Functional similarities between the small heat shock proteins
Mycobacterium tuberculosis
HSP 16.3 and human aB-crystallin
Melissa M. Valdez
1
, John I. Clark
1,2
, Gabrielle J. S. Wu
3
and Paul J. Muchowski
4
1
Departments of Biological Structure, and
2
Ophthalmology, University of Washington, Seattle, WA, USA;
3
Seattle Genetics, Bothell,
WA, USA;
4
Department of Pharmacology, University of Washington, Seattle, WA, USA
Mycobacterium tuberculosis heat shock protein 16.3 (MTB
HSP 16.3) accumulates as the dominant protein in the latent
stationary phase of tuberculosis infection. MTB HSP 16.3
displays several characteristics of small heat shock proteins
(sHsps): its expression is increased in response to stress, it
protects against protein a ggregation in v itro, and it contains
the core Ôa-crystallinÕ domain found in all sHsps. In this study
we characterized the c haperone activity of r ecombinant
MTB HSP 16.3 in several different assays and compared the
results to those obtained with recombinant human
aB-crystallin, a well characterized member of the sHsp

MTB during p rolonged periods of infection [5–7]. It w as
shown that MTB HSP 16.3, initially described as the
immunodominant 14- or 16-kDa antigen [8–11], was a
major component in tuberculosis infection in humans and
played an important role in enhancing protein stability and
survival [5]. Eighty-five percent of patients with active
tuberculosis showed a positive reaction to this a ntigen,
suggesting that this protein expressed in vivo had a key role
in MTB infection [11,12]. The 14K antigen was later
renamed MTB HSP 16.3 [13]. MTB HSP 1 6.3 accumulates
to become t he dominant protein in the la tent stationary
phase of M. tube rculosis infection [7]. Over-expression of
HSP 16.3 in log phase growth of M. tuberculosis slowed the
growth rate and protected ag ainst stationary phase autolysis
in v itro [7]. MTB HSP 16.3 h as been charac terized a s a
membrane associated protein [12] having sequence homo-
logy to other proteins i n the small heat s hock protein (sHsp)
family [11,14]. All sHsps share se quence similarity in a
conserved 80–100 amino-acid Ôa-crystallinÕ domain region
found in the C-terminus which is thought to be important
for c haperone function s [14–16]. MTB HSP 16.3 h as been
shown to contain an oligomeric, active structure which may
form a trimer of trimers and pos sesses in vitro molecular
chaperone activity [13].
Up-regulation o f l arge and small sHsps is t hought to be a
universal response to s tress. In vitro, human aB-crystallin
and other sHsps function as molecular chaperones by
suppressing unfolding and aggregation o f polypeptides in
response to s tress [17,18]. MTB HSP 16.3 modulates its
chaperone activity by exposing hydrophobic s urfaces a nd

manner [24] and indicating the involvement of ATP in
substrate release [25]. In this s tudy, t he effect of ATP on
recombinant MTB HSP 16.3 was compared with the effect
of ATP on recombinant human aB-crystallin.
We report t he expression and purification of recombinant
MTB HSP 16.3 from E. coli and studies of its f unction as a
molecular chaperone. MTB HSP 16.3 was compared with
aB-crystallin in vivo and in vitro biochemically and in
functional assays. Although only 18% sequence identity is
shared between the two s Hsps, MTB HSP 16.3 functioned as
effectively as aB-crystallin as a molecular chaperone in v itro.
The molecular chaperone activity of recombinant MTB
HSP 1 6.3 was enh anced in the presence of ATP which is
consistent with previous findings of the effect of ATP on
recombinant human aB-crystallin [23]. The expression of
MTB HSP 16.3 in E. coli exposed to high temperatures
resulted in a very impressive level of survival. Our results
suggest the chaperone activity ofMTB H SP 16.3may play an
important role in the survival and stability of M. tuberculo sis.
MATERIALS AND METHODS
Expression and purification of MTB HSP 16.3
HSP 1 6.3 was subcloned into the pET-20b(+) expression
vector which w as provided by H. McHaourab (Department
of Molecular Physiology a nd Biophysics, Vanderbilt
University School of Medicine, Nashville, TN, USA). T he
pET-20b(+)-HSP 16.3 expression plasmid was used to
transform E. coli BL21 (DE3) competent cells (Novagen,
Inc., Milwaukee, WI, USA). The expression of HSP 16.3
was based on a method described previously [26]. One litre
of Luria–Bertani broth containing 10 g NaCl, 5 g yeast

stirring for 30 min. The sample was then placed in a 50-mL
tube and centrifuged at 18 000 g for 1 h. The supernatant
from this sample was transferred to a new beaker with
constant stirring at room temperature with the addition of
400 lL 5% polyethylenimine a nd 800 lL 200 m
M
dithio-
threitol for 10 min. The sample was then c e ntrifuged at
35 000 g for 2 h at 4 °C. The supernatant was decanted and
the insoluble pellet w as discarded. The s upernatant was then
ready for purification using the Pharmacia FPLC system.
The supernatant containing the soluble protein was
filtered through a 0 .22 lm filter and was loaded onto a High
Trap Q Anion Exchange Column (Pharmacia), pre-equil-
ibrated with Buffer A (20 m
M
Bis/Tris, pH 6.5). The
proteins were eluted using a linear gradient of
0–1.0
M
NaCl. T he protein fractions were analyzed using
SDS/PAGE (Invitrogen). Proteins were analyzed on a
4–12% polyacrylamide e lectrophoretic gel in the presence of
0.1% SDS a nd Mes buffer and were stained with C oomasie
blue R-350 (Amersham Pharmacia). Fractions containing
the 16.3-kDa p rotein were th en pooled and concentrated
using a 1 0 000 molecular mass cut-off concentrator
(Amicon). Concentrated protein (5 mL) was loaded onto
a Phenyl Superose Hydrophobic Interaction Column,
preequilibrated with a buffer containing 50 m

HSP 1 6.3 and ligated into pET16b (Pharmacia) digested
with the same restriction enzymes. Restriction mapping
analysis was performed on pET16b-HSP 16.3 to ensure
that the HSP 16.3 gene was i nserted in the proper
orientation, and the coding sequence of HSP 16.3 was
verified by DNA sequence analysis. The pET16b-HSP 16.3
vector was then t ransformed into BL21 (DE3) competent
cells.
For t he thermal killing experiment, an equal number of
cells were gr own containing either pET16b-HSP 16.3
vector, pET16b (empty vector control), or pET16b-aB
(positive control). Equal numbers of cells from overnight
cultures were inoculated into 50 mL of
L
-broth medium
containing 100 lgÆlL
)1
carbenicillin and grown at 37 °C
until they reached an D
600
¼ 0.8. Protein expression was
then induced with 1.0 m
M
IPTG. After a 2 h i nduction,
samples were s hifted to a s haking water bath a t 4 8 °C.
Samples w ere r emoved at 3-h t ime points postinduction
and scored f or cell viability by p lating on Luria–Bertani
broth plates containing carbenicillin. Cell viability was
determined by counting the number of colony forming
units (CFUs) on each plate after heat shock at 48 °C

was equilibrated w ith 3.5 m
M
ATP (or A TPcS), 3 .5 m
M
MgCl
2
and 1 0 m
M
KCl b efore addition of HSP 16.3 o r
CS. The protection from thermal aggregation of CS at
45 °C w ith HSP 16.3 was a lso c ompared with t he same
molar ratios of aB-crystallin (predicted from monomeric
molecular weights).
Chymotrypsin digestion of HSP 16.3
Chymotrypsin digestion with HSP 16.3 was base d on the
methods used for GroEL and aB-crystallin [27,28]. In
summary, f or each reaction 70 lgMTBHSP16.3were
diluted into a final volume of 100 lL buffer c ontaining
100 m
M
Tris/HCl, pH 7.4, 3.5 m
M
MgCl
2
,10m
M
KCl,
and 0.01% Tween-20. For reactions with ATP or ATPcS,
a final concentration of 3.5 m
M

(a-16 kDa)] raised against native MTB HSP 1 6.3
(Fig. 1 B). IT-4 (a-16 kDa) recognized the recombinant
MTB HSP 16.3 from E. coli and did not react with
recombinant aB-crystallin (Fig. 1 B). The purification
yield obtained from 1 L of Luria–Bertani broth culture
was between 10 and 3 0 mg o f MTB HSP 16.3.
Fig. 1. Expression and Purification of MTB HSP 16.3. (A) The
expression an d p urification of MTB HSP 16.3 was a nalyz ed b y S DS/
PAGE using 4–12 % Bis/Tris polyacrylamide gels in the presence of
Mes buffer. Lanes 1 and 7, m olec ular mass markers; lane 2, expression
protein in E. coli cells not induced by IPTG; lane 3, protein expression
in E. coli after induction with IPTG; lane 4, following purification on
the High trap Q anion exchange c olumn; lane 5, MTB HSP 16.3
enriched after purification using a Phenyl Superose H ydro phobic
Interaction Column; lane 6, following Tris/HCl buffer exchange of
MTB HSP 16.3 on a Superdex 200 Size Exclusion column (to remove
salt). A protein assay performed after desalting the sample showed that
the y ields varied between 10 and 30 mgÆL
)1
of MTB HSP 16.3 cell
culture. (B) SDS/PAGE (left) and Western immunoblot (right) on a
4–12% polyacrylamide gel of recombinant MTB HSP 16.3 and
recombinant h uman aB-crystallin. Lane 1, molecular mass markers;
lane 2, MTB HSP 16.3; lane 3, aB-crystallin. In the Western immu-
noblot, th e I T-4 antibody to MTB HSP 16.3 d etected r ecom binant
MTB HSP 16.3 in lane 2 only. No reactivity with anti-(MTB
HSP 16.3) I g was obser ved in lan e 3, which contained human
aB-crystallin (right s ide).
1808 M. M. Valdez et al. (Eur. J. Biochem. 269) Ó FEBS 2002
Amino-acid sequence determination of recombinant

decreased by two orders of magnitude after the 6-h heat
shock. The protective effect of the induced aB-crystallin was
stronger than MTB HSP 16.3. SDS/PAGE analysis s howed
strong expression of MTB HSP 16.3 and aB-crystallin at
time points 0 , 3, and 6 h following heat shock at 48 °C
(Fig. 2 B,C).
In vitro
chaperone activity of MTB HSP 16.3
and aB-crystallin
In Fig. 3, we observed the effects of different concentrations
of MTB HSP 16.3 in the presence and absence of ATP on
the aggregation of CS. In the absence of added MTB
HSP 1 6.3, aggregation o f CS increased af ter a short delay to
reach a maximum after approximately 25 min at 4 5 °C
Fig. 2. Cell viability of MTB HSP 16.3. The pET 16b-MTB HSP 16.3
vector, the control pET 16b-aB vector and the pET 16b vector con-
taining no inserted gene were e xpressed at 37 °C and induced with
IPTG when the cell c ultures reached D
600
¼ 0.8. After induction fo r
2 h and heat shock to 48 °C, the cells were incubated for a further 6 h.
Samples were taken at concurrent and sequential time points beginning
at the time of heat shock, plated and CFU counted. The proportions of
viable cells expressing the pET 16b-MTB HSP 16.3 vector and the two
control vectors were plotted for 0 , 3 and 6 h following h eat s hock. At
48 °C, the proportion of surviving cells expressing pET 16b vector only
or aB-crystallin uninduced c ells was negligible and viability of cell
cultures decreased more than fourfold (A). By 6 h post heat shock, the
cultures that over-expressed MTB HSP 16.3 and the aB-crystallin in-
duced c ells remained viable. Protein expression was analyzed b y SDS/

next compared to aB-crystallin at identical molar ratios
(Fig. 4). In ge neral, aB-crystallin was more effective as a
molecular chaperone than MTB HSP 16.3 under the
conditions of these experiments (Fig. 4). Complete s uppres-
sion of CS aggregation by MTB HSP16.3 required a molar
ratio of 15 : 1, while aB-crystallin required a molar ratio of
5 : 1 for complete suppression of aggregation.
Chymotrypsin proteolysis of MTB HSP 16.3
in the absence and presence of ATP
MTB H SP 16.3 was digested with chymotrypsin in the
absence and presence of ATP at 42 °C (Fig. 5A–C).
Proteolysis of MTB HSP 16.3 increased with chymotrypsin
concentration as expected (data not shown). Each individ-
ual lane is a sample of MTB HSP 16.3 plus chymotrypsin
Fig. 4. Comparison of molecular chaperone activity between recombin-
ant MTB HSP 16.3 and recombinan t human aB-crystallin. The
molecular chaperone activity of MTB HSP 16.3 o n CS aggregation
was compared to the effect of human aB-crystallin on C S aggregation.
The aggregation of CS was mea sured in the presence o f d ifferent
concentrations of MT B HSP 16.3 o r human aB-crystallin after a
30-min period. The bar graphs measure the a ggregation of C S in
arbitrary units vs. t he ratios of MTB HSP 16.3/CS and h uman
aB-crystallin/CS. With increased ratios of the molecular chaperone
protein to CS, there was increased protection aga inst CS aggregation.
Recombinant human aB-crystallin demonstrated better p rotection
against CS aggregation than MTB HSP 16.3. At the 15 : 1 molar ratio
of MTB HSP 16.3/CS, protection aga inst CS aggregation was almost
complete. Similar protection was observed at a ratio of 5 : 1 for human
aB-crystallin/CS.
Fig. 3. Molecular Chaperone Activity of MTB HSP 16.3. To test the

MTB HSP 16.3. [d,CSalone;,,HSP16.3:CS (10:1)+1m
M
KCl and 3.5 m
M
MgCl
2
; s, HSP 16.3 : CS (10 : 1); .,HSP 16.3 : CS
(10 : 1) + 3.5 m
M
ATPcS; j, HSP 16.3 : CS (10 : 1) + 3.5 m
M
ATP].
1810 M. M. Valdez et al. (Eur. J. Biochem. 269) Ó FEBS 2002
taken at 5-min time intervals over a 30-min period. Nearly
all intact MTB HSP 16.3 was degraded after 15 m in in the
absence of ATP (Fig. 5A). In the presence of 3.5 m
M
ATP,
MTB HSP 16.3 was stabilized against proteolysis, and
intact MTB HSP 16.3 could be detected even after 3 0 min
of proteolysis (Fig. 5B). The digestion pattern of MTB
HSP 1 6.3 was similar in the absence and presence of ATP,
where two major proteolytic fragments at M
r
 8and
13 kDa were observed. The specificity of the effect of ATP
on stabilization of MTB HSP 16.3 against p roteolysis was
confirmed with nonhydrolyzable ATP analog ATPcS,
which had no stabilizing e ffect against proteolysis (Fig. 5C).
DISCUSSION

protein, MTB HSP 16.3 was effective as a chaperone,
although less effective than aB-crystallin in suppressing CS
aggregation. It is possible that additional cofactors found
only in M. tuberculosis cytosol could i ncrease the chaper one
activity of MTB HSP 16.3. It is also likely t hat the efficiency
of MTB HSP 16.3 as a chaperone may b e improved using
target proteins that are native to M. tuberculosis.
The e ffects of ATP on the c haperone activity of MTB
HSP 1 6.3 were similar to aB-crystallin, a sHsp whose
Fig. 6. Sequence alignment of recombinant MTB HSP 16.3 and
recombinan t hum an aB-crystallin. Amino-acid sequence a lignment
between recombinant MTB H SP 16.3 and h uman aB-crystallin was
aligned using the
MULTALIN MULTIPLE SEQUENCE ALIGNMENT
program
(PBIL, Franc e) with the h elp of S. Yarfitz (Unive rsity o f Washington
Health Science s L ibrary, Seattle, WA, USA). Shading indicates
chemically identical and similar a mino-acids residues (
BOXSHADE
program from the European Molecular Biology Network). Residues
highlighted black indicate amino-acid residues that are chemically
identical a nd residu es h igh lighted gray indicate amino-acid residues
that are c hemically similar. Between MTB HSP 16.3 and human
aB-crystallin there was an 18 % sequence identity and a n overall 30%
shared sequence similarity between the tw o proteins. The conserved
core a-crystallin domain observed in proteins of the sHsp spans resi-
dues of E67–I161 in the aB- crystallin sequence.
Fig. 5. The chymotrypsin proteolysis of MTB HSP 16.3. SDS/PAGE used 4–12% Bis/Tris polyacrylamide gels in the presence of Mes buffer (A–C).
Arrows indicate the MTB HSP 16.3 band. Lane 1 of each gel contains the molecular mass markers. Each individual lane is a sample of MTB
HSP 16.3 plus 0.51 lg chymotrypsin taken at 5-min intervals over a 30-min period. MTB HSP 16.3 was readily degraded by chymotrypsin and

phosphate buffer systems [13,19]. The conditions used in
this study were the same as those used successfully to
demonstrate the ATP effect on human aB-crystallin [23].
In separate experiments the chymotrypsin proteolytic
digestion pattern of MTB H SP 16.3 in the presence and
absence of ATP was evaluated. Similar to aB-crystallin [28]
and Hsp27 [32], chymotrypsin cleavage sites in MTB
HSP 16.3 appeared to be shielded in the presence of ATP.
The s imilarity of the chymotrypsin d igestion pattern for
MTB HSP 16.3 to previous studies with aB-crystallin and
Hsp27 may indicate similar domain structures and assembly
properties that are stabilized in the presence of ATP. As
with aB-crystallin, ATPcS (a nonhydrolyzable A TP analog)
did not enhance the chaperone function of MTB HSP 16.3,
and did not protect against i ts proteolysis by c hymotrypsin.
Although there is on ly 18% sequen ce identity, the c ore
Ôa-crystallinÕ domain in M TB HSP16.3 may have functional
significance similar to that of aB-crystallin [28].
MTB HSP 16.3 of M. tuberculosis may be ideally suited
for studies of the structure and function of the core
Ôa-crystallinÕ domain of sHsps because the quaternary
structure is more monodisperse than aB-crystallin and
other sHsps, that are known to have highly variable
quaternary structures [ 41]. T he crystal s tructure of HSP
16.5 from Methanococcus janaschii demonstrates a mono-
mer containing a core domain that consists largely of
b sheets [42]. The molecules o f HSP 16.5 form dimers that
assemble into a spherical complex of octahedral symmetry,
while MTB HSP 16.3 is reported t o consist of a trimer of
trimers [13]. Spin labeling o f MTB HSP 16.3 in solution is

for technical assistance. This work was supported by National Eye
Institute Grant E Y0452 (to J. I. C.).
REFERENCES
1. Sudre, P., te n Dam, G. & K ochi, A. (1992) Tuberculosis: a global
overview of the situation t oday. Bull. World H ealth Organ. 70,
149–159.
2. Ginsberg, A.M. (2000) A proposed national strategy for
tuberculosis vaccine development. Clin. Infect. Dis. 30 (Suppl. 3),
S233–S242.
3. Bloom, B.R. & Murray, C.J. (1992) Tube rculosis: commentary on
a reemergent killer. Science 257, 1055–1064.
4. Wayne, L.G. (1994) Dormancy of M ycobacterium tuberculosis
and latency of disease. Eur. J. Clin. Microbiol. I nfect. Dis. 13,
908–914.
5. Yuan, Y., Crane, D.D. & Barry, C.E., III (1996) Stationary phase-
associated protein expression in Mycobacterium tuberculosis:
function of the mycobacterial alpha-crystallin homolog. J. Bac-
teriol. 178, 4484–4492.
6. Cunningham, A.F. & Spreadbury, C.L. ( 1998) Mycobacterial
stationary phase induced by low oxygen tension: cell wall thick-
ening and localization of the 16-kilodalton alpha-crystallin
homolog. J. Bacteriol. 180, 801 –808.
7. Yuan, Y., Crane, D.D., Simpson, R.M., Zhu, Y.Q., Hickey, M.J.,
Sherman, D .R. & Barry, C.E. (1998) The 16-kDa alpha-crystallin
(Acr) protein of Mycobacterium tuberculosis is required for growth
in macrophages. Proc.NatlAcad.Sci.USA95, 9578–9583.
8.Engers,H.D.,Houba,V.,Bennedsen,J.,Buchanan,T.M.,
Chaparas, S.D., Kadival, G., Closs, O., David, J.R., van Embden,
J.D.A.,Godal,T.et al. (1986) Results of a World Health
Organization-sponsored workshop to characterize antigens

16. de Jong, W.W., Caspers, G.J. & Leunissen, J .A. (1998) Genealogy
of the alpha-crystallin-small heat-shock protein superfamily. Int.
J. Biol. Macromol. 22, 151–162.
17. H orwitz, J. (1992) Alpha-crystallin can function as a molecular
chaperone. Proc. Natl Acad. Sci. USA 89, 10449–10453.
18. Jakob,U.,Gaestel,M.,Engel,K.&Buchner,J.(1993)Smallheat
shock proteins are molecular chaperones. J. Biol. Chem.
268, 1517–1520.
19. Y ang, H., H uang, S ., D ai, H., Gong, Y., Zheng, C. & Chang, Z.
(1999) The Mycobacterium tuberculosis small heat shock protein
HSP 16.3 exposes hydrophobic surfaces at mild conditions: con-
formational flexibility and molecular chaperone activity. Protein
Sci. 8, 1 74–179.
20. Dai, H ., Mao, Q., Yang, H., Huang, S. & Chang, Z. (2000)
Probing the roles of the only universally conserved l eucine residue
(Leu 122) in the oligomerization and chaperone-like activity of
Mycobacterium tuberculosis small heat s hock protein H SP 16.3.
J. Protein Chem. 19, 319–326.
21. H artl, F.U. ( 1996) Molecular c haperones i n cellular p rotein fold-
ing. Nature 381, 571–579.
22. Gething,M.J.&Sambrook,J.(1992)Proteinfoldinginthecell.
Nature 355, 33–45.
23. Muchowski, P.J. & Clark, J.I. (1998) ATP-enhanced molecular
chaperone f unctions of the small he at shock pro tein human
alphaB crystallin. Proc. Natl A cad. Sci. US A 95, 1004–1009.
24. Wang, K. & Spector, A. (2000) Alpha-crystallin prevents irrever-
sible protein denaturation and acts cooperatively with other
heat-shock proteins to renature the stabilized partially denatured
protein in an ATP-dependent manner. Eur. J. Biochem. 267, 4705–
4712.

decrease molecular chaperone functions. J. Mol. Biol. 289,
397–411.
34. Rawat, U . & Rao, M. (1998) Interactions of chaperone alpha-
crystallin with the molten globule state of xylose reductase.
Implications for reconstitution of the active enzyme. J. Biol. Chem.
273, 9415–9423.
35. R eddy, M.C., Palmisano, D.V., Groth-Vasselli, B. & F arnsworth,
P.N. (1992) 31P NMR studies of the ATP/alpha-crystallin com-
plex: f unctional implications. Bio chem. Biophys. Res. Commun.
189, 1578–1584.
36. Smykal, P., Masin, J., Hrdy, I., Konopasek, I. & Zarsky, V. (2000)
Chaperone activity of tobacco HSP18, a small heat-shock protein,
is inhibited by A TP. Plant J. 23, 703–713.
37. Palmisano, D.V., Groth-Vasselli, B., Farnsworth, P.N. & Reddy,
M.C. (1995) Interaction of A TP and lens alpha crystallin char-
acterized by equilibrium binding studies and intrinsic tryptophan
fluorescence spectroscopy. Bioc him. Biophys. Acta. 1246, 91–97.
38. K antorow, M. & Piatigorsky, J. (1994) Alpha-crystallin/small heat
shock protein has autokinase a ctivity. Proc. Natl Acad. Sci. USA.
91, 3112–3116.
39. K antorow, M., Horwitz, J ., van Boekel, M.A., de Jong, W.W. &
Piatigorsky, J. (1995) Conversion from oligomers to tetramers
enhances autophosphorylation by lens alpha A-crystallin. Speci-
ficity between alpha A - and alpha B-crystallin s ubunits. J. Biol.
Chem. 270, 17215–17220.
40. Kantorow, M. & Piatigorsky, J. (1998) Phosphorylations of alpha
A- and alpha B-crystallin. Int. J. Biol. Macromol. 22, 307–314.
41. Haley, D.A., Horwitz, J. & Stewart, P.L. (1998) The small heat-
shock protein, alphaB-crystallin, has a variable quaternary struc-
ture. J. M ol. B iol. 27 7, 27–35.