MOLECULAR

EEMKAL
PARASITOLOGY
ELSEVIER Molecular and Biochemical Parasitology 71 (1995) 203-210

Predicted disulfide-bonded structures for three uniquely related

proteins of Plasmodium falciparum, Pfs230, Pfs48/45 and Pf12
Richard Carter a,*, Andrew Coulson b, Sunita Bhatti ‘, Brian J. Taylor ‘,
John F. Elliott ’
aInstitute of Cell, Animal and Population Biology, Division of Biological Sciences, University of Edinburgh, West Mains Road,
Edinburgh EH9 3JT, UK
’ Institute of Cell and Molecular Biology, Division of Biological Sciences, University of Edinburgh, West Mains Road,
Edinburgh EH9 3JR, UK
’ Department of Medical Microbiology and Infectious Diseases, Glaxo Heritage Research Institute, 621 HMRC, University of Alberta,
Edmonton, Alberta, Canada T6G 2S2

Received 7 October 1994; accepted 9 March 1995

Abstract

Pfs230 is a surface protein of the gametes of Plasmodium fulciparum and has been demonstrated to be a target of malaria
transmission-blocking antibodies; it is an important candidate antigen for a transmission-blocking vaccine. The target
epitopes of transmission-blocking antibodies against Pfs230 are almost all reduction sensitive suggesting that disulfide bonds
are critical for folding the native molecule. Following the cloning of the Pfs230 gene attempts are now underway to express
subunits of the protein for use in vaccine trials. It will be important to understand the disulfide-bond structure of the Pfs230
to achieve this goal. In this paper we present a model for this structure based on the observation that the Pfs230 molecule
contains a series of regularly repeated cysteine-containing motifs. Four such motifs have been identified, together with a fifth
cysteineless motif, which occur in the same relative order, with regular alternating omission of specific motifs, 14 times
throughout the length of the protein. Each of the 14 sets of motifs contains an even number of cysteine residues (2, 4 or 6).
We postulate that each set folds into a separate disulfide-bonded domain in which corresponding pairs of cysteines form an
equivalent disulfide bond in every such domain. The postulated bonding arrangements in the different domains are mutually
confirmatory throughout the sequence of Pfs230. We have identified two other malaria proteins, Pfs48/45 and Pf12, which
share the same arrangements of motifs and conform to the same disulfide-bond structure proposed for Pfs230; no other
proteins in the sequence data base share these characteristics. Although having similarities to the ‘cystine knot’ described in
other proteins, the arrangement proposed here appears to be unique among described structures.

Keywords: Plasmodium falciparum; Sexual-stage antigen; Transmission-blocking vaccine; Protein structure

1. Introduction

Transmission-blocking vaccines have the potential

to make a significant impact on malaria transmission
* Corresponding author. Tel.: (44-131) 6.50-5558; Fax: (44-131) rates. Pfs230, a sexual stage surface antigen of Plas-
668-3861; e-mail: r.carter@ed.ac.uk modium falciparum, is an important candidate anti-

0166-6851/95/$09.50 0 1995 Elsevier Science B.V. All rights reserved

SSDI 0166-6851(94)00054-2
204 R. Carter et al. /Molecular and Biochemical
gen for use in a transmission-blocking vaccine against
this parasite. The gene encoding Pfs230 was cloned
independently by Williamson and colleagues [l]
(GenBank accession No. M8135) and also in one of
our laboratories (Bhatti and Elliott, unpublished,
GenBank accession No. LO4162). These two se-
quences for Pfs230 were derived from DNA libraries
prepared independently from the same cloned line of
P. fulciparum, 3D7 [2]. Except for one correction [l],
both sequences were in complete agreement.
The sequence of Pfs230 contains 70 cysteines.
Since almost all target epitopes of anti-Pfs230 mono-
clonal antibodies are destroyed by reduction [3-51,
we expect that the antigenicity of these epitopes will
be found to be dependent upon correct formation of
disulfide bonds. An understanding of the disulfide-
bond structure of Pfs230 will almost certainly be
crucial, therefore, to attempts to express regions of
the molecule in immunogenic form for vaccine de-
velopment. Based on an analysis of the sequence of
Pfs230, we present here a postulated arrangment for
the disulfide bonds in Pfs230 and also for those of
two other proteins of P. falciparum, Pfs48/45, an-
other gamete surface protein and transmission-block-
ing vaccine candidate [6] (GenBank accession No.
222145) and Pf12, a putative asexual stage, mem-
brane-associated protein cloned and sequenced in
one of our laboratories [7] (GenBank accession No.
M28889).
In a previous analysis of the sequence of Pfs230
Williamson and colleagues [ll recognised a 7-cy-
steine-containing motif with a loosely conserved
amino-acid sequence. They described 6 copies of this
motif distributed through the C-terminal four-fifths
of the molecule following the region of glutamate-
rich repeats. In the present study we have re-ex-
amined the sequence of Pfs230 and have made a new
analysis of the motifs and their relationships to each
other. This analysis suggests a logically consistent
model for the disulfide-bond structure of Pfs230
which applies equally to that of Pfs48/45 and Pf12.
2. Materials and methods
2.1. Parasites
The sequences used in the analysis in this paper
are from P. falciparum clone 3D7 for Pfs230 [l],
Parasitology 71 (1995) 203-210
from P. falciparum isolate NF54 (the parent isolate
of 3D7) for Pfs48/45 [6] and from the FMG isolate
of P. falciparum for Pf12 [7].
2.2. Criteria for identifying and defining the amino-
acid motifs and their representative sequences in
Pfs230, Pfs48 / 45 and Pf12
By visual inspection of the amino-acid sequences
of Pfs230, Pfs48/45 and Pf12, four distinct sets of
cysteine-containing motifs and a fifth cysteineless
motif were recognised. Each set of motifs is charac-
terised by the presence of more or less conserved
amino acids at specific positions in the motif. In
order to confirm that these sequences, identified by
eye as representing a given motif, were truly related,
the following tests were carried out.
Each amino acid in each sequence representing
the motif is given a score, the Individual Position
Score. This score is the number of times that that
particular amino acid occurs at that position in the
motif in all three proteins (in this analysis leucine
and isoleucine have been treated as equivalent). The
minimum Individual Position Score is 1, i.e., the
specific amino acid would have been found at that
position only once among all the representatives of
the motif. The maximum possible Individual Position
Score is obtained when the same amino acid occurs
at the position in question in all representatives of
the motif. By averaging the Individual Position
Scores at a particular position for all representatives
of the motif, a number is obtained which reflects the
degree to which the same amino acids recur at that
position. The higher this Mean Individual Position
Score, the more often the same amino acids have
occurred at that position in the motif.
In order to determine the extent to which these
Mean Individual Position Scores deviate from those
which would have resulted from random distribu-
tions of amino acids, they were divided by a ‘con-
trol’ score representing a random selection of amino
acids from the motif. Thus, the Control Individual
Position Score is the average of the Mean Individual
Position Scores obtained for several sets (the number
in each set being the number of representatives of
the motif among all 3 proteins) of amino acids
selected at random from the motif. We have desig-
nated the ratios derived in this way as Position
Conservation Scores, as they represent the mean
 R. Carter et al. /Molecular and Biochemical Parasitology 71 (1995) 203-210 205

number of times that the amino acids at the relevant
position in the motif recur more frequently than they
would have done had the amino acids of the putative
representatives of the motif been distributed ran-
domly.
Because the Position Conservation Scores (data
not shown, but submitted to editorial scrutiny) are
almost all greater or much greater than unity we
conclude that the sequences of members of each
putative motif are, on average, related to each other
and that the motifs are, therefore, real entities.
It was also necessary to determine whether each
putative representative of an individual motif was
sufficiently related to the other members of the motif
to justify its inclusion in the motif. This was done by
determining a Sequence Conservation Score. For
each putative representative of a motif the Individual
Position Scores were averaged and divided by the
Control Individual Position Score. This ratio is the
Sequence Conservation Score for that putative repre-
sentative of the motif and represents the mean num-
ber of times that the amino acids in that particular
sequence recur more frequently than they would
have done had the amino acids in the putative repre-
sentative of the motif been distributed at random.
All the putative representatives of each motif had
Sequence Conservation Scores of greater than unity
(Table 1) confirming the validity of their inclusion as
representatives of the motif in question.
Tables with full sequences and Individual Position
Scores for all representatives of the motifs are avail-
able on request to the authors.
As further proof of the validity of the motifs and
the identity of their representative sequences it is
noted that throughout all three proteins in which we
have identified them, the representatives of the mo-
tifs always occur in the same order i.e., Motif 1 to 2
to 3 to 4 to 5 from the N-terminal end of a protein.

Table 1
Sequence Conservation Scores of representatives of motifs in Pfs230, Pfs48/45 and Pf12
Domain Motif Mean Sequence
1 2 3 4 5 Conservation Score
for each Domain

Pfs230
I _ 3.4 1.9 2.3 2.5
II 1.6 3.2 1.9 2.7 1.8 2.2
III _ 3.0 1.9 - 2.3 2.4
IV 2.2 3.2 2.0 3.2 2.5 2.6
V _ 2.3 2.0 2.1 2.1
VI 1.6 3.2 2.1 2.4 2.6 2.4
VII _ 3.0 1.9 2.9 2.6
VIII 1.9 3.7 2.0 3.1 2.3 2.6
IX _ 3.1 1.6 2.3 2.3
X 1.2 3.0 2.4 2.9 2.2 2.3
XI 2.4 2.3 2.6 2.4
XII 1.7 3.4 2.1 3.3 1.6 2.4
XIII _ 1.5 1.2 2.4 1.7
XIV 1.6 3.6 1.2 3.2 2.3 2.4

Pfs48/45
II _ 2.7 1.5 3.4 2.6 2.6
III - 1.9 1.4 2.7 2.0 2.0
IV 1.4 3.4 2.0 3.0 2.0 2.4

Pf12
I _ 2.9 1.8 3.4 2.1 2.6
II 1.9 3.7 1.9 2.2 2.4 2.4

Mean Sequence Overall mean

Conservation Score for
each Motif 1.7 3.0 1.8 3.0 2.3 2.3
206 R. Carter et al. /Molecular and Biochemical Parasitology 71 (1995) 203-210

In instances where representatives of individual mo-
tifs are absent, the representatives of the remaining
motifs still occur in numerical order (Figs. 1 and 2).
2.3. Numbering of motifs and domains
Domains are numbered, using Roman numerals,
in the order in which they occur from the N-terminal
end of a protein. Motifs are numbered, using Arabic
numerals, in the order in which they occur from the
N-terminal end of a domain in which a full set of the
possible motifs is present. The full possible set of
motifs is five and the motifs are thus numbered 1 to
5. However, as already noted, not all domains con-
tain the full set of possible motifs. In such cases the
representatives of a motif retain their generic num-
ber, i.e., the number they would have had if the full
set of motifs had been present in that domain.
(with one exception in Pfs 48/45, see Results) con-
tain an even number of cysteines from 2 to 6. The
proposed model for disulfide-bonding structures in
the three proteins Pfs230, Pfs48/45 and Pf12 is
based on two assumptions: (i) that each set of motifs
forms a disulfide-linked structural domain containing
1, 2 or 3 disulfide bridges and (ii) that in each
domain analogous pairs of cysteines are always
bonded to each other.
3. Results
Following the methods of analysis described in
Materials and methods the disulfide-bond arrange-
ments illustrated in Figs. 1 and 2 were deduced for
the three proteins Pfs230, Pfs48/45 and Pf12. These
structures have the following features.

2.4. Criteria for deducing the arrangement of disul- 3.1. Pfs230

fide-bond pairing of cysteines in Pfs230, Pfs48 / 45
There are 14 motif-defined, cysteine-containing
and Pf12
domains in Pfs230 numbered from the N-terminal
Representatives of each set of cysteine-containing end of the protein (Fig. 1, see also Tables 2 and 3).
motifs, defined and tested as described above, all Domains of the odd-numbered series lack the cys-

Table 2
Positions of Motifs l-5
Domain Motif No. of Cys

I 2 3 4 5
Pfs230
_ 589- 597 611- 627 _ 691- 706 4
II 718- 732 733- 741 781- 796 796- 809 849- 864 4
III _ 918- 926 944- 959 _ 1097-1114 4
IV 1121-1135 1136-1144 1161-1176 1192-1205 1235-1251 6
V _ 1285-1293 1311-1327 1393-1408 4
VI 1420-1434 1435-1443 1459-1474 1475-1488 1519-1534 6
VII - 1694-1702 1726-1741 _ 1866-1881 4
VIII 1895-1909 1910-1918 1938-1953 1955-1968 2002-2017 6
IX 2052-206 1 2074-2089 _ 2163-2179 2
X 2190-2203 2204-2212 2229-2244 2246-2259 2341-2356 6
XI 2448-2456 2476-2491 _ 2623-2638 4
XII 2651-2666 2666-2674 2706-2721 2721-2735 2789-2804 6
XIII 2831-2839 2846-2862 _ 2936-2951 2
XIV 2967-2981 2982-2990 3010-3025 3027-3040 3075-3090 6

Pfs48/45
II - 45- 53 71- 87 94- 107 143- 158 5
III 182- 190 210- 228 226- 239 259- 275 4
IV 279- 293 294- 302 327- 345 345- 357 397- 412 6

Pf12
27- 35 53- 68 73- 86 123- 137 6
II 160- 174 175- 183 211- 226 228- 241 271- 286 6
 R. Carter et al. /Molecular and Biochemical Parasitology 71 (1995) 203-210 207

teine-containing Motif 4. The odd- and even-num- Table 3

bered domains form a series of paired structures Lengths of ‘a’ and ‘b’ loops and of B and J segments

repeated seven times. Odd- and even-numbered do- Domain J segment B segment ‘a loop’ ‘b loop’
mains are connected from the C-termini of Motifs 5 Pfs230
in the odd-numbered domains by a bridge segment I _ 14 64
(B segment) which consists of a short sequence II _ 27 40 40
III 54 _ 18 138
followed by the cysteineless Motifs 1 to the N-termini
IV 22 17 30
of Motifs 2 of the even-numbered domains. In all 14 V 34 _ 18 66
domains Motifs 2 and 3 are connected by relatively VI _ 27 16 31
short loops which we refer to as ‘a loops’. We VII 160 _ 24 125
VIII - 29 20 34
predict that each pair of odd- and even-numbered
IX 35 _ 13 74
domains, associated ‘a loops’ and B segment, com-
X 25 17 82
XI 92 - 20 132
XII _ 28 32 54
XIII 27 7 74
XIV _ 31 20 35

Pfs48/45
II 18 36
III 24 20 20
IV _ 19 25 40

Pf12
I _ 18 37
II - 38 28 30

prises a single semi-conserved structure repeated

Fig. 1. Predicted disulfide-bond structure for Pfs230. Motifs are
indicated with Arabic numerals and domains with Roman numer-
als. The position of each cysteine is marked by a closed circle; the
position of each proline is marked by a stroke. The loop regions
associated with each domain are marked ‘a’ and ‘b’ respectively,
the ‘bridge’ sections connecting odd- to even-numbered domains
are marked ‘B’ and the ‘junction’ sections connecting even- to
odd-numbered domains are marked ‘J’. The glutamate-rich repeats
at the N-terminal end of the molecule are marked by the zigzag
patterns.
through the molecule.
Extending from these relatively conserved central
structures are cysteine-free loops connecting the C-
termini of Motifs 3 or 4, for odd-and even-numbered
domains respectively, to the N-termini of Motifs 5;
we refer to these as ‘b loops’. The ‘b loops’, while
varying considerably in size, are all larger or much
larger than all but the largest ‘a loop’ (in domain II)
and among them we have detected no significant
homologies.
Between the entire paired structures (semi-con-
served paired domains with ‘a loops’, B segments
and subtended ‘b loops’) are ‘junction’ (J) segments.
These are very variably sized cysteine-free sections
which connect the C-termini of Motifs 5 of the
even-numbered domains with the N-termini of Mo-
tifs 2 of the odd-numbered domains.
3.2. Pfs48 / 45
Pfs48/45 (Fig. 2A, see also Tables 2 and 3)
contains the equivalent of 1 l/2 paired domains; by
208 R. Carter et al. /Molecular and Biochemical Parasitology 71 (1995) 203-210

comparison with the arrangement described for

Pfs230 the first, odd-numbered domain of Pfs 48/45
and the linking B segment is absent. For this reason
we have designated the N-terminal domain of Pfs
48/45 (which must be even-numbered to follow the
convention described for Pfs 230) Domain II. This is
linked by a short J segment to Domain III which is
linked by a B segment to Domain IV. In contrast to
the consistently diminished odd-numbered domains
of Pfs230 (lacking Motif 4 throughout), Domain III
of Pfs48/45 contains the full complement of Motifs
2 to 5. However, Domain III lacks the cysteine pair
which putatively links Motifs 2 and 3 elsewhere.
Domain II is unique among all three molecules by
lacking the second cysteine in Motif 3 which is
postulated to link elsewhere to the second cysteine of
Motif 5.

3.3. Pf12

The motifs and two domains of Pf12 (Fig. 2B, see

also Tables 2 and 31, which follow the same general

Fig. 3. A comparison of the arrangement of (A) the ‘cystine knot’

with (B) a hypothetical structure related to the cystine knot and
consistent with the 6-cysteine-containing domains postulated for
Pfs230, Pfs48/45 and Pf12. Cysteines are indicated by a C
enclosed in a circle and are numbered in their linear order from
the N-terminal end of a protein. The disulfide bonds between the
cysteines are indicated by the zigzag lines. The curved lines
represent the linear amino-acid sequences between the cysteines.

pattern that we have described for Pfs230 and

Pfs48/45, are organised in the most conserved, pos-
sibly primordial, form. Both domains consist of a
full complement of cysteine-containing Motifs 2 to
5, and each has 6 cysteines; they are connected via a
B segment containing the cysteineless Motif 1. It is
noteworthy that the ‘b loops’ of Pf12 are very small
in comparison to those of Pfs230 and even Pfs48/45,
being similar in size to ‘a loops’ which are almost
uniformly small in all three molecules.

3.4. Relationship of proposed disulfide-bonding

structures to others described in the data base

Fig. 2. Predicted disulfide-bond structures for (A) Pfs48/45 and Small polypeptide structures are often stabilised
(B) Pf12. See the legend to Fig. 1 for explanation of symbols. by a cluster of disulfide bridges (Fig. 3). The pattern
 R. Carter et al. /Molecular and Biochemical Parasitology 71 (1995) 203-210 209

of disulfide bridges postulated here for the domains appear to be the most likely location of such disul-
of the Plasmodium proteins was compared with fide-bond-dependent epitopes. The cysteine-free ‘b
patterns in known structures, since a close analogy loops’ and junctional (J) regions of these molecules
might indicate structural and indeed functional simi- are less likely to form epitopes dependent upon
larity. disulfide bonding.
The ‘Iditis’ database (version 2.02) [8] was Because so few mAb are generated against epi-
scanned for all proteins containing the cyclic struc- topes which are not disulfide-bond dependent, we
ture Cys(a)-S.S-Cys(b)-X-Cys(c)-S.S-Cys(d)-X,- speculate that the ‘b loops’ and J regions of the
Cys(a) together with at least one more disulfide molecules may be inaccessible to the immune system
bridge (in the proposed structures of the Plasmodium perhaps because they are internal to the molecule or
proteins, n = 9,lO or, in one case, 24 (Pfs230); 6 or are in close contact with the gamete surface. As they
7 (Pfs 48/45); and 9 or 13 (Pfsl2)). The correspond- are so readily recognised on the surface of intact
ing entries were retrieved from the Brookhaven gametes of P. falciparum by antibodies, we propose
database and the proteins considered individually. that the disulfide-bonded domain structures form the
Three of these proteins (defensin, methylamine outfacing surfaces of these molecules in contact with
dehydrogenase and Bowman-Birk soybean trypsin the external milieu.
inhibitor) have at least one more disulfide bridge The arrangements of motifs and disulfide-bond
rooted in the Cys(d) to Cys(a) segment and this pairings predicted for Pfs230, Pfs48/45 and Pf12
makes a structural analogy to the Plasmodium pro- appear to be unique in the currently available
teins unlikely. databases. However, the general arrangement of the
In other cases secondary structure elements con- 6-cysteine, disulfide-bonded domains postulated here
strain n to values outside the range found here has features similar to, but nevertheless distinct from,
(neuraminidase, n = 1; scorpion neurotoxin and the ‘cystine knot’ [ll] described, for example, in the
transferrin, n = 4). hormone, human chorionic gonadotropin [9] and the
All remaining cases are examples of the cystine growth factors, transforming growth factor beta 2,
knot (9-11) in which a ring is formed by parallel platelet-derived growth factor BB and nerve growth
peptide segments and two disulfide bridges with a factor [lo].
third disulfide bridge threaded through it (Fig. 3A). In conclusion, we have derived a set of predicted
In the known cystine knots n = 3 or in one case 9. A disulfide-bonding arrangements for three uniquely
similar structure to the cystine knot could be pro- related proteins of P. falciparum, two of which,
posed for the Plasmodium domains but the peptide Pfs230 and Pfs48/45, are expressed on the surface
strands of the ring would be anti-parallel (Fig. 3B). of the gametes of the parasite and are important
candidates for malaria transmission-blocking vac-
cines [13]. Based on our present interpretation of
4. Discussion their disulfide-bond structure, together with data from
studies with transmission-blocking mAb against the
With rare exceptions [5,12] virtually all mono- proteins, we propose that most of the target epitopes
clonal antibodies (mAb) against the native forms of of these mAb will be found within the disulfide-
Pfs230 and Pfs48/45 react with epitopes whose bonded, motif-defined, domains in these molecules.
structures depend upon intact disulfide bonds [3,4,12]. Verification of the proposed bonding arrangements
These mAb all react with epitopes which are ex- will depend upon biochemical and structural chemi-
posed on the surface of the intact gametes. This cal analyses of these proteins.
suggests that the disulfide-bonded regions of the
molecules represent the aspects that are accessible to
the external milieu of the intact gametes. In our Acknowledgements
present interpretation of the structure of these pro-
teins, the disulfide-bonded domains, defined by Mo- We thank Prof. Kamini Mendis and Dr. Ali Sul-
tifs 2 to 5 together with the small ‘a loops’, would tan for discussion and suggestions which were of
210 R. Carter et al. /Molecular and Biochemical Parasitology 71 (1995) 203-210

great value to the development of this analysis. The
work was supported by the Medical Research Coun-
cil of the UK, the Medical Research Council of
Canada and by a grant from the UNDP/WORLD
BANK/WHO Special Programme for Research and
Training in Tropical Diseases (TDR).
References
Dl Williamson, KC., C&o, M.D. and Kaslow, DC. (1993)
Cloning and expression of the gene for Plasmodium falci-
parum transmission-blocking target antigen, Pfs230. Mol.
Biochem. Parasitol. 58, 355-358.
[21 Walliker, D., Quakyi, LA., Wellems, T.E., McCutchan, T.F.,
Szarfman, A., London, W.T., Corcoran, L.M., Burkot, T.R.
and Carter, R. (1987) Genetic analysis of the human malaria
parasite Plasmodium falciparum. Science 236, 1661-1666.
131 Quakyi, LA., Carter. R., Rener, J., Kumar, N., Good, M.F.
and Miller, L.H. (1987) The 230~kDa gamete surface protein
of Plasmodium falciparum is also a target for transmission-
blocking antibodies. J. Immunol. 139, 4213-4217.
t41 Read, D., Lensen, A.H.W., Bergamie, S., Haley, S., Raza, A.
and Carter, R. (1994) Transmission-blocking antibodies
against multiple, non-variant target epitopes of the Plasmod-
ium falciparum gamete surface antigen Pfs230 are all com-
plement-fixing. Parasite Immunol. 16, 511-520.
[51 Wizel, B. and Kumar, N.(1991) Identification of a continu-
ous and cross-reacting epitope for Plasmodium falciparum
transmission-blocking immunity. Proc. Natl. Acad. Sci. USA
88, 9533-9537
[61Kocken, C.H.M., Jansen, J., Kaan, A.M., Beckers, P.J.A.,
Ponnudurai, T., Kaslow, DC., Konings, R.N.H. and Schoen-
makers, J.G.G. (1993) Cloning and expression of the gene
coding for the transmission blocking target antigen Pfs48/45
of Plasmodium falciparum. Mol. Biochem. Parasitol. 61,
59-68.
[71 Elliott, J.F., Albrecht, G.R., Gilladoga, A., Handunetti, S.M.,
Neequaye, J., Lanllinger, G., Minjas, J.N. and Howard, R.J.
(1990) Genes for Plasmodium falciparum antigens cloned by
expression in COS cells. Proc. Nat1 Acad. Sci. USA 87,
6363-6367.
181Thornton, J.M. and Gardner, S.P. (1989) Protein motifs and
data base searching. Trends B&hem. Sci. 14, 300.
[91 Laptbome, A.L., Harris, D.C., Littlejohn, A., Lustbader,
J.W., Canfield, R.E., Machin, K.L., Morgan, F.J. and lsaacs,
N.W. (1994) Crystal structure of human chorionic go-
nadotropin. Nature 369, 455-461.
Murray-Rust, J., McDonald, N.Q., Blundell, T.L., Hosang,
M., Oefner, C., Winkler, F. and Bradshaw, R.A. (1993)
Topological similarities in TGF-beta 2, PDGF-BB and NGF
define a superfamily of polypeptide growth factors. Structure
1, 153-159.
McDonald, N,Q. and Hendrickson, W.A. (1993) A structural
superfamily of growth factors containing a cystine knot
motif. Cell 73, 421-424.
Carter, R., Graves, P. M., Keister, D.B and Quakyi, LA.
(19901 Properties of epitopes of Pfs48/45, a target of trans-
mission blocking monoclonal antibodies, on gametes of dif-
ferent isolates of Plasmodium falciparum. Parasite Immunol.
12,587-603.
Carter, R. (1994) Malaria transmission-blocking vaccines
against sexual and mosquito stages of the parasites: the
current state of development. In: Recombinant and Synthetic
Vaccines (Talwar, G.P., Roa, K.V.S. and Chauhan, V.S.,
eds.), Chapter 21. Narosa, New Delhi, pp. 147-158.