Molecular biology of Streptococcus pneumoniae: an

Res. Microbiol. 151 (2000) 407–411
© 2000 Éditions scientifiques et médicales Elsevier SAS. All rights reserved
Molecular biology of Streptococcus pneumoniae: an everlasting challenge
Michel Sicard*, Anne Marie Gasc, Philippe Giammarinaro, Jacques Lefrançois,
Frank Pasta, Mustapha Samrakandi
Laboratoire de microbiologie et de génétique moléculaire du C.N.R.S. et Université Paul Sabatier, 118, route de
Narbonne 31062 Toulouse cedex, France
Abstract – Streptococcus pneumoniae is a model for elucidating: 1) recombination steps of DNA, from its
discovery to polarity of integration; 2) long-patch mismatch repair, short-patch repair triggered by A/G and
exclusion of deletions; 3) resistance to β-lactam antibiotics; and 4) factors of virulence. Several of these topics
remain a challenge for future investigations. © 2000 Éditions scientifiques et médicales Elsevier SAS
Streptococcus pneumoniae / recombination / antibiotic resistance / virulence
1. Introduction
At the beginning of the century, pneumococcus was the major cause of death by bacterial
infection. In the USA, more than 50 000 people
died every year. Even today throughout the
world, millions of people are killed every year
by this virulent bacteria [1]. The mortality of
septic shock has been reduced during the last
30 years, but only from 70 to 45% for patients
admitted to intensive care for septicemia, with
an increasing proportion of Gram-positives
(45%) [2]. Septic shock from invasive pneumococci can kill, at a high frequency (10–15%) and
within a few days or hours, patients who had
been in good health. These pneumococci are
frequently not even resistant to antibiotics. This
is why these bacteria have been studied for
more than 80 years, resulting in several major
contributions to biology and requiring a worldwide effort to resolve many scientific mysteries
concerning public health.
2. DNA recombination
It is common practice to quote the famous
publication of Avery and his collegues in 1944
* Correspondence and reprints [email protected]
for the identification of nucleic acids as genetic
determinants. But it should be remembered that
it took 9 years to be accepted by the scientific
community, thanks to the efforts of several
scientists such as Rollin Hotchkiss who intensively purified DNA in order to refute the
criticisms of the biochemist A. Mirsky, who
argued that traces of proteins could account for
the transforming activity. In fact, it is likely that
the criticisms of Mirsky are responsible for the
failure to confer the Nobel prize upon Avery.
Even in 1951, some geneticists did not accept
that bacteria contain genes, since “Mendel laws
cannot be demonstrated” in these organisms [3].
When Watson and Crick defined the structure
of DNA accounting for its fantastic information
content required for genes, the replication and
the mutation of this molecule [4], these debates
became only of historical interest. The field of
genetics had developed over a 40-year period
by hybridization of eucaryotes, but was unable
to answer fundamental questions such as: what
is the structure, the nature, the mode of replication, recombination, mutation and expression of
the gene? Unexpectedly, bacteria solved these
genetic mysteries. The discovery that genes are
DNA was the first molecular biology experiment.
Before 1950 it was generally admitted that
crossing over, a key process in genetics, occurred
M. Sicard et al. / Res. Microbiol. 151 (2000) 407–411
between genes that could not be split. As it was
possible to handle millions of microorganisms
and select rare events, intragenic recombination
was easily demonstrated even between two
base pairs. As early as 1951, Harriett EphrussiTaylor isolated four partially capsule-deficient
mutants. Rare normal encapsulated pneumococci were obtained by reciprocal transformation as one of the first cases of intragenic
recombination. In the middle 1950s, to explain
recombination, two hypotheses were put forward: breakage and reuniting of DNA molecules without DNA replication or copy-choice
in which the replication machinery would shift
from one parental chromatid to the other. In
1960, Maury Fox demonstrated that pneumococcal transformation occurred without significant DNA synthesis showing that the model of
breakage and reunion between DNA molecules
was correct [5]. Almost at the same time [6],
Sandy Lacks observed that donor DNA is singlestranded inside the recipient bacteria before its
integration in the chromosome. This was the
first evidence for hybrid DNA during recombination, a step proposed much later to account
for gene conversion in fungi.
Obviously, most of our understanding of the
steps and enzymes of recombination results
from experiments on Escherichia coli. However,
concerning some specific items, Streptococcus
pneumoniae contributed significantly to this
knowledge. For example, what is the process of
recombination of long heterologies? Are they
excluded, or integrated as point mutations, or
are they submitted to some replication or repair
process? In pneumococci the efficiency of integration of deletions or insertions is fairly high as
long as their size is much smaller than the
donor DNA molecule. This is accounted for by a
normal pairing between the surrounding
homologous sequences. However, we have
observed that, when these deletions are transformed in a two-point cross, the frequency of
wild-type recombinants is tremendously
increased (more the ten-fold). Franck Pasta was
able to show that this observation is accounted
for by the exclusion of the donor heterologous
DNA at a frequency of 20% [7]. Taking advan-
Figure 1. Model accounting for the preponderance of 5’ site
tage of the interruption in recombination by
deletion, he compared the transformation efficiencies of the segments 5’- and 3’-ward from
this deletion. Using the refined tools of molecular genetics of the ami locus, artificial heteroduplexes containing the deletion, PCR amplification of the DNA from individual transformants,
restriction enzyme analysis to detect which
strand and which side of the deletion had
recombined, Franck Pasta found that in vivo the
5’ side of donor single-strand DNA is strongly
(80%) favored by recombination [8] (figure 1).
Therefore, the strand exchange is polarized from
5’ to 3’. It will be interesting to see if this
polarity can be found in other organisms.
3. DNA mismatch repair
A major contribution of pneumococcal studies was the discovery and understanding of
mismatch repair. One of the original observations was the isolation by Harriett EphrussiTaylor in 1959 of an optochin-resistant mutant
transforming eight-fold less than the standard
reference marker. She proposed that this mutation was a long deletion or insertion, or as an
alternative, that there were some weak points
on the DNA molecule near the position of this
mutation. How to test these possibilities? A
genetic map of such a gene could give an
answer but it was not possible to make it. We
decided to change a locus and to find a gene
where wild-type recombinants could be recovered, as well as mutant strains. We were able to
find such mutations that confer resistance to
amethopterin but could not grow in a defined
M. Sicard et al. / Res. Microbiol. 151 (2000) 407–411
medium containing an excess of isoleucine versus leucine and valine [9]. Having constructed
such a system, hundreds of mutants were isolated either spontaneously or by mutagenic
treatments. We could perform a one-point cross
in both directions and two-point crosses to
build a genetic map. Unexpectedly, the poorly
transformable mutants were not long heterologies, as they were spread all along the gene.
Jean-Pierre Claverys, Pedro García and AnneMarie Gasc showed that they were single-site
transitions or ± 1- to 3-base deletions [10]. Sandy
Lacks found the same results in the amylomaltase locus [11]. In 1966 Harriett Ephrussi-Taylor
[12] proposed that low efficiency markers were
excised and eventually one tenth of them would
integrate by a repair process. This explanation
resulted from the observation that optochinresistant mutations disappeared after their entry
into the cell and by analogy to the excision
repair of E. coli DNA after UV irradiation. Much
work has been devoted to confirming this daring hypothesis by J.M. Louarn, J.P. Claverys, G.
Tiraby and Anne-Marie Gasc in my laboratory
and in Sandy Lacks laboratory, especially by the
isolation of strains unable to discriminate
between high and low efficiency mutants, opening the way for cloning and sequencing the hexA
and hexB genes.
Gérard Tiraby demonstrated in 1973 that this
repair system was also antimutagenic [13]. Later
it was found in other bacteria, in yeast and in
man where these genes protect against specific
cancers. It should be emphasized how long it
took from the original unusual observations of
low efficiency markers by several scientists in
1959, and the studies of the first discriminating
strain by Green and Ravin in 1959 [14], to the
characterization of the specificity of the repair
system in 1981, and eventually to the description of the anticancerous properties of these
repair genes in 1993.
At the same time as Peggy Lieb described the
short-patch repair system of the A/G mismatch,
Lefèvre observed a 12-base-long repair for this
mismatch in pneumococcus [15] that is coded
by a mutY homologue, since Franck Pasta, who
cloned and mutated it, found that it complements the E. coli mutY gene.
4. β-Lactamine resistance
Penicillin-resistant mutants were obtained by
Rollin Hotchkiss in 1951 [16], i.e. 16 years before
the first hospital-resistant isolates. Most resistant mutants could be accounted for by a modification of penicillin-binding proteins (PBPs) as
demonstrated by Regine Hakenbeck, Alex
Tomasz and their collegues [17]. By serial transformation of the wild type by DNA from highly
resistant strains, we showed that genes coding
for PBP 2x and PBP 3 yielded increased resistance. Moreover, transformants to higher resistance were not modified in their PBP pattern.
Eric Guenzi in Regine Hakenbeck’s laboratory
showed that this is due to mutations in two
genes that encode a sensor protein belonging to
the family of signal-transducing histidine
kinases and a regulator responding to an environmental signal [18]. Philippe Giammarinaro
in my laboratory has shown that Ca++ is the
environmental signal of this system (ciaR/ciaH)
which is likely to affect cell-wall structure and
competence genes [19].
An unexpected observation was that a mutation of this system not only increases cefotaxime
resistance but blocks competence. Indeed, competence is not fully understood, even though
the startling discovery by Pakula in Poland in
1962 [20] of a competence hormone and the
characterization of its structure by Don Morrison, Sive Harvestein and their collegues in 1995
[21] tremendously contributed to the knowledge of DNA uptake in pneumococcus. In my
laboratory, Jacques Lefrançois [22], using electropermeation, found that it is not sufficient to
introduce chromosomal DNA inside the cell to
obtain recombinants even if we induce natural
competence. The integration of markers in the
chromosome in naturally competent cells must
require DNA processing during entry, which is
still to be discovered.
While it is well demonstrated that the dissemination of penicillin-resistant isolates is
directly related to the use of this antibiotic, the
M. Sicard et al. / Res. Microbiol. 151 (2000) 407–411
process of bacterial migration is not well understood. Using a very discriminating method,
FIGE analysis, Anne-Marie Gasc in my laboratory has shown that most of the 9V French
isolates, which were either resistant or susceptible to penicillin, are clonal [23]. What could be
the selective advantage of this clone in being
able to invade Spain, France and other countries
in a few years?
5. Factors of virulence
In the early 1920s, compelling evidence had
accumulated that virulent pneumococci are
encapsulated, appearing as smooth colonies on
plates, whereas nonencapsulated strains are not
virulent and are rough. At the same time, unexpectedly, Oswald Avery and Michael Heidelberger were able to show that this capsule was
not a protein but polysaccharides that protect
the bacteria from phagocytosis [24]. In 1949,
using intermediate mucoid colonies of type III,
Harriett Ephrussi-Taylor in Avery’s laboratory
reported that there is a direct relationship
between virulence and the size of the capsule
[25]. An average three-fold reduction in the
amount of polysaccharide is enough to shift by
a factor of six to seven logs the quantity of
bacteria required to kill approximately 50% of
mice [26]. Since then, scores of genes involved
in the biosynthesis of several capsular types
have been characterized. Why is virulence so
highly sensitive to relatively small differences in
the amount of capsular polysaccharides? How
does the capsule protect against phagocytosis?
Why is the capsule a barrier to DNA uptake?
Deficiency in several proteins was reported to
attenuate virulence without suppressing it, in
contrast to capsule deficiency that blocks virulence. However, significant conclusions require
isolation of well-characterized mutants, comparison between isogenic strains and reproducibility in other laboratories. The refined molecular genetics tools now available should be
intensively used. Autolysin coded by lytA is one
of these factors that partially affects virulence in
mouse models [27, 28]. Virulence of lytA– strains
can be further reduced by the addition of a
mutation of a gene involved in calcium transport [29]. The effects of autolysin, however,
seemtobemediatedbythereleaseofpneumolysin, another important virulence factor. Virulence
is reduced in pneumolysin-negative strains as
well as in autolysin-negative strains [28]. Recent
studies suggested that pneumolysin stimulates
nitric oxide production from macrophages more
efficiently than crude cell-wall preparation [30].
It was possible to isolate pneumolysin mutants
affected in either one of their dual functions:
lysis or complement activation. Strains producing competent activation-deficient pneumolysin
are fully virulent, whereas lysis-deficient pneumolysin significantly reduces virulence [31].
Although pneumolysin-deficient mutants are
still partially virulent, it seems that inflammation, a major step during invasion, is triggered
by this molecule. Obviously, other factors of
virulence remain to be discovered. A survey of
virulence-negative mutants was performed by
Daniel Simon and coworkers [32] by random
inactivation of genes. Mutation of several new
Figure 2. Hybridization using cap3A probes on type 3 DNA
restricted by EcoRI. Lane 1: original encapsulated strain; lane 2:
the same strain after five passages on mice; lane 3: the same
encapsulated strain transferred 35 times on Petri dishes; lanes 4
and 5: two capsule-deficient strains isolated on Petri dishes after
35 transfers.
M. Sicard et al. / Res. Microbiol. 151 (2000) 407–411
genes reduced septicemia, opening the way to
future investigation. With Mustapha Samrakandi we initiated a program to detect genomic
modifications resulting from loss of virulence
when virulent strains were subcultured in vitro
by serial transfers. This procedure as well as
prolonged storage in the refrigerator had already
been described in the 1920s as selecting virulentdeficient mutants. By FIGE or restriction enzyme
analysis using several probes (PspA, lytA, recA,
ply, nanaA, hyal, cap3A), deletions or amplifications could not be detected. A modification
was, however, observed in a capsule-deficient
strain obtained by this method (figure 2). It
seems likely that by using random probes, new
virulent genes can be found. It may also be
possible that invasive septicemia could result
from even limited amplifications of capsular
loci, an unstable structure in serial subcultures.
It could account for the origin and the end of
epidemics, which are still a mystery. Thanks to
molecular genetics tools, future generations of
microbiologists may hopefully throw light on
[1] McCarthy M. (Ed.), The Transforming Principle. Discovering that
genes are made of DNA, W.W. Norton and Co, New York,
London, 1985.
[2] Friedman G., Silva E., Vincent J.L., Has the mortality of septic shock
changed with time?, Crit. Care Med. 26 (1998) 2078–2086.
[3] Hotchkiss R.D., The identification of nucleic acids as genetic determinants, Ann. N.Y. Acad. Sci. 325 (1979) 321–342.
[4] Watson J.D., Crick F.H.C., The structure of DNA, Nature 171
(1953) 737–738.
[5] Fox M., Fate of transforming deoxyribonucleate following fixation
by transformable bacteria, Nature 187 (1960) 1004–1006.
[6] Lacks S., Molecular fate of DNA in genetic transformation of
Pneumococcus, J. Mol. Biol. 5 (1962) 119–131.
[7] Pasta F., Sicard A.M., Exclusion of long heterologous insertions and
deletions from the pairing synapsis in pneumococcal transformation, Microbiology 142 (1996) 695–705.
[8] Pasta F., Sicard A.M., Polarity of recombination in transformation of
Streptococcus pneumoniae, Proc. Natl. Acad. Sci. USA 96 (1999)
[9] Sicard A.M., A new synthetic medium, and its use for the study of
reciprocal transformation at the amiA locus, Genetics 50 (1965)
[10] Claverys J.P., Méjean V., Gasc A.M., Galibert F., Sicard A.M., Base
specificity of mismatch repair in Streptococcus pneumoniae: relationship between base mismatches and transforming efficiencies, Nucleic
Acids Res. 9 (1981) 2267–2280.
[11] Lacks S.A., Dunn J.J., Greenberg B., Identification of base mismatches recognized by the heteroduplex-DNA-repair system of
Streptococcus pneumoniae, Cell 31 (1982) 327–335.
[12] Ephrussi-Taylor H., Gray T.C., Genetic studies of recombining DNA
in pneumococcal transformation, J. Gen. Physiol. 49 (1966) 211–231.
[13] Tiraby G., Sicard A.M., Integration efficiencies of spontaneous
mutant alleles of amiA locus in pneumococcal transformation,
J. Bacteriol. 116 (1973) 1130–1135.
[14] Green D., Ravin A., A host-specific variation affecting relative
frequency of transformation of two markers in Pneumococcus, Exp.
Cell Res. 18 (1959) 466–480.
[15] Lefèvre J.C., Gasc A.M., Burger A.C., Mostachfi P., Sicard A.M.,
Hyperrecombination at a specific DNA sequence in pneumococcal
transformation, Proc. Natl. Acad. Sci. USA 81 (1984) 5184–5188.
[16] Hotchkiss R.D., Transfer of penicillin resistance in pneumococci by
the deoxyribonucleate derived from resistant cultures, Cold Spring
Harbor Symp. Quant. Biol. 16 (1951) 457–461.
[17] Hakenbeck R., Tarpay M., Tomasz A., Multiple changes in penicillinbinding proteins in penicillin-resistant clinical isolates of Streptococcus pneumoniae, Antimicrob. Agents Chemother. 17 (1980)
[18] Guenzi E., Gasc A.M., Sicard A.M., Hakenbeck R., A two component signal-transducing system is involved in competence and
penicillin susceptibility in laboratory mutants of Streptococcus pneumoniae, Mol. Microbiol. 12 (1994) 505–515.
[19] Giammarinaro P., Sicard M., Gasc A.M., Genetic and physiological
studies of the CiaH-CiaR two-component signal-transducing system involved in cefotaxime resistance and competence of Streptococcus pneumoniae, Microbiology 145 (1999) 1859–1869.
[20] Pakula R., Wasczak K.W., On the nature of competence of transformable streptococci, J. Gen. Microbiol. 31 (1963) 125–133.
[21] Havarstein L.S., Coomaraswamy G., Morrison D.A., An unmodified
heptadecapeptide pheromone induces competence for genetic
transformation in Streptococcus pneumoniae, Proc. Natl. Acad. Sci.
USA 92 (1995) 11140–11144.
[22] Lefrançois J., Samrakandi M., Sicard A.M., Electrotransformation
and natural transformation of Streptococcus pneumoniae: requirement of DNA processing for recombination, Microbiology 144
(1998) 3061–3068.
[23] Sicard M., Selakovitch-Chenu L., Giammarinaro P., Geslin P.,
Gasc A.M., Genetic studies of cefotaxime resistance in Streptococcus pneumoniae, in: Horaud, et al. (Eds.), Streptococci and the Host,
Plenum Press, New York, 1997, pp. 429–433.
[24] Heidelberger M., Avery O.T., The specific soluble substance of
pneumococcus, Proc. Soc. Exp. Biol. Med. 20 (1923) 434.
[25] Ephrussi-Taylor H., Additive effects of certain transforming agents
from some variants of pneumococcus, J. Exp. Med. 89 (1949)
[26] MacLeod C.M., Krauss M.R., Relation of virulence of pneumococcal
strains for mice to the quantity of capsular polysaccharide formed in
vitro, J. Exp. Med. 92 (1950) 1–9.
[27] Berry A.M., Yother J., Briles D.E., Hansman D., Paton J.C., Reduced
virulence of a defined pneumolysin-negative mutant of Streptococcus
pneumoniae, Infect. Immun. 57 (1989) 2037–2042.
[28] Berry A.M., Paton J.C., Hansman D., Effect of insertional inactivation of the genes encoding pneumolysin and autolysin on the
virulence of Streptococcus pneumoniae type 3, Microb. Pathog. 12
(1992) 87–93.
[29] Azoulay-Dupuis E., Rieux V., Rivier C., Trombe M.C., Pleiotropic
mutations alter the kinetics of calcium transport, competence
regulation, autolysis and experimental virulence, Res. Microbiol.
149 (1998) 5–13.
[30] Braun J., Novack R., Gas G., Murray P., Shenep J., Pneumolysin, a
protein toxin of Streptococcus pneumoniae, induces nitric oxide
production from macrophages, Inf. Immun. 67 (1999) 3750–3756.
[31] Paton J.C., The contribution of pneumolysin to the pathogenicity of
Streptococcus pneumoniae, Trends Microb. 4 (1996) 103–106.
[32] Polissi A., Pontiggia A., Feger G., Altrieri M., Mottl H., Ferrari L.,
Simon D., Large-scale identification of virulence genes from Streptococcus pneumoniae, Inf. Immun. 66 (1998) 5620–5629.