA polymerase I has
an exonuclease activity that removes the sections of RNA primer.
In order to do this, DNA polymerase I first recognizes the space or "nick"
between the 3'-end of the newly synthesized DNA strand and the 5'-end of the
adjacent RNA primer. The final linkage
between the 5'-phosphate group on the DNA chain that was synthesized by DNA
polymerase III and the 3'-hydroxyl group on the chain made by DNA polymerase I
is catalyzed by another enzyme, DNA ligase, plus one molecule of ATP. 12
The surrounding environment is constantly
The damaging agent is usually chemical, but radiation is also a
significant mutagen. Sometimes spontaneous changes
also occur that are unrelated to any other agent.
Such changes or mutations are usually neutral or harmful.
"Beneficial" mutations are extremely rare.
So, it is fortunate that cells are remarkably efficient at damage
control. Most of these repair mechanisms involve
recognizing the lesion, removing the damaged section of DNA, and, using the
sister strand of DNA as a template, filling in the gap left by the excision of
the abnormal section of DNA.
Ultraviolet light specifically causes the
covalent joining of two adjacent pyrimidines (usually thymines), producing a "pyrimidine
These dimers prevent DNA polymerase from replicating DNA beyond the site
of the dimer. In order to remove these dimers a
specialized UV-specific endonuclease recognizes the dimer and cleaves the
damaged strand on the 5' side of the dimer.
Next, an excision exonuclease recognizes the incision made by the
endonuclease. In E. coli, this 5' to 3'
exonuclease activity is associated with DNA polymerase I.
The gap left by the removal of the thymine dimer is filled, using the
sister strand as a template, by DNA polymerase I.
And, as with normal DNA synthesis, the last phosphodiester bond is formed
with the help of DNA ligase. Without this repair
mechanism in place, dimers build up in cells, rapidly resulting in cancers such
as skin cancer. Those with increased rates of such
cancers commonly have a problem with UV-specific endonuclease activity.12
Other forms of damage to DNA occur
For example, over time cytosine slowly looses its amino group.
When this happens, it is no longer cytosine, but uracil.
Other types of damage can be caused by chemical interactions with DNA.
For example, nitrous acid (formed by the cell from precursors such as
nitrosamines, nitrites, and nitrates), removes the amino group from cytosine,
adenine, and guanine bases. Bases may also be
completely lost spontaneously.
For example, around 10,000 purine bases are lost spontaneously per day in
an average cell. DNA lesions involving alteration or
loss must be corrected rapidly in order to keep up with the rapid rate at which
they occur. And in fact, this is what happens.
Abnormal bases are recognized by specific glycosidases that hydrolytically
cleave the base out of the strand. Specific
endonucleases recognize that a base is missing and fill in the gap as usual.
In addition, other endonucleases recognize abnormal bases and initiate
the nucleotide excision without first requiring that the base be cleaved from
its sugar backbone by a specific glycosidase.12
So, vital DNA repair is accomplished by a large
number of unique and highly specific enzymes working in concert.
Their importance is quickly realized if even one or two of them are lost,
because sickness or even cell death quickly follows. (Back to Top)
(Back to Top)
Antibiotics that Specifically Target Bacterial Nucleic Acid Replication and Repair
Fluoroquinolones (ie: ciprofloxacin) are synthetic antibiotics that target both DNA gyrases as well as topoisomerases (type "IV"). The gyrases are heterotetramers (A2B2) whose subunits are encoded by the gyrA and gyrB genes. Topoisomerase IV is also a heterotetramer composed of distinct subunits called ParC and ParE. ParC is a homologue of gyrA while ParE matches gyrB. Specifically, topoisomerase IV is the main enzyme that removes the linking of daughter chromosomes after a round of DNA replication (allows cellular segregation to occur).19
Fluoroquinolones specifically trap doubly cleaved DNA on the gyrase structure. Double-stranded breaks accumulate as a consequence since the gyrase enzymes are now incapable of reattaching the broken strands. This blocks the progress of the replication fork and the cells dies. Fluoroquinolones also stabilize the DNA-topoisomerase IV complex, thus hindering chromosomal separation.19,20 Interestingly enough, gyrases are the primary target in Gram-negative bacteria while topoisomerase IV is the primary target in Gram-positive bacteria.19,21
In Gram-positive bacterial species fluoroquinolone resistance is the result of a single point mutation (although several different point mutations between "residues 67-106" can result in quinolone resistance). The beneficial point mutations generally involve the "quinolone resistance-determining region" (QRDR) of the A subunit of DNA gyrase. The QRDR is located in the N-terminal region of the gyrase protein. This region is located on a specific area of the gyrase enzyme where it bends or distorts the strand of DNA that is cleaved. It is thought that mutations to this area reduce the probability of enzyme-quinolone-DNA complex formation. A mutation in this area, that results in just one amino acid change, can cause a 4 to 8-fold increase in fluoroquinolone resistance.20,21 The theory put forward to explain Gram-negative resistance suggests that specific mutations could change the stability of the topoisomerase-DNA-quinolone complex without affecting the probability of its formation.22 (Back to Top)
As briefly touched on earlier, proteins are not
made directly from the DNA or "master copy." Instead,
"working copies" are needed in the form of messenger RNA (mRNA) molecules.
The process of making mRNA from DNA is called transcription while the
process of making proteins from mRNA is called translation.10
A central feature of transcription is that the process of selecting regions of DNA to copy into mRNA is highly specific. This specificity is primarily due to "signals" that are embedded in nucleotide sequences of the DNA. These signals instruct the RNA polymerase where to start transcription, how often to start transcription, and where to stop transcription. These signals are extremely important because the biochemical differentiation of different cellular structures and ultimately tissues and organ systems is due to the selectivity of this transcription process.10
A second important feature of transcription is
that the mRNA transcripts, which initially are faithful copies of the DNA,
undergo various modifications (terminal additions, base modifications, trimming,
internal segment removal, and splicing) that convert the inactive primary
transcript into a functional molecule.10
In bacteria RNA polymerase there is a
multi-subunit enzyme that recognizes specific nucleotide sequences such as the
â€œpromoter regionâ€ at the beginning of a stretch of DNA.
After it makes a complementary RNA copy of the DNA template, the polymerase
enzyme also recognizes the end or "terminator region" of the DNA sequence that
is to be transcribed. Four of the RNA polymerase
subunits (2a, 1b, and 1b' ) are responsible for the 5' to 3' RNA polymerase
activity. However this core enzyme complex cannot
recognize the promoter region of the DNA template by itself.
This recognition requires the action of the "s-subunit" or "s-factor" that
enables the polymerase to recognize the promoter regions on the DNA.
As far as termination recognition goes, some termination regions are
recognized by the RNA polymerase direction while others require yet another
subunit called the "r-factor" (in E. coli).12
So, there are three basic steps involved in RNA synthesis. These are initiation, elongation and termination.12 Consider the following illustration of these steps:
As previously mentioned, the initiation of transcription involves the binding of RNA polymerase to a promoter region on the DNA. Certain specific promoter sequences are known. One such sequence is called the, "Pribnow box." This is a sequence of seven nucleotides centered around ten bases upstream from the initial base of the mRNA coding sequence. A second nucleotide sequence, "TGTTG" is also recognized by RNA polymerase. This sequence is located about 35 bases upstream. Of course this is based on a prokaryotic model. The structure of eukaryotic promoters is much more complex than this.12
Once the promoter is recognized the RNA
polymerase begins to synthesize the RNA transcript.
Unlike DNA polymerase, RNA polymerase does not require a primer and has no known
endonuclease or exonuclease activities. The binding of RNA polymerase to the
double-stranded DNA template results in a local unwinding of the DNA helix.
Therefore, RNA polymerase does not require additional unwinding enzymes
or helix destabilizing enzymes as does DNA polymerase.
Like DNA polymerase, RNA polymerase uses nucleoside triphosphates and
releases a molecule of pyrophosphate each time a nucleotide is added to the
growing chain. The growing RNA molecule is
synthesized from its 5'-end to its 3'-end, antiparallel to its DNA template (The
only difference between the nucleotides used in DNA and RNA is the substitution
of a uracil (U) in RNA for a thymine (T) in DNA).
The process of elongation continues until a termination signal is
RNA polymerase can sometimes recognize termination regions all by itself.
However, as previously mentioned, the r-factor may be required for the
release of the RNA product as well as the RNA polymerase.
As the figure illustrates, the termination region of the DNA exhibits a
twofold symmetry owing to the presence of a palindrome.
The RNA transcript of the DNA palindrome forms a stable "hairpin turn"
that slows down the progress of the RNA polymerase at this terminator site.
Meanwhile, the r-factor binds to the 5'-end of the RNA transcript and moves
along the newly forming RNA molecule, following the polymerase at some distance
as it forms the RNA. When the polymerase pauses at the terminator site, the
r-factor catches up to the polymerase and causes termination.12 (Back to Top)
(Back to Top)
Antibiotics that Specifically Target RNA Synthesis
Rifampin, a semi-synthetic derivative of
rifamycin, is an important antibiotic in the treatment of mycobacterial
infections (i.e., Tuberculosis).
Rifampin specifically targets the b-subunit of RNA polymerase.
The formation of a drug-enzyme complex inhibits the initiation of chain
formation in the process of RNA synthesis.
Interestingly, the existence of rifamycin-inactivating enzymes has not
been reported (an unusual event for drugs which are natural products).
Thus, antibiotic resistance mechanisms rely solely on mutagenesis of the
Rifampin target. A single mutation to this target
results in resistance. Obviously then, susceptible
organisms, such as mycobacteria, can develop rifampin resistance at incredibly
Specific insertions and deletion mutations of bacterial RNA polymerase
usually occur within 3 short, highly conserved regions of the b subunit, and
encompass the "hot spot" area between residues 505 and 534 (E. coli
numbering).23 New synthetic derivatives
such as rifabutin, rifapentine, and KRM-1648 display strong cross-resistance
with rifampin and also are not any more effective than rifampin.14,24
So, although rifampin is highly effective against M. tuberculosis
infections, it should always be used with at least one other antibiotic.
Actinomycin D also acts by interfering with RNA
synthesis by binding to the DNA template and interfering with the movement of
RNA polymerase along the DNA.12
I am not aware of a target mechanism of resistance to actinomycin D.
(Back to Top)
(Back to Top)
Cell Wall Synthesis
In a bacterium the "cell
wall" is the outermost protective covering (as is illustrated in the figure).
Almost all bacteria have cell walls except for a few, such as the Mycoplasma
species, which are bound by thin cell membrane instead of an actual wall.11 The cell wall is a multi-layered structure located
externally to the cytoplasmic membrane. It is
composed of an inner layer of peptidoglycan surrounded by an outer membrane that
varies in thickness and chemical composition depending upon the bacterial type.
The peptidoglycan layer provides structural support and maintains the
characteristic shape of the cell.11
The structure, chemical composition, and
thickness of the cell wall differ in gram-positive and gram-negative bacteria.
Perhaps the most obvious difference is the dramatically thicker peptidoglycan
layer of gram-positive bacteria. Some gram-positive
bacteria also have a layer of teichoic acid outside the peptidoglycan layer
whereas gram-negative bacteria do not.
There are many other key differences and important features between
gram-positive and gram-negative bacteria, but since antibiotics primarily target
the peptidoglycan layer, we will concern ourselves primarily with its structural
Peptidoglycan is a complex, interwoven network of sugars and proteins that surrounds the entire cell wall. It is found only in bacterial cell walls. It provides rigid support for the cell, is important in maintaining the characteristic shape of the cell, and allows the cell to withstand environments with low osmotic pressure - such as pure water.
The term, "peptidoglycan" is derived from the peptides and the sugars (glycans) that make up the molecule. The peptidoglycan molecule has a carbohydrate backbone composed of alternating N-acetylmuramic acid (NAM) and N-acetylglucosamine (NAG) molecules. Attached to each of the muramic acid molecules is a tetrapeptide consisting of both D- and L-amino acids - the precise composition of which differs from one bacterium to another.
Two of these amino acids, diaminopimelic acid
and D-alanine, are worthy of special attention.
Diaminopimelic acid is unique to bacterial cell walls and D-alanine is involved
in the cross linking of the
tetrapeptides (and is specifically targeted by penicillin type antibiotics).
Note also that this tetrapeptide contains rare D-isomers of amino acids (most
proteins contain the L-isomer). Also, since peptidoglycan is present in
bacteria but not in humans, it is an excellent target for antibiotics.
Without a functional peptidoglycan layer, the bacteria swell due to osmotic
pressures and rupture in low osmotic environments.11
The biosynthesis of
peptidoglycan involves a number of cytoplasmic steps, after which the
disaccharide-pentapeptide intermediate is attached to a lipid carrier and is
translocated across the membrane into the bacterial periplasm (The space between
the membrane and the cell wall). In the periplasm,
the glycosyl transfer and transpeptidation activities of "penicillin-binding
proteins" (PBPs), or "transpeptidases", complete the polymerization of the
sugar backbone as well as the cross-linking of the pentapeptide chains (or "stem
Beta-lactam antibiotics (penicillins, cephalosporins, etc.) kill
gram-positive bacteria by inhibiting PBPs - the enzymes that catalyze the final
cross-linking steps in the synthesis of peptidoglycan.25
For example, in Staphylococcus aureus, transpeptidation occurs
between the amino group on the end of the pentaglycine cross-link and the
terminal carboxyl group of the D-alanine on the tetrapeptide side chain.
Since the structure and chemical nature of
penicillin is similar to that of the dipeptide D-alanyl-D-alanine,
penicillin can bind to the active site of the transpeptidases as a "pseudosubstrate"
and inhibit their activity.26 The PBP
active sites, once acylated by penicillin, deacylate very slowly and are
subsequently incapable of performing regular cross-linking functions. Since PBPs
are the enzymes that link the sugar backbone together as well as the cross-links
of the protein "stems" that hold the sugar chains together, anything that
interferes with their action would result in a loss of cell peptidoglycan
integrity. A second factor that comes into play concerns the action of
peptidoglycan hydrolases or "murein hydrolases" (murein is a synonym for
Murein hydrolases degrade peptidoglycan and are naturally produced by
many types of bacteria for the purpose of peptidoglycan upkeep and remodeling.
However, if the action of these hydrolases is limited in some way, the
integrity of the cell wall will remain intact a great deal longer even in the
presence of antibiotics that destroy the building functions of PBPs. For example, if a wall cannot be torn down very easily, it is
not as important that one be able to repair wall damage quickly.
In fact, some bacteria, such as S. aureus, are tolerant to the
action of penicillins due to the fact that autolytic murein hydrolases are not
activated.11 A "tolerant" organism is
one that is inhibited, but not killed by a usually bactericidal antibiotic.
When one thinks of penicillin-type antibiotic
resistance, the enzyme "penicillinase" often comes to mind.
However, there are several other ways that bacteria become resistant to
A notable example is occurs S. pneumoniae.
beta-lactamases have never been identified in S. pneumoniae.
And yet, these bacteria are capable of penicillin resistance due to modification
of their PBPs.27,28,29 Many
specific point mutations have been noted to confer penicillin resistance, but
all of them are located in the transpeptidase domain - with high recurrence at
locations "338 and 552". "Thr 338" is immediately
adjacent to the catalytic "Ser 337" and is buried in a small cavity shielded
from the active site.30,31 This cavity also harbors a buried water
molecule, which is coordinated to the side chain hydroxyl group of Thr 338,
among others. The mutation of Thr 338 to glycine, alanine, proline or valine
lowers the acylation efficiency of PBP for b-lactams, even though this residue
is not in direct contact with the antibiotic.
Consequently, in the absence of a crucial hydroxyl group, the buried
water molecule can no longer be stabilized. This reduces the antibiotic
acylation efficiency.32 Further
mutagenesis efforts identified that mutation of "Gln 552" into a glutamate
residue caused a severe reduction in acylation efficiency for both cefotaxime
and penicillin G - probably due to electrostatic effects between the incoming
antibiotic and the residue, which borders the active site.33
Another example of PBP-related antibiotic
resistance is detected in methicillin-resistant
S. aureus (MRSA), which is not naturally transformable and
therefore does not develop resistance by increasing gene mosaicity (as occurs in
In fact, MRSA produces four unmodified PBPs as well as an additional high
molecular weight form (PBP2'), which has low affinity for almost all b-lactam
molecules in use and is not found in normal Gram-positive isolates.
In the presence of methicillin, the cell wall of MRSA is altered,
generating a peptidoglycan with fewer oligomeric peptides. This new murein
structure is believed to be a direct consequence of the fact that PBP2' takes
over the regular peptidoglycan transpeptidation functions normally performed by
methicillin-sensitive PBPs.34 The gene coding for PBP2' (mecA)
is found in the same chromosomal location in all MRSA isolates, an observation
which suggests that its introduction occurred only once.23,26
Glycopeptides, such as vancomycin, the famous antibiotic of "last
resort", do not target PBPs, but instead they interfere with peptidoglycan
synthesis by binding to the terminal D-Alanyl-D-Alanine (D-Ala-D-Ala) terminus
of stem peptides.
Binding of the glycopeptides to the C-terminal ends of the "stem"
peptides impedes their subsequent recognition by transpeptidases (PBPs).
Obviously then, peptidoglycan cross-linking is blocked.
In addition, it has also been suggested that glycopeptides block the
transfer of peptidoglycan precursors through the cytoplasmic membrane by steric
hindrance.35 This blockade causes a
hindering buildup of peptidoglycan precursors in the cytoplasm of the cell.
The complex between vancomycin and D-Ala-D-Ala involves a set of five crucial hydrogen bonds between the elongated backbone of vancomycin and the stem peptide. As would be expected, vancomycin resistance involves a single structural change in the cell wall where the D-Ala-D-Ala portion is replaced with D-ala-D-Lactate. This inclusion of an ester instead of an amide bond in the stem peptide structure profoundly alters its interaction with the vancomycin "backbone" structure in that there is a loss of one of the five hydrogen bonds. The resulting binding affinity is three orders of magnitude lower than it was with D-Ala-D-Ala, and bacterocidal activity is lost.36
In several bacterial species, such as in Enterococcal species (VRE), vancomycin resistance is achieved through the function of a five-gene operon that is located on a transposon (mobile genetic element). The mobility of this element suggests that this resistance mechanism could be acquired by a large number of pathogenic bacteria. The five genes include three structural genes named, vanH, vanA, and van X.36 VanH converts pyruvate to D-lactate. VanA takes the D-lactate produced by vanH and synthesizes D-Ala-D-Lactate with it. VanX selectively hydrolyzes all D-Ala-D-Ala peptides that are naturally produced by the host. Without the availability of D-Ala-D-Ala, the concentration of D-Ala-D-Lactate builds up in the cell and the resulting peptidoglycan is made entirely of D-Ala-D-Lactate - given the host resistance to vancomycin.37 Of side interest are the two other proteins coded by the operon - vanS and vanR. VanS is a sensor kinase and vanR is a regulator that controls transcription. VanS senses the presence of vancomycin and transmits this message to vanR via autophosphorylation and phosphoryl transfer. VanR then upregulates the production of D-Ala-D-Lactate. 35,37
In this light, it is interesting to note that Enterococcal bacteria do not ever gain vancomycin resistance spontaneously in an isolated gene pool. The VanHAX genes happen to be on a transposable element in VRE. And, it just so happens that these genes, in the same orientation, are also found in various bacteria that historically synthesize and secrete vancomycin. Other types of bacteria, like E. coli also have very similar genes.38,39 Such similarity is thought to represent an evolutionary relationship.
In short, vancomycin resistance did not evolve in real time. The genes coding for vancomycin resistance were already in the bacterial gene pool before humans started using vancomycin as an antibiotic. VRE did not evolve over the course of 15 or so years of vancomycin use in Enterococcal bacteria. They simply gained vancomycin resistance via lateral transfer of a pre-formed genetic element that was already there in the gene pool. Exactly the same thing happened with penicillin resistance that results from penicillinase production. The penicillinase gene does not evolve in real time. It is only inherited via vertical or horizontal transfer - preformed. (Back to Top)
There is one other cell wall component that we have not yet
discussed which is a very important aspect of mycobacteria.
Mycobacteria have a similar cell wall to gram-positive bacteria, but in
addition, they have an outer leaflet that contains mostly mycolic acids.
Mycolic acids are long chain fatty acids that are covalently linked to an
"arabinogalactan" polysaccharide layer.
The polysaccharide layer is in turn linked to the underlying
peptidoglycan structure. Mycolic acids provide a
formidable barrier against most antibiotics and are essential for bacterial
viability.11 Antibiotics that affect
mycolic acid stability are key to fighting mycobacterial infections.
Antibiotics, such as isoniazid, are thought to
disrupt the biosynthesis of mycolic acids through the inhibition of "inhA," an enzyme (enoyl-ACP reductase) essential in the biosynthesis of long chain
fatty acids and thus mycolic acids.39,40
Before isoniazid can act, it must be activated by the
catalase-peroxidase enzyme coded by the KatG gene.
Once activated, isoniazid can bind to a target site on the inhA enzyme
and inactivate it. Thus, isoniazid resistance is
often the result of a single point mutation in the inhA gene.
This mutation changes the serine occupying position 94 in the inhA enzyme
to an alanine. This change lowers the enzyme's
affinity for a vital NADH molecule by 5 fold.
It just so happens that the activated isoniazid molecule covalently
attaches to the NADH molecule that is located within the inhA enzyme's active
site. If the NADH molecule becomes more loosely
attached to this enzyme, the isoniazid antibiotic along with the NADH molecule
fall away from the enzyme and the bacterium gains immunity.
This same mutation also gives resistance to isoniazid's structural
Restriction of Antibiotic Access to the Target
There are several physical restrictions that
limit antibiotic access to a bacterial target.
Several of these have been mentioned already, to include the thick mycolic acid
layer of protection produced by mycobacteria, and an outer lipid membrane
produced by gram-negative bacteria.11
Such structures limit a large variety of antibiotics from reaching their
For gram-negative bacteria in particular glycopeptides are specifically
limited. Glycopeptides have limited access to their
peptidoglycan precursor targets in gram-negative bacteria because glycopeptides
have large hydrophobic structures in their molecular makeup that cannot readily
cross the gram-negative cell's outer membrane.35
A more complex method of limiting antibiotic
access to target sequences is found in cases of macrolide resistance where the
macrolide antibiotics (like erythromycin and azithromycin) are actually "pumped"
out of the cell.
The pump can get rid of the antibiotic faster than it can accumulate in
the cell. One type of pump is produced by the "MefE"
gene. This gene produces a 12-trasmembrane-helix
macromolecule which exports 14 and 15-membered macrolides from the cell - giving
the cell resistance to antibiotics such as erythromycin.
In fact, as many as 85% of erythromycin-resistant S. pneumoniae strains
harbor the MefE gene.18,41
Tetracycline resistance is also the result of a failure of the antibiotic to reach an inhibitory concentration inside the bacterium. This is due to plasmid-encoded processes that either reduce uptake of the antibiotic or enhance the antibiotic's transport out of the cell. 11
Likewise, one of two methods of sulfonamide
resistance is mediated by a plasmid-encoded transport system that actively
exports the antibiotic out of the cell.
The other mechanism involves a point mutation in the gene coding for the
target enzyme "dihydropteroate synthetase", which reduces the binding affinity
of the antibiotic.11
Quinolone resistance, although often the result
of mutations that modify DNA gyrase, can also be the result of changes in outer
membrane proteins that result in reduced uptake of the antibiotic.11
Inactivation of the Antibiotic
Perhaps the most famous example of direct
inactivation of an antibiotic by bacteria involves the neutralization of
penicillin and penicillin-like antibiotics via the action of beta-lactamases.
beta-lactamases can be divided into four major groups based on primary structure
alignments, molecular size, and active sites.42,43
Class A enzymes harbor a serine in their active site and have an
approximate molecular weight of 30Kda. Also, they
are usually plasmid-encoded and produce the "TEM-1" enzyme.26
Class B enzymes are Zn2+-metalloenzymes,
and usually exhibit a broad spectrum of activity.
Class C b-lactamases are chromosomally encoded, and, like their class D
counterparts, also harbor a serine in their active sites.
Most beta-lactamases produced by gram-positive species are class A
enzymes.25 Approximately 90% of all
beta-lactam resistant S. aureus produce beta-lactamases with the
structural and regulatory genes (blaZ, blaI, and blaR1) being harbored by a
The active sites of class A b-lactamases are very similar to those of PBPs. This similarity poses a challenge for designing new beta-lactam antibiotics, which might be less sensitive to beta-lactamase inactivation, but must still be specifically bound by the PBP active site.34,44 Significant changes to the antibiotic's structure might avoid inactivation by beta-lactamases, but at the same time it would not be able to bind to its PBP target site. This creates what might be called a "win the battle but loose the war" situation. (Back to Top)
Aminoglycoside Resistance Through
Resistance to aminoglycosides is achieved
primarily through chemical modification of the antibiotic.
Three main classes of aminoglycoside-modifying enzymes have been identified in
O-phosphotransferases (APHs), which add on phosphate groups; N-acetyl
transfereases (AACs), which acetylate amino groups on the antibiotics, and
nucleotidyl transferases (ANTs), which add "AMP-moieties." 17,45
Antibiotics that have are attached by strategically placed groups of various
kinds no longer have affinity for the target RNA molecule and are thus incapable
of interrupting protein synthesis.
Research done by those
like James R. Knox et al. is commonly used to support the theory that the
penicillinase enzymes evolved from existing bacterial components.
Knox proposes that the penicillinase enzyme is the result of the
duplication of the D-Ala-D-Ala ligase protein that penicillin binds to.
This explains the substrate specificity of the penicillinase enzyme.
X-ray crystallography clearly demonstrates the striking "homologous
three-dimensional structures that extend well beyond the beta-lactam binding
site" - as well as similar catalytic activities.
Knox goes on to say, "From these results, it seems likely that the
beta-lactamase evolved from an enzyme with DD-peptidase activity." 8
Of course, a hypothesis that "seems likely" is very much different from one that is actually testable in the lab. A lot of things might seem likely, but this is what makes something "hypothetical." All hypothesis seem likely on paper, but until they are tested, they remain hypothetical as far as the scientific method is concerned. Knox showed some seemingly reasonable hypothetical evolutionary pathways that could give rise to the penicillinase enzyme, but none of these evolutionary pathways or portions thereof was ever demonstrated by direct experimentation. If put to the test, they just might run into a few snags along the way. . .
For example, it seems that not just one, but
several key changes are needed for "DD-peptidases" to achieve "beta-lactam
hydrolysis." It seems like a key substitution at "position
69" and a hydrogen bond from B3 to Asn 170 in the W-loop are fairly important
for lactamase activity. Knox's own evolutionary path for the simpler
"Class C" beta-lactamases from a DD-peptidase requires an additional mutation in
the YXN sequence. The more complicated "Class A" beta-lactamases require a
mutation in the SXN sequence plus the addition of a general base on the W-loop.
Knox describes some other "important differences which exist between the two
molecules [DD-peptidase and beta-lactamase] [that] must be accommodated by the
new binding site." These include the following: "Bonds about the nitrogen
atom of the b-lactam ring are nonplanar; the C6 b-acylamido group of the
b-lactam ring is quite apart in direction from the corresponding peptide group
of the DD-peptide; and the beta-lactam lacks a methyl group corresponding to the
penultimate D-methyl group of the DD-peptide." 8
Some might argue that DD-peptidases have a very small amount of beta-lactamase activity to begin with. However, this is not the point. The point is that it is not clear that each and every mutation needed to achieve selectively advantageous beta-lactamase activity is selectively advantageous. In other words, there might be some neutral gaps in function that need to be crossed. Of course, potential neutral mutations would not be selectively advantageous and therefore would take much more time to randomly occur - especially if the neutral gap were more than two or three mutations wide. Consider the follow amino acid sequences structures for a DD-peptidase and a beta-lactamase:
Refined crystal structure of beta-lactamase from citrobacter freundii Class - Alpha Beta 46,47
The refined crystallographic structure of a dd-peptidase penicillin-target enzyme.45,47
Carl Wieland, Antibiotic Resistance in Bacteria, Creation
Ex Nihilo Technical Journal, 8:1 (1994), p. 5.
Ewan More, Rats! Another Case of Sickle-Cell Anaemia,
Creation Ex Nihilo, 17:2 (March-May 1995), pp. 44-45.
Francisco J. Ayala, The Mechanisms of Evolution,
Scientific American, 239 (September 1978), p. 65.
R. McQuire, Eerie: human Arctic fossils yield resistant
bacteria, Medical Tribune, 12/29/1988, pp. 1, 23.
Maceirj Giertych, Professor of Genetics Says "No!" to
Evolution, Creation Ex Nihilo, 17:3 (June-August 1995), pp. 47-48.
Lewis Kleinsmith and Valerie Kish, Principles of Cell and
Molecular Biology, New York: Harper-Collins College Publishers, 1995, pp.
Felix Konotey-Ahulu, The Sickle Cell Disease Patient (New York:
Macmillan, 1991), p. 106-108.
Knox JR, Moews PC, and Frere JM, Molecular evolution of bacterial
beta-lactam resistance, Chemistry & Biology 3, 1996, p. 937-947.
B.G. Hall, Evolution on a Petri Dish.
The Evolved B-Galactosidase System as a Model for Studying
Acquisitive Evolution in the Laboratory, Evolutionary Biology, 15(1982):
Lewin, Benjamin, Genes V, Oxford University Press and Cell Press,
Livinson, W.E., Jawetz, E., Medical Microbiology and Immunology 3rd
Ed. Appleton & Lange, 1994.
Champe, Pamela C., Harvey, Richard A., Biochemistry, Lippincottâ€™s
Illustrated Reviews, J.B Lippincott Co., 1987.
Chambers, H. F.; Sande, M. A. Goodman & Gilman's The
Pharmacological Basis of Therapeutics, 9th Ed. International Edition.1996,
Blanchard, J. S. Annu. Rev. Biochem., 1996,
Mandell, G. L.; Petri Jr, W. A. Goodman & Gilman's The
Pharmacological Basis of Therapeutics, Ninth Edition, International Edition
pp. 1155-1174. J.G. Hardman, L.E. Limbird (eds) 1996.
Ross, J. I.; Eady, E. A.; Cove, J. H.; Jones, C. E.; Ratyal, A. H.;
Miller, Y. W.; Vyakrnam, S.;Cunliffe, W. J. Antimicrob. Agents
Chemother., 1997, 41, 1162-1165.
Walsh, C. Nature, 2000, 406, 775-781.
Gram-positive Pathogens pp. 694-710. V.A. Fischetti et al. (eds),
, Am. Soc. Microbiol., Washington, D.C. 2000.
19. Hooper, D. C. Clin. Infect. Dis., 31 Suppl 2, S24-S28. 2000.
Hooper, D. C.
In Gram-Positive Pathogens pp. 685-693. V.A. Fischetti et al.
(eds) Am. Soc. Microbiol., Washington, D.C., 2000.
Ferrero, L.; Cameron, B.;
Crouzet, J. Antimicrob. Agents Chemother., 39,
Hiasa, H.;Shea, M. E.
J. Biol. Chem., 275, 34780-34786, 2000.
Spratt, B. G.
Science, 264, 388-393, 1994.
Williams, D. L.; Spring,
L.; Collins, L.; Miller, L. P.; Heifets, L. B.; Gangadharam, P. R.; Gillis, T.
P. Antimicrob. Agents Chemother., 42, 1853-1857,
Frare, J. M.
Mol. Microbiol., 16, 385-395, 1995.
Spratt, B. G.
Bacterial cell wall pp. 517-534. J.-M. Ghuysen and R. Hakenbeck (eds),
Elsevier Science B.V., Amsterdam, The Netherlands,
Ghuysen, J. M.; Dive, G.
In Bacterial Cell Wall
pp. 103-129. J.M. Ghuysen and R. Hakenbeck (eds), Elsevier Science, The
Di Guilmi, A. M.; Mouz,
N.; Andrieu, J. P.; Hoskins, J.; Jaskunas, S. R.; Gagnon, J.; Dideberg,
O.;Vernet, T. J. Bacteriol.,
180, 5652-5659, 1998.
Di Guilmi, A. M.; Mouz,
N.; Martin, L.; Hoskins, J.; Jaskunas, S. R.; Dideberg, O.; Vernet, T.
J. Bacteriol., 181, 2773-2781, 1999.
ParÃ¨s, S.; Mouz, N.;
PÃ©tillot, Y.; Hakenbeck, R.;Dideberg, O. Nature Struct. Biol.,
Gordon, E.; Mouz, N.;
DuÃ©e, E.;Dideberg, O. J. Mol. Biol., 299,
Mouz, N.; Gordon, E.; Di
Guilmi, A. M.; Petit, I.; Petillot, Y.; Dupont, Y.; Hakenbeck, R.; Vernet, T.;
Dideberg, O. Proc. Natl. Acad. Sci. USA, 95,
Mouz, N.; Di Guilmi, A.
M.; Gordon, E.; Hakenbeck, R.; Dideberg, O.;Vernet, T. J. Biol.
Chem., 274, 19175-19180, 1999.
Kernodle, D. S.
In Gram-Positive Pathogens V.A. Fischetti et al. (eds), Am.
Soc. Microbiol., Washington, D.C., 2000.
Fraimow, H. S.; Courvalin,
P. In Gram-positive Pathogens pp. 621-634. V.A.,
Fischetti et al. (eds), Am. Soc. Microbiol., Washington, D.C.,
Walsh, C. T.; Fisher, S.
L.; Park, I. S.; Prahalad, M.;Wu, Z. Chem. Biol., 3,
Lessard, I. A.; Walsh, C.
T. Proc. Natl. Acad. Sci. USA, 96, 11028-11032,
Roper, D. I.; Huyton, T.;
Vagin, A.; Dodson, G. Proc. Natl. Acad. Sci. USA, 97,
Banerjee, A.; Dubnau, E.;
Quemard, A.; Balasubramanian, V.; Um, K. S.; Wilson, T.; Collins, D.; de Lisle,
G.; Jacobs, W. R. Science,
263, 227-230, 1994.
Quemard, A.; Sacchettini,
J. C.; Dessen, A.; Vilcheze, C.; Bittman, R.; Jacobs, W. R.;Blanchard, J. S.
Biochemistry, 34, 8235-8241, 1995.
Tait-Kamradt, A.; Clancy,
J.; Cronan, M.; Dib-Hajj, F.; Wondrack, L.; Yuan, W.;Sutcliffe, J.
Antimicrob. Agents Chemother., 41, 2251-2255,
Massova, I.; Mobashery,
S. Antimicrob. Agents Chemother., 42, 1-17,
Matagne, A.; Dubus, A.;
Galleni, M.; FrÃ¨re, J. M. Nat. Prod. Rep., 16,
Joris, B.; Ghuysen, J.
M.; Dive, G.; Renard, A.; Dideberg, O.; Charlier, P.; Frere, J. M.; Kelly, J.
A.; Boyington, J. C.; Moews, P. C. et al. Biochem.
J., 250, 313-24, 1988.
Shaw, K. J.; Wright, G. D. Gram-Positive
Pathogens pp. 635-646. V.A. Fischetti et al. (eds), Am. Soc.
Microbiol., Washington, D.C., 2000.
48. Sophie Maisnier-Patin, Dan I. Andersson, Adaptation to the deleterious effects of antimicrobial drug resistance mutations by compensatory evolution, Research in Microbiology 155 (2004) 360-369 (weblink)
49. Ivan A. D. Lessard and
50. Lai-King Ng, Irene Martin, Gary Liu, and Louis Bryden, Mutation in 23S rRNA Associated with Macrolide Resistance in Neisseria gonorrhoeae, Antimicrobial Agents and Chemotherapy, September 2002, p. 3020-3025, Vol. 46, No. 9 ( http://aac.asm.org/cgi/content/full/46/9/3020 )
51. HIV Animation :HIV Info Source access September, 2005
(Back to Top)
. Home Page . Truth, the Scientific Method, and Evolution
. Maquiziliducks - The Language of Evolution . Defining Evolution
. DNA Mutation Rates . Donkeys, Horses, Mules and Evolution
. Amino Acid Racemization Dating . The Steppingstone Problem
. Harlen Bretz . Milankovitch Cycles
Since June 1, 2002