Mechanism of Bactericidal and Fungicidal
Activities of Textiles Covalently Modified
With Alkylated Polyethylenimine
Jian Lin,1 Shuyi Qiu,1 Kim Lewis,2 Alexander M. Klibanov1
1
Department of Chemistry and Division of Biological Engineering,
Massachusetts Institute of Technology, Cambridge, Massachusetts 02139;
telephone: 617-253-3556; fax: 617-252-1609; e-mail: [email protected]
2
Department of Biology, Northeastern University, Boston, Massachusetts
Received 13 November 2002; accepted 16 December 2002
DOI: 10.1002/bit.10651
Abstract: Our previous studies have led to a novel “nonrelease” approach to making materials bactericidal by
covalently attaching certain moderately hydrophobic
polycations to their surfaces. In the present work, this
strategy is extended beyond the heretofore-used nonporous materials to include common woven textiles (cotton, wool, nylon, and polyester). Pieces of such cloths
derivatized with N-hexylated+methylated highmolecular-weight polyethylenimine (PEI) are strongly
bactericidal against several airborne Gram-positive and
Gram-negative bacteria. In contrast, the immobilized and
N-alkylated PEIs of low molecular weight have only a
weak, if any, bactericidal activity. These findings support
a mechanism of the antibacterial action whereby highmolecular-weight and hydrophobic polycationic chains
penetrate bacterial cell membranes/walls and fatally
damage them. The bactericidal textiles prepared herein
are lethal not only to pathogenic bacteria but to fungi as
well. © 2003 Wiley Periodicals, Inc. Biotechnol Bioeng 83: 168–
172, 2003.
Keywords: polycations; immobilization; microbicidal materials; fibers; cotton; molecular weight of polymers
INTRODUCTION
There is an ever-growing need for prevention of bacterial
infections in the hospital environment, in common daily
situations, and in (tragically real) acts of biological terrorism and warfare. In our recent studies, a new strategy has
been designed and verified for rendering common materials,
such as plastics and glass, permanently bactericidal (Lin et
al., 2002a, 2002b; Tiller et al., 2001, 2002). This strategy,
involving covalent attachment of certain hydrophobic polycations to material surfaces, has proved very effective
Correspondence to: Prof. A. M. Klibanov
Contract grant sponsor: U.S. Army Research Office (Institute for Soldier
Nanotechnologies at MIT)
The content of the information reported in this work does not necessarily
reflect the position or policy of the U.S. Government, and no official
endorsement should be inferred.
© 2003 Wiley Periodicals, Inc.
against a variety of Gram-positive and Gram-negative bacteria, whether airborne or waterborne (Lin et al., 2002a,
2002b; Tiller et al., 2001, 2002). Thus far, however, this
approach has been validated for only nonporous materials
(Borman, 2002) and only against bacterial pathogens.
The defining mechanistic feature of the aforementioned
strategy is that it is not based on a gradual release of an
entrapped antiseptic agent (Tiller et al., 2001). Similar nonrelease claims have been made for antibacterial materials
developed by other investigators (Borman, 2002; Isquith et
al., 1972; www.microbeshield.com), whereby fatty alkyl
groups were attached to surfaces via ammonium anchors.
However, in contrast to our immobilized polymers (Lin et
al., 2002a, 2002b; Tiller et al., 2001, 2002), it is difficult to
imagine how these relatively short chains can fatally penetrate bacterial membranes and/or cell walls.
In the present investigation, we extend our bactericidal
technology to textiles made of cotton, wool, nylon, or polyester. Moreover, such porous, woven materials bearing covalently attached alkylated polyethylemine (PEI) have been
found to be not only highly bactericidal but also equally
deadly to airborne fungi. Finally, mechanistic experiments
carried out with PEIs of widely different sizes have revealed
that surface-immobilized chains indeed must be polymeric
(with average molecular weights at least in the multiple
thousands) to exert their bactericidal effect.
MATERIALS AND METHODS
Materials
Cotton, wool, nylon (Nylon-66), and polyester [poly(ethylene terephthalate), or PET] textiles were purchased from a
fabric store in the Chinatown section of Boston, Massachusetts. Amino-glass slides (aminopropyltrimethoxysilanederivatized microscope slides) and fluorescein (Na salt)
were purchased from Sigma Chemical Co. PEI (Mw of 750,
25, 2, and 0.8 kDa), bromohexane, iodomethane, cetyltrimethylammonium chloride, tert-amyl alcohol, and all other
chemicals and solvents were obtained from Aldrich Chemical Co. and used without further purification.
Immobilization of PEI Onto Textile and Glass
Surfaces Followed by N-Alkylation
A 5×5-cm cotton or wool cloth was thoroughly washed with
distilled water, dried, and placed in a solution of 5 mL of
4-bromobutyrylchloride in 95 mL of dry chloroform, followed by stirring at room temperature for 5 h (Arai et al.,
2001). The acylated cloth was rinsed with chloroform to
remove unreacted 4-bromobutyrylchloride and immersed in
a mixture containing 10 g of PEI and 0.1 g of KOH in 90
mL of tert-amyl alcohol. After stirring at 75°C for 9 h, the
PEI-derivatized cloth was rinsed with methanol, air dried,
and placed in a solution of 0.1 g of KOH and 10 mL of
bromohexane in 90 mL of tert-amyl alcohol (Noding and
Heitz, 1998). After stirring at 75°C overnight, the cloth was
rinsed with methanol and further N-methylated in a solution
of 10 mL of iodomethane in 90 mL of tert-amyl alcohol at
40°C for 5 h in a sealed bottle (Lin et al., 2002b). The
resultant N-alkylated cloth was washed with methanol and
distilled water. In addition, when so stated hereafter, it was
further exhaustively washed with a 1% liquid soap (Softsoap, Colgate-Palmolive Co.) solution in water at 55°C
overnight.
A 5×5-cm nylon cloth was partially hydrolyzed in a 3 M
aqueous solution of HCl at 45°C for 4 h (Berezin et al.,
1974), and a polyester cloth of the same size was partially
hydrolyzed in a 1 M aqueous solution of NaOH at 45°C for
5 h (Chen and McCarthy, 1998). These treatments had no
noticeable effect on the appearance and physical strength of
the textiles. They were subsequently covalently derivatized
using the same procedure as described earlier for cotton and
wool.
An amino-glass slide was ultrasonicated in isopropyl alcohol for 5 min, dried at 80°C, and placed in a solution of
5 mL of 4-bromobutyryl chloride in 95 mL of dry chloroform, followed by stirring at room temperature for 5 h. The
N-acylated amino-glass slide was rinsed with methanol and
immersed in a mixture containing 20 g of PEI and 0.5 g of
KOH in 80 mL of tert-amyl alcohol. After stirring at 90°C
for 9 h, the PEI-modified slide was removed, rinsed with
methanol, air dried, and placed in a solution of 0.5 g of
KOH and 10 mL of bromohexane in 90 mL of tert-amyl
alcohol. After stirring at 90°C overnight, the slide was
rinsed with methanol and further N-methylated in a solution
of 20 mL of iodomethane in 80 mL of tert-amyl alcohol at
60°C for 9 h in a sealed bottle.
Titration of Primary and Quaternary
Amino Groups
Primary amino groups (Lee and Loudon, 1979). A 2×2-cm
wool cloth was immersed in 2 mL of a 0.1 M picric acid
solution in methylene chloride, followed by washing with
20 mL of the solvent, drying, and addition of 2 mL of dry
N,N-diisopropylethylamine. After a 5-min shaking, the
cloth was washed twice with 5 mL of methanol. The combined 2 mL of N,N-diisopropylethylamine and 10 mL of
methanol were diluted to 40 mL with ethanol. The concentration of the immobilized NH2 groups, 0.44 ± 0.01 nmol/
cm2, was calculated from the literature extinction coefficient of the amine–picrate complex of 14.5 mM−1 cm−1 at
358 nm.
The partially hydrolyzed nylon was also analyzed using
the foregoing procedure, yielding 0.70 ± 0.01 nmol of NH2
groups per square centimeter of cloth. The carboxylic
groups of the partially hydrolyzed polyester cloth were converted into amino groups via activation with 1,3dicyclohexylcarbodiimide, followed by coupling with 1,2diamonoethane. The density of the resultant NH2 groups
was found to be 0.92 ± 0.02 nmol/cm2.
Quaternary amino groups (Ledbetter and Bowen, 1969).
A 2×2-cm alkyl-PEI-derivatized cloth was dipped in a 1%
aqueous solution of fluorescein (Na salt), shaken for 5 min,
thoroughly rinsed with distilled water, and placed in 20 mL
of a 0.25% aqueous solution of cetyltrimethylammonium
chloride. After removing the fabric and adding 10% (v/v) of
a 0.1 M aqueous phosphate buffer (pH 8.0) to the solution,
the absorbance of the resultant aqueous solution was measured at 501 nm. The independently measured extinction
coefficient of fluorescein in this solution was 77 mM−1
cm−1. The densities of the quaternary amino groups on the
N-alkylated PEI-derivatized cotton, wool, nylon, and polyester cloths were found to be 17.3 ± 0.4, 3.9 ± 0.8, 15.9 ±
0.8, and 9.1 ± 0.6 nmol/cm2, respectively.
Determination of Microbicidal Activity
All bacteria (Tiller et al., 2001) were grown in yeast–
dextrose broth at 37°C, with aeration and agitation at 200
rpm for 6 to 8 h. The airborne (aerosolized) bacteria were
generated as described by Tiller et al. (2001, 2002). The
bacterial cells were harvested by centrifugation at 5,160xg
for 10 min and washed twice with distilled water. A 105cell/mL bacterial suspension in distilled water was finely
sprayed onto a cloth in a fume hood using a 10-mL thinlayer chromatography sprayer (VWR). After drying under
air, the fabric was placed in a Petri dish and immediately
covered with a layer of solid growth agar (1.5% agar in the
yeast–dextrose broth), autoclaved, poured into a Petri dish,
and dried under reduced pressure at room temperature overnight. The cloth-containing Petri dish was sealed and incubated at 37°C for 2 days. The bacterial colonies grown as a
result were counted on a light box.
Saccharomyces cerevisiae was standard laboratory strain
BY4742 (ATCC 4040004) (MAT␣ his3-⌬1 leu2-⌬0 lys2⌬0 ura3-⌬0). Candida albicans was purchased from the
American Type Culture Collection (ATCC No. 90028). The
fungi were grown in yeast–peptone–dextrose broth at 30°C
with agitation at 50 rpm overnight. The airborne fungi were
LIN ET AL.: RENDERING TEXTILES PERMANENTLY MICROBICIDAL
169
generated, and experimented with, in the same way as the
airborne bacteria.
RESULTS AND DISCUSSION
With the goal of making clothing and uniforms permanently
microbicidal, in the present work we have extended our
previous studies (Lin et al., 2002a, 2002b; Tiller et al., 2001,
2002) to a new class of materials, namely woven textiles.
Specifically, we subjected pieces of cloth made of four common fabrics (both natural and synthetic: cotton, wool, nylon, and polyester) to the covalent derivatization with Nperalkylated polyethylenimine (PEI), as shown in Figure 1.
In the case of cotton and wool, the procedure was analogous to that previously employed by us for nonporous materials (Lin et al., 2002a). Cotton’s OH and wool’s NH2
groups were acylated with 4-bromobutyryl chloride, and
then commercially available PEIs of different molecular
weights were N-alkylated with the resultant brominated surfaces. The textile-immobilized PEIs thus obtained were subsequently N-hexylated with bromohexane, followed by Nmethylation of the remaining nonquaternized amino groups
(presumably inaccessible to the hexyl moieties) with iodomethane. In the case of nylon and polyester, this procedure was preceded by partial acid and alkaline hydrolyses,
respectively, to form free NH2 and OH groups (Fig. 1). In
all instances, the derivatized textiles were titrated (Ledbetter
and Bowen, 1969) to determine the density of the quaternary amino groups formed (see Methods).
The bactericidal activity of cotton, wool, nylon, and polyester derivatized with N-alkylated 750-kDa PEI was tested
against airborne Staphylococcus aureus and epidermidis,
Escherichia coli, and Pseudomonas aeruginosa. An aqueous bacterial suspension (a few tens of microliters) was
sprayed onto a piece of cloth, and the latter was air dried.
An overlapping slab of a nutrient-containing agar gel was
then placed on the cloth, followed by incubation at 37°C.
The bacterial colonies grown were subsequently counted,
and their number was compared with that formed on the
parent (i.e., nonderivatized) cloth under identical conditions.
As can be seen in the first rows of Table I, all derivatized
fabrics were lethal to the bacteria tested; that is, well over
90% of the deposited bacteria were invariably killed, with at
least 98% killed in most instances. (Following microbiological convention, microorganisms throughout this work
were considered dead if they failed to multiply on derivatized materials over prolonged periods of time under favorable conditions; i.e., those promoting vigorous growth and
division of the same bacteria on the corresponding untreated
materials.) These bactericidal activities are similar to those
previously observed by us for other, nonporous materials
covalently modified using optimal N-alkylated poly(vinyl
pyridines) or PEIs (Lin et al., 2002a, 2002b; Tiller et al.,
2001, 2002).
Next, we explored whether the textiles with immobilized
750-kDa N-alkyl-PEI would be effective against airborne
fungi. (Fungi/yeasts have a single membrane covered by a
thick cell wall composed of glucan and chitin. This design
is similar to the cell envelope of Gram-positive bacteria,
composed of a membrane covered with a peptidoglycan
wall. In Gram-negative bacteria, an additional outer membrane is found above the peptidoglycan wall [Lewis, 2001].)
To this end, experiments analogous to the foregoing were
carried out with a model fungus, Saccharomyces cerevisiae,
and a major pathogenic fungus, Candida albicans. One can
see in the last two rows of Table I that the fungicidal activities of the derivatized textiles are nearly as high as their
bactericidal activities. It is worth mentioning that the
slightly lower (compared with cotton’s and wool’s) fungicidal efficiencies of the derivatized nylon and polyester can
possibly be boosted further by increasing the density of the
hydrolytically generated free NH2 and OH groups (steps 1
and 2, respectively, in Fig. 1).
Figure 1. Schematic representation of the synthesis of N-alkylated PEIs covalently immobilized onto cotton, wool, nylon, or polyester cloths. Step 1:
alkaline hydrolysis of polyester [poly(ethylene terephthalate)]; step 2: acid hydrolysis of nylon (Nylon-66); step 3: acylation of original or partially
hydrolyzed textiles’ hydroxyl or amino groups with 4-bromobutyryl chloride; step 4: covalent attachment (N-alkylation) of PEI; step 5: N-hexylation of
the immobilized PEI with bromohexane; and step 6: N-methylation of the immobilized N-hexyl-PEI with iodomethane (see text).
170
BIOTECHNOLOGY AND BIOENGINEERING, VOL. 83, NO. 2, JULY 20, 2003
Table I. Microbicidal efficiencies of four textiles derivatized with Nalkylated 750-kDa PEI against airborne bacteria and fungi.a
Microbicidal activities (%)b
Microorganism
Type
Cotton
Wool
Nylon
Polyester
S. aureus
S. epidermidis
E. coli
P. aeruginosa
S. cerevisiae
C. albicans
Gram(+)
Gram(+)
Gram(−)
Gram(−)
Fungus/yeast
Fungus/yeast
98 ± 1
97 ± 1
99 ± 1
98 ± 1
97 ± 1
96 ± 1
94 ± 2
98 ± 1
98 ± 1
97 ± 1
98 ± 1
96 ± 3
96 ± 2
98 ± 1
99 ± 1
98 ± 1
90 ± 2
91 ± 2
93 ± 1
97 ± 1
96 ± 2
98 ± 1
88 ± 9
92 ± 7
a
Immobilization and N-alkylation of PEI were carried out as outlined in
Figure 1. For experimental details, see text.
b
All measurements were done in duplicate with the mean values and
standard errors obtained shown.
To shed light on the mechanism of bactericidal action of
the immobilized N-alkylated PEI, we varied the average
molecular weight of the polymer. Specifically, in addition to
the 750-kDa PEI employed thus far (Table I), we also immobilized onto cotton cloth (Fig. 1) PEIs with average molecular weights of 25, 2, and 0.8 kDa. The bactericidal potencies of the resultant derivatized cotton cloths against airborne S. aureus are presented in the first data column of
Table II. To our surprise, even very short (2- and 0.8-kDa)
PEIs were found to possess significant killing efficiencies
— some half of that with the 750-kDa polymer.
In our standard immobilization/N-peralkylation procedure (Fig. 1), the derivatized cotton was thoroughly washed
with methanol and distilled water to remove physically adsorbed (as opposed to covalently attached) polymers. We
hypothesized that this washing, although quite sufficient
with, say, monolithic nonporous plastics (Lin et al., 2002a),
may not wash off all the adsorbates because of the interwoven, multithread character of cotton fibers making up the
cloth and/or strong affinity of N-alkylated PEIs to cellulose.
To test this hypothesis, all four derivatized cotton cloths
were subjected to an additional exhaustive washing that
included stirring overnight in soapy, warm water, followed
by thorough rinsing with distilled water.
The bactericidal activities of derivatized cotton cloths af-
ter this exhaustive “laundry cycle,” as well as after a subsequent one, are shown in the second and third data columns
of Table II and reveal some important conclusions. One can
see that neither the first nor the second cycle had any appreciable effect on the bactericidal potency of the cottonimmobilized and N-alkylated 750-kDa PEI (first row in
Table II). (Likewise, this exhaustive washing of the Nalkylated high-molecular-weight PEI immobilized onto
wool, nylon, and polyester did not noticeably diminish the
bactericidal potency.) In contrast, this washing substantially
cut the bactericidal efficiencies of the immobilized and Nalkylated PEIs of smaller sizes. Moreover, the lower the
latter, the greater the fraction of bactericidal efficiency that
was lost: 16%, 30%, and 67% after the second laundry cycle
for the 25-, 2-, and 0.8-kDa PEIs, respectively (see Table
II). On the basis of these observations we hypothesized that
the significant bactericidal activity obtained with immobilized and N-alkylated 2- and 0.8-kDa PEIs was largely, and
perhaps exclusively (if even the two exhaustive laundry
cycles were not sufficient), due to the polycations that
bound noncovalently to cotton.
Table II. Bactericidal potencies against airborne S. aureus of cottonimmobilized and N-alkylated PEIs of different molecular weights as a
function of washing.a
Bactericidal activities (%) after:
Molecular weight
of PEI (kDa)
Regular
washb
First
exhaustive washc
Second
exhaustive washc
750
25
2
0.8
98 ± 1
83 ± 3
53 ± 17
58 ± 10
98 ± 1
75 ± 11
35 ± 15
24 ± 4
95 ± 3
70 ± 6
37 ± 8
19 ± 13
a
See footnotes to Table I.
Consisted of washing with methanol and distilled water.
c
Consisted of additional stirring in soapy water at 55°C overnight, followed by thorough rinsing with distilled water.
b
Figure 2. Bactericidal activities against airborne S. aureus of the Nalkylated PEIs of different molecular weights, as well as of 1,2diaminoethane, covalently attached (as in Fig. 1) to NH2-glass slides (see
text).
LIN ET AL.: RENDERING TEXTILES PERMANENTLY MICROBICIDAL
171
This conclusion was further confirmed with the Nalkylated PEIs of different molecular weights covalently
attached to NH2-glass slides (Lin et al., 2002a). Because the
glass slides are monolithic and nonporous with a modest
and readily accessible surface, washing off adsorbed polymers from them, unlike from cotton cloth, should be easy
and unequivocal, thus eliminating putative artifactual bactericidal activities. Figure 2 shows the dependence of bactericidal activity on the molecular weight of PEIs immobilized, followed by the N-alkylation (Fig. 1) and washing,
onto glass slides. One can see that immobilized 750- and
25-kDa PEIs were nearly completely lethal to airborne S.
aureus. In contrast, their 2- and 0.8-kDa counterparts had
negligible, if any, antibacterial activities, as did the immobilized and N-peralkylated PEI monomer analog 1,2diaminoethane.
The results presented in Table II and Figure 2 demonstrate that the PEI used for immobilization must be truly
polymeric to be bactericidal; that is, it must have very long
chains (see later). This conclusion supports the previously
proposed mechanism of bactericidal action (Tiller et al.,
2001) whereby hydrophobic polycationic chains traverse
and irreparably damage the bacterial cellular membrane/
wall.
Because the commercial PEIs used herein are branched,
the exact polymer chain length is impossible to calculate.
Nevertheless, we estimated the lengths of hypothetical linear, fully extended PEIs listed in Table II and related these
to the dimensions of a typical bacterial cell (some 1 to 2
␮m) and of its relevant boundary components. The 750-kDa
PEI was thus estimated to be approximately 6 ␮m in length;
that is, more than enough to penetrate even the whole bacterial cell. The 25-kDa PEI is up to some 0.2 ␮m in length
(although many chains are longer still because the molecular weight given is average), which is conceivably enough to
damage cellular membranes/walls. On the other hand, the 2and 0.8-kDa PEIs, merely some 0.02 and 0.007 ␮m in
length, respectively, are presumably too short to inflict appreciable harm.
It should be emphasized that fatty alkyl chains attached to
surfaces through quaternary amine anchors in the reported
172
alternative antibacterial materials (Borman, 2002; Isquith et
al., 1972; www.microbeshield.com) are far shorter than
even the 0.8-kDa PEI. Therefore, such materials cannot
function via the same “hole-poking” mechanism as described in this study and elsewhere (Lin et al., 2002a,
2002b; Morgan et al., 2000; Tiller et al., 2001, 2002).
References
Arai T, Freddi G, Innocenti R, Kaplan DL, Tsukada M. 2001. Acylation of
silk and wool with acid anhydrides and preparation of water-repellent
fibers. J Appl Polym Sci 82:2832–2841.
Berezin IV, Klibanov AM, Martinek K. 1974. The mechanochemistry of
immobilized enzymes. How to steer a chemical process at the molecular lever by a mechanical device. Biochim Biophys Acta 364:193–199.
Borman S. 2002. Surfaces designed to kill bacteria. Chem Eng News
80:36–38.
Chen W, McCarthy TJ. 1998. Chemical surface modification of poly(ethylene terephthalate). Macromolecules 31:3648–3655.
Isquith AJ, Abbott EA, Walters PA. 1972. Surface-bonded antimicrobial
activity of an organosilicon quaternary ammonium chloride. Appl Microbiol 24:859–863.
Lee CCY, Loudon GM. 1979. Quantitative determination of amino groups
on derivatized controlled pore glass: A comparison of methods. Anal
Biochem 94:60–64.
Ledbetter JW, Bowen JR. 1969. Spectrophotometric determination of the
critical micelle concentration of some alkyldimethylbenzylammonium
chlorides using fluorescein. Anal Chem 41:1345–1347.
Lewis K. 2001. In search of natural substrates and inhibitors of MDR
pumps. Mol Microbiol Biotechnol 3:247–254.
Lin J, Qiu S, Lewis K, Klibanov AM. 2002a. Bactericidal properties of flat
surfaces and nanoparticles derivatized with alkylated polyethylenimines. Biotechnol Progr 18:1082–1086.
Lin J, Tiller JC, Lee SB, Lewis K, Klibanov AM. 2002b. Insights into
bactericidal action of surface-attached poly(vinyl-N-hexylpyridinium)
chains. Biotechnol Lett 24:801–805.
Morgan HC, Meier JF, Merker RL. 2000. Method of creating a biostatic
agent using interpenetrating network polymers. U.S. Patent No.
6,146,688.
Noding G, Heitz W. 1998. Amphiphilic poly(ethylenimine) based on longchain alkyl bromides. Macromol Chem Phys 199:1637–1644.
Tiller JC, Lee SB, Lewis K, Klibanov AM. 2002. Polymer surfaces derivatized with poly(vinyl-N-hexylpyridinium) kill both air- and waterborne bacteria. Biotechnol Bioeng 79:465–471.
Tiller JC, Liao C-J, Lewis K, Klibanov AM. 2001. Designing surfaces that
kill bacteria on contact. Proc Natl Acad Sci USA 98:5981–5985.
BIOTECHNOLOGY AND BIOENGINEERING, VOL. 83, NO. 2, JULY 20, 2003