Expression and characterization of the protein Rv1399c
from
Mycobacterium tuberculosis
A novel carboxyl esterase structurally related to the HSL family
Ste
´
phane Canaan
1
, Damien Maurin
1
, Henri Chahinian
2
,Be
´
ne
´
dicte Pouilly
1
,Ce
´
cile Durousseau
1
,
Fre
´
de
´
ric Frassinetti
1
,Lore
´

homologous three-dimensional structures reveals a canon-
ical catalytic triad (Ser162, His290 and Asp260) located at
the bottom o f a solvent accessible pocket lined by neutral or
charged residues. Based on this m odel, kinetic data o f the
Arg213Ala mutant partially explain the role of the guanid-
inium moiety, located close to His290, to confer an unusual
low pH s hift of the catalytic histidine in the wild type
enzyme. Overall, these data identify Rv1399c as a new
nonlipolytic hydrolase from M. t uberc ulosis and we thus
propose to reannotate its gene product as NLH-H.
Keywords: e sterase; nonlipolytic hydrolase; Rv1399c; tuber-
culosis.
The recent elucidation of the complete sequence of Myco-
bacterium tuberculosis genome [1] has offered new perspec-
tives for the search of novel drugs against tuberculosis. The
disease was responsible for around 2.5–3 million deaths in
2002 and t he World Health Organization (WHO) estimates
8 million new tuberculosis patients each year [2]. The
current genome annotation consists of  4000 predicted
proteins, classified into 11 distinct protein groups [3], of
which 48% have unknown function. Comparative sequence
analysis of the M. tuberculosis genome has revealed that it
contains 250 enzymes involved in lipid metabolism com-
pared to only 50 in Escherichia coli. Among these enzymes,
a family of 21 carboxyl este r h ydrolases, called Lip (A to W,
except K and S), have been annotated as putative esterases
or lipases, based on the presence of the consensus sequence
GXSXG characteristic of members of the a/b hydrolase
fold family. Within this family, the recent crystal structure
of the M. tuberculosis antigen 85C (Ag85C) [4], a mycolyl-

mole
´
cules Biologiques, AFMB UMR 6098, CNR S, 3 1 C hemin J oseph
Aiguier, 13402 Marseille Cedex 20, France. Fax: +33 491 16 45 36,
Tel.: +33 491 16 45 12, E-mail:
Abbreviations : HSL, h ormone-sensitive lipase; NLH, nonlipolytic
hydrolase; LH, lipolyt ic hydrolase; IB, inclusion bodies; IPTG, iso-
propyl thio-b-
D
-galactoside; CMC, c ritical m icellar c oncentration;
E600, diethyl para-nitrophenyl phosphate; D LS , dynamic light
scattering.
(Received 7 June 2004, accepted 17 August 2004)
Eur. J. Biochem. 271, 3953–3961 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04335.x
characterization, the Lip term could be c onfused to refer to
a NLH, and the original accession number should b e used
to avoid confusion.
Here we report the cloning, expression and refolding of
the M. tuberculosis Rv1399c gene product, a member o f the
Lip family. The biochemical characterization of Rv1399c,
using triacylglycerols and vinyl esters as substrates, demon-
strates that this carboxyl ester hydrolase must be classified
as an NLH instead of an LH as proposed initially by
bioinformatic tools [1,3]. Rv1399c efficiently hydrolyzes
short-chain t riacylglycerols and vinyl esters and has no
detectable activity against emulsified substrates. We thus
propose to rename Rv1399c NLH-H instead of LipH.
Sequence alignment has revealed that this enzyme defines a
new NLH that is structurally related to the hormone-
sensitive lipase (HSL) family and a model derived from a

confirmed by DNA sequencing.
In our hands, protein expression using the commercially
available pDest 17 plasmid frequently appeared to be
constitutive and occurred whether or not the isopropyl t hio-
b-
D
-galactoside (IPTG) inducer was added to the medium.
This problem, which was unsolved using the BL21(DE3)-
pLysS cells known to over-express lysozyme as a compet-
itive inhibitor of the T7 RNA polymerase, was most
probably due to the absence of the lac operator dowstream
of the T7 p romoter. To overcome this problem, a derivative
of the pDest 17 plasmid, pD est 1 7O/I, was constructed from
the pET-Dest42 plasmid and contains a P shAI-XbaI-
digested fragment that encompassed the Lac I gene (under
the control of a constitutive promoter) upstream the T7
promoter followed by Lac O (the DNA binding site of Lac
I). The resulting construct constitutively expresses Lac I,
which inhibits the T7 RNA polymerase in b inding to its
specific DNA binding site on Lac O.
Mutagenesis. Site-directed mutagenegis, which was per-
formed using the QuickchangeÒ site-directed mutagenesis
system (Stratagen, La Jolla, CA, USA), was used for the
mutation of Arg213 fi Ala. The oligonucleotides u sed
were: 5¢-gcgccaatcctggacgctgacgtcatcgacgcg-3¢ (forward)
and 5¢-cgcgtcgatgacgtcagcgtccaggattggcgc-3¢ (reverse). The
bases in bold indicate the location of the mutation. DNA
sequence of the mutant was confirmed by DNA sequencing
(Millegen, Prologue Biotech, France).
Protein expression. BL21(DE3)pLysS cells were trans-

of lysozyme and
1m
M
phenylmethanesulfonyl fluoride] and stored at
)80 °C overnight. T he pellet was thawed on ice for 1 h
andthen10lgÆmL
)1
DNase a nd 20 m
M
MgSO
4
(final
concentration) were added to the cell s uspension f or 30 min.
Cells were disrupted by ultrasonication (10· with a 15 s
cycle) using a Branson Sonifier 450. Inclusion bodies were
separated from t he cell e xtract by centrifugation at 17 000 g
for 30 m in. The pellet was then washed with 10 m
M
Tris/
HCl, pH 8.0 and 150 m
M
NaCl, sonicated (4· with a 15 s
cycle) and collected by centrifugation at 17 000 g for
20 min, this procedure was repeated three times. Inclusion
bodies were solubilized by stirring at 4 °C overnight in a
40 mL solution containing 10 m
M
Tris/HCl, pH 8.0,
150 m
M

Refolding. Purified Rv1399c was refolded by a dilution
method consisting of diluting the concentrated protein in
buffers with different pH and various compositions. The
refolding c onditions were determined by a refolding method
3954 S. Canaan et al. (Eur. J. Biochem. 271) Ó FEBS 2004
based on the measurement of the turbidity at 340 nm using
a 96-well plate [12]. The final refolding conditions consisted
of dilut ing Rv1 399c 20· in 50 m
M
Trisbuffer,pH 7,at4 °C
for 2 days. Rv1399c was concentrated up to 2–3 mgÆmL
)1
and traces of urea and imidazole were removed using a
desalting column ( HiPrep
TM
26/10, Amersham Bioscienc es).
Rv1399c was then concentrated up to 3 mgÆmL
)1
and
stored overnight at )20 °C, and after thawing, the ÔactiveÕ
refolded material was recovered by centrifugation at
17 000 g for 15 min. Rv1399c is stable in the refolding
buffer for at least 1 month at 4 °Candcanbestoredat
)20 °C for several months. Protein concentration was deter-
mined by measuring at A
280
using e
280
¼ 1.399 mg
)1

manufacturer’s instructions. Experiments were performed
at 20 °C with filtered (Millex syringe filters, 0.22 lm;
Millipore Corp.) p rotein samples (12 lLat3mgÆmL
)1
)
using a Dynapro MSTC-200 (Protein Solutions). All
calculations were performed using the software provided
by the manufacturer.
Kinetic assays. Enzymatic hydrolysis of solutions and
emulsions of various esters was followed potentiometrically
at 2 5 °C using a pH-stat (TTT 80, Radiometer, Copenha-
gen, Denmark) for at least 5 min. Assays were performed in
30 mL of 2.5 m
M
of Tris/HCl buffer, pH 7.0 and 0.1
M
NaCl. Release of fatty acid was titrated with 0.1
M
NaOH.
Enzymatic activity was expressed in units (U) per mg of
protein where 1 U corresponds to the liberation of 1 lmol
acidÆmin
)1
under standard conditions [13]. Assays using
olive oil and vinyl laurate were performed at pH 8.5 due to
the high pKa values of the oleic and l auric acids (8.1 and 7.4,
respectively).
pH stability – pH and temperature dependence activ-
ity. Tripropionin was chosen as substrate to perform the
pH and temperature dependence experiments rather than

M
) as a substrate. The pH value
was adjusted and the enzyme (7.4 lg) was added in the
vessel. Activity assays at acidic pH were corrected using the
calculated pKa value of 4.87 for butyric acid.
Inhibitor assay. Purified Rv1399c (2.51 nmol) was pre-
incubated at 25 °C with different E600 inhibitor/enzyme
molar ratio (2 and 10) in a final v olume of 32 lL. The
remaining activity was measured potentiometrically as a
function of time using tripropionin as above described.
Model building. The Rv1399c model was generated from
the automatic protein structure homology-modeling server
using
SWISS
-
MODEL
software (Biozentrum) [14–16], and
validated by the
PROCHECK
program [17]. Sequence align-
ment was performed initially using t he multiple sequence
alignment s oftware
T
-
COFFEE
[18], d isplayed wi th
ESPRIT
[19]
and then manually adjusted. The E600 inhibitor, which was
taken from the crystal structure of the cutinase–E600

buffer composition identified in the analytical s tep (50 m
M
Tris/HCl, pH 7.0).
Ó FEBS 2004 Biochemical characterization of Rv1399c from M. tuberculosis (Eur. J. Biochem. 271) 3955
The annotation of Rv1399c as a putative LH prompted
us to find a s uitable substrate, such a s t riglycerides and vinyl
esters, known to be s electively hydrolyzed by a l arge number
of NLH as well as LH. The preliminary biochemical
experiments allowed us to follow the specific activity of
Rv1399c along the refolding process a nd to estimate the
refolding y ield (Table 1). Using tripropionin as a substrate,
a specific activity of 80 UÆmg
)1
was obtain ed a fter the
dilution step and increased 6· after two days. After the
concentration s tep ( 2–3 mgÆmL
)1
), the specific activity
increased to 995 UÆmg
)1
, suggesting that some insoluble
material was s till refolded during t his process. The last
freezing/thawing step appeared to be an efficient and
powerful purification procedure. Indeed, whereas the total
activity was not affected significantly, the specific activity
increased from 1050 to 1350 UÆmg
)1
. This was due to the
presence of precipitated material accounting for 13% of the
total protein and arising from unfolded or misfolded

). Similarly, the K
1/2
value of Rv1399c (2.76 m
M
)
using soluble t riacetin at a concentration far b elow its CMC
(105 m
M
) [24] is similar to that of pig liver esterase (4 m
M
)
but is lower to that of acetylcholinesterase (30 m
M
). The
kinetic behavior of Rv1399c using short-chain soluble vinyl
esters and triacylglycerols along with the lack of d etectable
activity using insoluble vinyl esters (vinyl laurate) or
triacylglycerols (trioctanoin, olive oil) unambiguously clas-
sify Rv1399c as a NLH rather than a LH. However,
Rv1399c is able to hydrolyze a wide range of ester bonds
and does not show a substrate specificity towards the
alcohol or the acid moiety of short-chain esters (Table 2).
The pH (Fig. 2A) and temperature (Fig. 2B) stability
profiles of Rv1399c have been also investigated. The choice
of tripropionin (9.3 m
M
) as sub strate was g uided by t he high
stability of this compound in the w ide range of temperature
used. Our data indicate that Rv1399c is very sensitive to t he
pH as no activity was recorded after 1 h incubation in

116 kDa
66 kDa
45 kDa
35 kDa
25 kDa
234 5 6 7
Fig. 1. SDS/PAGE for expression and refolding of Rv1399c in E. coli.
Protein samples were lo aded on a 14% S DS/PAGE under reducing
and Coomassie-blue staining conditions. Lanes 1 and 7, molecular
mass markers (Fermentas); lane 2, E. coli proteins ( 30 lg) before
IPTG induction; lane 3, E. coli proteins ( 35 lg) after IPTG in duc-
tion; lane 4, purified Rv1399c (14.5 lg) eluted from the Ni-nitrilotri-
acetic acid column; lane 5, partially refolded Rv1 399c ( 6.8 lg); lane 6,
refolded Rv1399c after the freeze/thaw step (11 lg). The apparent
molecular mass of 36313 Da as estimated by Electrospray mass
spectrometry is due to the p resence, at the N-terminus, o f t he His
6
-tag
and the additional 21 residues from the expression plasmid.
3956 S. Canaan et al. (Eur. J. Biochem. 271) Ó FEBS 2004
hydrolases (Fig. 3). These enzymes share a functional
catalytic triad made of a catalytic nucleophile serine,
associated to a proton c arrier histidine and a c harge r elaying
aspartic (or glutamic) acid. To further investigate the
biochemical characterization of the enzyme, we have
titrated these key residues that form the catalytic triad of
Rv1399c.
Catalytic serine. Diethyl p aranitrophenyl phosphate
(E600), a specific powerful inhibitor of s erine hydrolases
was assayed on Rv1399c. As shown in Fig. 2D, the purified

Tripropionin
(9.28)
Tributyrin
(9.1)
Trioctanoin
–
Olive
oil
–
Acetylcholine
(33)
Rv1399c 610 1618 1123 0 184 1350 456 0 0 0
Acetylcholinesterase 970 210 0 0 450 0 0 0 0 1420
Pig liver esterase 320 300 470 0 60 50 70 0 0 ND
100
AB
CD
80
60
40
20
0
100
80
60
40
20
0
100
80

pKaWT
pKaMut
Fig. 2. Kinetic assays of Rv1399c. (A) pH stability, (B) temperature dependence. (C) Titration curves of wt Rv1399c (s) and the Arg213Ala mutant
(h) catalytic histidine residue. The curve profiles corresponding to pH values below 3.75 have been extrapolated due to Rv1399c instability at acidic
pH. (D) Rv1399c inhibitory effect using the E600 inhibitor as function of time. The protein/inhibitor ratio is indicated. Enzyme a ctivity was
determined as described in the Methods section.
Ó FEBS 2004 Biochemical characterization of Rv1399c from M. tuberculosis (Eur. J. Biochem. 271) 3957
Rv1399c is strongly inhibited by E 600 with a K
I
value in the
7to30· 10
)10
M
range, suggesting that the catalytic serine
is highly reactive. Moreover, the inhibition by the E600 is
not influenced by the presence of detergent, in contrast to
lid-containing human gastric and pancreatic lipases, sug-
gesting that the catalytic s erine o f R v1399c is fully accessible
to the solvent.
Catalytic histidine. It is widely assumed that serine
carboxyl ester hydrolases, like serine proteases, require an
appropriate protonation state of essential catalytic residues
in the active p H range. The imidazole ring of the catalytic
His must be in a neutral state to capture the hydrogen of the
catalytic serine for an efficient nucleophilic attack of the
substrate ester bond by the serine alcoholate. The shape of
the titration curve of the catalytic histidine shows an
apparent pKa value of the essential histidine estimated to
4.1 (Fig. 2C). This acidic pKa shift of the histidine residue
has been described previously for carboxypeptidase Y [25]

SWISS MODEL
server using, as templates, the co ordinates of structural
homologues of the a/b hydrolase fold family [33] that
present the highest sequence homologies (see above) with
Rv1399c: serine hydrolase (accession code 1evq) [29], and
carboxylester hydrolase (1jji) [30]. This m odel reveals the
overall topological organization of Rv1399c, predicts the
location of the catalytic triad, provides a valuable template
for further structure-function studies and is consistent with
our biochemical data. As expected, Rv1399c consists of
three domains. The largest domain encompasses strands b1
to b6 a nd strand b7 to the C-terminus. The central b sheet is
composed of 7 parallel b strands associated to an anti-
parallel strand (b2) and is surrounded by 5 helices (a1, a2,
a3, a7anda8). T he second domain c onsists of helices a4, a5
and a6 a ll clustered on the top of t he enzyme, as described
by Wei et al. [31] (Fig. 4). The third domain, which consists
of the 50 first amino acids, could not been modeled. This
Fig. 3. Amino-acid sequence alignment
between Rv1399c and four non lipolytic hydro-
lases (NLH) of known 3D structure. The
alignment was performed using the
T
-
COFFEE
and
ESPRIT
programs (available from the
expasy web site). Conserved residues are
boxed and those similar are indicated with a

Arg213
Arg213
Phe219
Phe219
Trp222 Trp222
Trp92
Trp92
Gly91
Gly91
Gly90
Gly90
Gly89
Gly89
Gly291
Gly291
T
y
r292 T
y
r292
Tyr190 Tyr190
Tyr190
Tyr190
His290
His290
His88
His88
Ser162
Ser162
Rv1399c Ag85c

from the b8-a8andb7-a7 loops, respectively. Another
important feature of the catalytic machinery is the so-called
oxyanion hole necessary for the stabilization of the oxyan-
ion transition state [34]. In Rv1399c, the oxyanion hole is
likely to be formed by the amides of Ala163, Gly89 and
Gly90, the latter’s being found in the HGGG consensus
sequence motif characteristic of members of the HSL family
[35]. Our model shows that the architecture of the active site
could accommodate an E600 inhibitor molecule covalently
bound to the catalytic Ser162 without d rastic conforma-
tional changes (Fig. 4C).
The acyl binding pocket (Fig. 4C) is delimited at one end
by the three hydrophobic side chains Trp92, Phe219 and
Trp222 that could play the role of filter thus preventing
binding of substrate with an acyl chain larger than eight
carbon atoms. This feature could explain the absence of
Rv1399c activity against triacylglycerols or esters with
longer carbon chain (Table 2). Moreover, t he three charged
residues Asp212, Arg213, Asp161 located at the periphery
of the pocket could have a role in substrate recognition,
suggesting a preference for a polar noncationic substrate as
a natural substrate since Rv1399c does not exhibit any
catalytic activity against acetylcholine (Table 2).
A structural s imilarity search, performed with DALI
using the coordinates of t he Rv1399c model as a template
identified numerous homologous proteins from the a/b
hydrolase fold family ( Z score v alues >10), including
Ag85C from M. t uberculo sis, a major protein component
of the cell w all [4]. The rmsd value between these two
structures is 3.2 A

bond could be reduced due to the proximity of the Arg213
guanidinium moiety that favor salt bridge formation with
theAsp260sidechain( 5.5A
˚
¢
) and induces a charge
repulsion with the His290 imidazolium (  2.5A
˚
¢
). This long
range hydrogen bond distance observed between His290
and Asp260, along with the close proximity of His290 to
hydrophobic residues from the b8-a8 loop (Trp189,
Tyr190), could f avor the displacement of the equilibrium
toward the neutral state of His290.
To attest this hypothesis, an Arg213Ala mutant has been
expressed and characterized. The specific activity of the
Ala213 mutant, using tripropionin as substrate, remains
identical (1280 UÆmg
)1
)tothewt enzyme (1350 UÆmg
)1
)
but the apparent pKa increases to  5.5. However, the
titration curve shows a plateau at acidic pH indicating that
Arg213 is not the only residue that dictates the a cidic
catalytic profile of His290 (Fig. 2C).
Concluding remarks
In summary, we have expressed, purified and refolded the
M. tu berculosis Rv1399c protein from inclusion bodies with

national network of Genopole.
3960 S. Canaan et al. (Eur. J. Biochem. 271) Ó FEBS 2004
References
1. Cole, S.T., Brosch, R., Parkhill, J., Garnier, T., Churcher, C.,
Harris, D., Gordon, S.V., Eiglmeier, K., Gas, S., Barry 3rd, C.E.,
et al. (1998) Deciphering the biology of Mycobacterium tuber-
culosis from the c omplete genome sequence. Nature 393, 537–544.
2. Dye, C., Sheele, S., Dolin, P., Pathania, V. & Raviglione, M.C.
(1999) Global burden o f tuberculosis: estimated incidenc, pre-
valence, and mortality by country. J. Am. Medical Assoc. 282.
3. Camus, J., Pryor, M., Medigue, C. & Cole, S. (2002) Re-annota-
tion o f the genome sequence of Mycobacterium tuberculosis
H37Rv. Microbiology 148, 2967–2973.
4. Ronning, D., Klabunde, T., Besra, G., Vissa, V., Belisle, J. &
Sacchettini, J. (2000) Crystal structure of the secreted form of
antigen 85C reveals potential targets for m ycobacterial drugs and
vaccines. Nat. Struct. Biol. 7, 141–146.
5. Wilson, R., Maughan, W., Kremer, L., Besra, G. & Futterer, K.
(2004) The structure of Mycobacterium tuberculosis MPT51
(FbpC1) defines a new family of non-catalytic alpha/beta hydro-
lases. J. Mol. Biol. 335, 519–530.
6. Sarda, L. & Desnuelle, P. (1958) Ac tion de la lipase pancre
´
atique
sur les esters en e
´
mulsion. Biochim. Biophys. Acta 30, 513–521.
7. Chahinian, H., Nini, L., Boitard,E.,Dubes,J.P.,Comeau,L.C.&
Sarda, L. (2002) Distinction between esterases and lipases: a
kinetic study with vinyl esters a nd TAG. Lipids 37, 653–662.

J.M. (1993) PROCHECK : a program to check th e stereochemical
quality of protein structures. J. Appl. Cryst. 26, 283–291.
18. Notredame, C., Higgins, D.G. & Heringa, J. (2000) T-Coffee: a
novel method for fast and accurate multiple sequence alignment.
J. Mol. Biol. 302, 205–217.
19. Gouet,P.,Courcelle,E.,Stuart,D.I.&Metoz,F.(1999)ESPript:
multiple sequence alignm ents in Post Script. Bioinformatics 15,
305–308.
20. Martinez, C., Nicolas, A., van Tilbeurgh, H., Egloff, M P.,
Cudrey, C., Verger, R. & Cambillau, C. (1994) Cutinase, a lipolytic
enzyme with a preformed oxyanion hole. Biochemistry 33, 83–89.
21. Christopher, J.A. (1998) SPOCK: the Structural Properties
Observation and Calculation Kit Program Manual. The Center
for Macromolecular Design, Texas A & M University, College
Station, TX.
22. Longhi, S., Mannesse, M., Verheij, H.M., de Haas, G.H.,
Egmond, M., Knoops-Mouthuy, E. & Cambillau, C. (1997)
Crystal structure of cutinase covalen tly inhibited by a triglyceride
analogue. Protein Sci. 6, 275–286.
23. Kawasaki, K., Kondo, H., Suzuki, M., Ohgiya, S. & Tsuda, S.
(2002) Alternate conformations observed in catalytic serine of
Bacillus subtilis lipase determined at 1.3 A
˚
resolution. Acta Crys-
tallogr. D Biol Crystallogr. 58, 1168–1174.
24. Entressangles, B. & Desnuelle, P. (1968) Ac tion of pancreatic
lipase on aggregated glyceride molecules in an isotropic system.
Biochim. Biophys. Acta 159, 285–295.
25. Stennicke, H., Mortensen, U. & Breddam, K. (1996) Studies on
the hydrolytic properties of ( serine) c arboxypeptidase Y. Bio-

Biol. 6, 340–345.
32. Zhu,X.,Larsen,N.A.,Basran,A.,Bruce,N.C.&Wilson,I.A.
(2003) Observation of an arsenic adduct in an acetyl esterase
crystal structure. J. Biol. Chem. 278, 2008–2014.
33. Ollis, D.L., Cheah, E., Cygler, M., Dijkstra, B., Frolow, F.,
Franken, S.M., Harel, M., Remington, S.J., S ilman, I., Schr ag, J.,
Sussman, J.L., Verschueren, K.H.G. & Goldman, A. (1992) The
a/b hydrolase fold. Protein Eng. 5, 197–211.
34. Kraut, J. (1977) Serine proteases: Structure and mechanism of
catalysis. Ann. Rev. Biochem. 46, 331–358.
35. Contreras, J.A., Karlsson, M., Osterlund, T., Laurell, H., Svens-
son, A. & Holm, C. (1996) Hormone-sensitive lipase i s structurally
relatedtoacetylcholinesterase,bile salt-stimulated lipase, and
several fungal lipases. Building of a three- dimensional model for
the catalytic domain o f hormone-sensitive lipase. J. Biol. Chem.
271, 31426–31430.
36. Fojan, P., Jonson, P., Petersen, M. & Petersen, S. (2000) What
distinguishes an esterase from a lipase: a novel structural
approach. Biochimie 82, 1033–1041.
37. Camacho, L., Ensergueix, D., P er ez, E., Gicquel, B. & G uilhot, C.
(1999) Identification o f a virulence gene cluster of Mycobacterium
tuberculosis by signature-tagged transposon mutagenesis. Mol.
Microbiol. 34, 257–267.
Ó FEBS 2004 Biochemical characterization of Rv1399c from M. tuberculosis (Eur. J. Biochem. 271) 3961