You are on page 1of 63

By DANIEL C. CARTER and JOSEPH X.

HO

Space Sclence Laboratory, Blophyslcs Branch, Natlonal Aeronautics and Space


Admlnlatratlon, Marshall Space Fllght Center, Huntsville, Alabama 35812

1. Introduction ....................................................... 153


11. Albumin Structure .................................................. 155
A. Primary Structure .............................................. 155
B. Early Concepts of Albumin Structure ............................. 161
C. Three-Dimensional Structure .................................... 166
111. Nature of Ligand Binding ........................................... 176
A. Small Organic Compounds: Sites I and I1 ......................... 181
B. Long-Chain Fatty Acids: Sites Ill and IV .......................... 186
C. Metals: Sites V and VI ........................................... 188
IV. Evolution of Albumin Structure ...................................... 190
V. Summary and Future Directions ...................................... 194
References ......................................................... 196

I. INTRODUCTION
Often utilized as a substitute for a typical protein, albumin needs no
introduction to the protein chemist. Because of its availability, low cost,
stability, and unusual ligand-binding properties, serum albumin has been
one of the most extensively studied and applied proteins in biochemistry.
However, as a protein, albumin is far from typical, and the widespread
interest in and application of albumin have not been balanced by an
understanding of its molecular structure. Indeed, for more than 30 years
structural information was surmised based solely on techniques such as
hydrodynamics, low-angle X-ray scattering, and predictive methods.
Serum albumin was recognized as a principal component of blood as
early as 1839 (Ancell, 1839). Early researchers generally referred to
protein as “albumen” stemming from Latin albus after the white color of
flocculant precipitates produced by various proteins. Today, one should
be aware that several proteins share this name but they are structurally
and functionally unrelated to serum albumin, e.g., ovalbumin and preal-
bumin. Hence, care should be taken to distinguish ablumen, which refers
to egg whites, from albumin or serum albumin.
As the most abundant protein in the circulatory system and with typical
blood concentrations of 5 g/100 ml, albumin contributes 80% to colloid
osmotic blood pressure. In addition, it has now been determined that
albumin is chiefly responsible for the maintenance of blood pH (Figge et

ADVANCES IN 153 Copyright 01994 by Academic Press, lnc.


PROTEIN CHEMISTRY, Vol. 45 All rights of reproduction in any form reserved.
154 DANIEL C. CARTER A N D JOSEPH X. HO

al., 1991). It is located in every tissue and bodily secretion, with the
extracellular protein comprising 60% of total albumin. In mammals
albumin is synthesized by the liver (Peters and Anfinsen, 1950) and
possesses a half-life in circulation of 19 days (Waldmann, 1977). Despite
the major role of albumin in circulating plasma, the absence of albumin,
or “analbuminemia,” though rare, has been observed in several individ-
uals (Russi and Weigand, 1983) and one strain of rats (Nagase et al.,
1979). Still, albumin is expressed in these individuals, but at a much lower
level. Clinical manifestations of analbuminemia are not completely un-
derstood, but include chronic fatigue, hyperlipidemia, and, in rats, an
increased susceptibility to cancer (Kakizoe and Sugimura, 1988).
Perhaps, the most outstanding property of albumin is its ability to bind
reversibly an incredible variety of ligands. (See later, Table V, which
provides a selective listing of the extensive literature regarding ligand
binding to albumin; see also Section 111.) Befitting its inclusion in this
volume, albumin is the principal carrier of fatty acids that are otherwise
insoluble in circulating plasma. But albumin performs many other func-
tions as well, such as sequestering oxygen free radicals and inactivating
various toxic lipophilic metabolites such as bilirubin (Emerson, 1989).
Although albumin has a broad affinity for small negatively charged
aromatic compounds, it has high affinities for fatty acids, hematin, and
bilirubin. Additionally, it forms covalent adducts with pyridoxyl phos-
phate, cysteine, glutathione, and various metals, such as Cu(11), Ni(II),
Hg(II),Ag(II), and Au(1). The participation of albumin as the key carrier
o r reservoir of nitric oxide, which has been implicated in a number of
important physiological processes, including neurotransmission, serves
to illustrate further the continuing recognition of the utility of albumin
(Stamler et al., 1992). Without question, albumin is the most multifunc-
tional transport protein known to date.
Albumin belongs to a multigene family of proteins that includes a-
fetoprotein (AFP) and vitamin D-binding protein (VDP),also known as G
complement (Gc) protein. Although AFP is considered the fetal counter-
part of albumin, its binding properties are distinct and it has been sug-
gested that AFP may have a higher affinity for some unknown ligands
important for fetal development. VDP plays an important role in calcium
regulation. Additionally, AFP and VDP both interact with the class I1
major histocompatibility complex. These and other data suggest that
these proteins may play an important role in modulating the immune
system (Oers et al., 1989). Albumin, although homologous in structure,
shares no immunological properties with these proteins. The occurrence
of AFP in adult serum is usually associated with disease (Adinolfi et al.,
1975).
A high degree of conformational flexibility of albumin has been recog-
SERUM ALBUMIN STRUCTURE 155

nized for many years. Foster (1960) documented several isomeric forms
of albumin that can be induced reversibly as a function of pH (see Section
II,C,2,d). Although there has been speculation about the possible func-
tion of one of the transitions, the physiological significance of these
isomeric forms remains in question. Depsite the conformational adapt-
ability of albumin, it is not readily denatured and survives heat pasteur-
ization at temperatures of 60°C for 10 hr. without deleterious effects
(Shrake et al., 1984). In addition to potential conformational isomeriza-
tion, other forms of albumin microheterogeneity occur as well. In circu-
lating plasma approximately 30% of free sulfhydryl, Cys-34, is oxidized
by cysteine and glutathione (Peters, 1985). Additional forms of albumin
impurities arise from its high affinity for fatty acids, bilirubin, hematin,
and metals, such as nickel and copper. Moreover, microheterogeneity
can be introduced by the isolation procedures. For example, some al-
bumin preparations contain up to 20% of dimerized albumin. The per-
centage of this dimer increases with the age of the protein unless Cys-34
has been blocked with cysteine or glutathione. Albumin for which Cys-34
is free is referred to as “mercaptalbumin.” Furthermore, 7 to 10% of
normal human albumin in circulation is glucosylated and a much higher
percentage is observed for individuals with diabetes (Guthrow et al.,
1979).
Over the years there have been many outstanding reviews on albumin
(Hughes, 1954; Foster, 1960; Kragh-Hansen, 1981; Fehske et al., 1981;
Brown and Shockley, 1982; Peters, 1980, 1985, 1992; Rothschild et al.,
1988) and several other reviews that describe more specialized studies.
Because of the extensive nature of the literature regarding albumin, the
authors have relied heavily on these reviews to provide many of the key
works on albumin. They were essential both during the structure deter-
mination process and in the preparation of this review. It is the aim of this
review to correlate many of the findings of past significant papers with
the current knowledge of the recently determined X-ray structure (He
and Carter, 1992). Clearly, such an immense body of literature predating
the structure determination provides an opportunity and privilege af-
forded few crystallographers.

11. ALBUMIN
STRUCTURE

A. Primary Structure
In recent years there has been a flurry of albumin sequence informa-
tion, and now amino acid sequences have been deduced for several
albumins as well as other members of the multigene family (Table I). The
156 DANIEL C. CARTER AND JOSEPH X. HO

TABLE I
Determindion of Amino Acid Seguences

Protein and source Abbreviation Ref.

Albumin
Human HSA Behrens et al. (1975),Meloun
et al. (1975),Lawn et al.
(198l),Dugaiczyk et al.
(1982)
Bovine BSA Brown (1975),Holowachuk
(1991)
Equine ESA Ho et al. (1993)
Rat RSA Sargent etal. (1981)
Mouse MSA Minghetti et al. (1985)
Pig PSA Weinstock and Baldwin (1988)
Sheep OSA Brown et al. (1989)
Frog XSA Haefliger et al. (1989)
Salmon SSA Byrnes and Gannon (1990)
Lamprey LSA Gray and Doolittle (1992)
a-Fetoprotein
Human AFP Law and Dugaiczyk (1981),
Morinaga el al. (1983)
Rat - Jagodzinski et al. (1981)
Mouse - Gorinetal. (1981)
Vitamin D-binding protein
Human VBP Yang et al. (1985),Schoentgen
et al. (1986)
Rat - Cooke and David (1985)
Mouse - Yang et al. (1990)

sequences of all albumins are characterized by a unique arrangement of


disulfide double loops that repeat as a series of triplets. Mammalian
albumins have nine such loops formed by 17 disulfides. From observa-
tions of these loops and other homologies, albumin was deduced to be the
product of three homologus domains (I, 11, III), a result of some early
gene fusion event (Brown, 1977; McLachlan and Walker, 1977). Interest-
ingly, the only known exception to the three-domain structure is lamprey
albumin, which has seven domains. Brown, studying bovine serum al-
bumin (BSA), further divided these domains into three subdomains, A,
B, and C, after the three disulfide loops, and utilized this nomenclature to
discuss hypothetical structure and ligand-binding properties. Other in-
vestigators have preferred to associate ligand-binding activities with the
individual loops (Ll-Lg) (Peters, 1985) (Fig. 1). From examinations of
SERUM ALBUMIN STRUCTURE 157

11

12
DOMAIN1

13

14

15
DOMAIN I1

16

17

18

DOMAIN 111

19

FIG. 1. Illustration of the loop-link-loop pattern in the primary structure of serum


albumin. The three domains ( I , 11, and 111) are shown; loops are numbered 1 to 9 (LI-Lg).
TABLE I1
Aligned Albumin Sequences Illustrating Conserued Amino Acids and Invariant ResidueP

SEQUENCE HOMOUXSY: HOMAN/BOVINE/EQUINE/OVINE/RAT/FRW/SALMON


TABLE
I1 (continued)

Light shading indicates conserved amino acids; dark shading indicates invariant residues. The
approximate positions of the 28 helical segments of HSA are also indicated; *, deleted.

159
160 DANIEL C. CARTER AND JOSEPH X. HO

the internal sequence homology of BSA, McLachlan and Walker (1977)


correctly deduced that albumin is composed of six subdomains, which
were later verified by the crystal structure (Carter et al., 1989; Carter and
He, 1990; He and Carter, 1992). A modified nomenclature was adopted
to reflect the domain structure observed in the three-dimensional struc-
ture better, and it should not be confused with the earlier, similar nomen-
clature of Brown (see Section II,C,2,a). It should be also be noted that
throughout the remainder of this review all albumin sequences (see, e.g.,
Table 11) are normalized to that of HSA.
T h e amino acid sequence of albumin is further characterized by un-
usually high percentages of Cys (35%) and charged amino acids, and low
percentages of tryptophan, glycine, and methionine (Brown and Shock-
ley, 1982; Peters, 1985). The primary structure of albumin is unusual
among extracellular proteins in possessing a single free sulfhydryl (Cys-
34) and in having no sites for enzymatic glycosylation. However, this is
not the case for other albumin family members, such as a-fetoprotein
(Law and Dugaiczyk, 1981; Morinagaetal., 1983), VDP (Yangetal., 1985;
Schoentgen et al., 1986), and lamprey and frog albumins, which are
glycoproteins.
The sequence homologies are high for all albumins determined thus
far except for salmon, frog, and lamprey (Table 11). Still, all conserve the
characteristic repeating disulfide pattern. High percentages of sequence
identities have been noted between HSA and BSA (76%) (Peters, 1985),
and the recently determined amino acid sequence deduced from cDNA
for horse serum albumin (ESA), which reveals sequence identities be-
tween ESA and HSA, BSA, and RSA for 76, 73, and 76%, respectively
(Ho et al., 1993). HSA and human AFP exhibit 39% homology (Morinaga
et al., 1983). One of the major differences between AFP and serum
albumin is the absence of a double disulfide bridge located in L6. Simi-
larly, VDP lacks L9 in its entirety, and overall VDP proteins exhibit much
lower homology with albumin, suggesting an early divergence of this
gene. Homology among VDPs, however, is quite high (Gray and Doo-
little, 1992). Gray and Doolittle (1992) recently constructed a phy-
logenetic tree for the albumin superfamily based on their work with the
more distantly related lamprey serum albumin (LSA).From the sequence
they deduce that LSA, like all other members of the albumin family, is
composed of 190 amino acid repeats, suggesting, “that the invention of
albumin antedates the vertebrate radiation.” From these data and com-
parative sequence alignments with other members of the albumin multi-
gene family, they place the gene duplication event that gave rise to other
members of the albumin three-domain structure prior to the appearance
of vitamin D-binding protein, approximately 450 million years ago. It has
SERUM ALBUMIN STRUCTURE 161

been theorized that the 190-amino acid “protoalbumin” originated by an


incomplete gene fusion event involving some primordial globin or globin
precursor (Brown, 1976). A more detailed discussion of the structural
evolution of albumin is presented in Section IV.
Polymorphisms in the sequences of human albumin are quite rare, but
are relatively common in cattle, horses, and sheep (Tucker, 1968). T o
date, the positions of 43 single-site point mutations in HSA have been
determined, largely due to the efforts of F. W.Putnam, and these data
are presented in Table 111. T h e majority of the variants have been
identified by anomalous electrophoretic migration (Tarnoky, 1980) or, in
one case, by the unusually high affinity for thyroxin (Borst et al., 1983).
Therefore, typical commercial preparations of HSA produced from out-
dated, pooled blood can be considered essentially homogeneous in amino
acid sequence.

B . Early Concepts of Albumin Structure


Over the years, albumin has been probed with diverse experimental
methods, including hydrodynamics, low-angle X-ray scattering, fluo-
rescence energy transfer, electrophoretic methods, and NMR, IR, UV,
mass, and Raman spectroscopies. Virtually every measurable property
has been determined more than once (Table IV), with each investigator
seeking to establish important insight into albumin structure and chem-
istry.
Based largely on hydrodynamic experiments (Hughes, 1954; Squire et
al., 1968; Wright and Thompson, 1975) and low-angle X-ray scattering
(Bloomfield, 1966), the early conception of serum albumin began to
converge on the shape of an oblate ellipsoid having dimensions or axes of
140 x 40 A (Fig. 2). Experiments have continued to support these di-
mensions (Bendedouch and Chen, 1983; Feng et al., 1988). Brown and
Shockley ( 1982), with their demonstrated knowledge of albumin, con-
structed a “cigar-shaped” three-domain model of BSA that complied with
an incredibly diverse variety of data. In the absence of an X-ray structure,
this model has served as a frequently referenced conceptual model, albeit
an incorrect one, for the past decade. In fact, in the early stage of
crystallographic analysis of HSA at low resolution, due to the special
packing arrangement of the molecules in the crystal, it seemed to us that
the cigar-shaped model was plausible (Carter et al., 1989; Carter and He,
1990).
Despite the general acceptance of the cigar-shaped structure of serum
albumin (Fig. 2), several research groups provided evidence otherwise.
Electron microscopy indicated that BSA was a roughly spherical molecule
TABLE I11
Amino Acid Substitutions and Codon Chunges fm Albumin Cenctu Variant9

Minimum
Substitution Variant name Codon changeb change Ref.

-2 Arg+ His Lille, etc. TICST + CAT G+A Abdo et al. (1981). Arai et 01. (1990), Takahashi el al. (1987b)
- 2 Arg + Cys Malmo I, etc. TIGGT + I G T C+T Brennan el al. (1990)
- 1 Arg + Gln Christchurch, etc. TlCGA -+ CAM G-A Arai et al. (1990), Brennan and Carrel1 (1978)
- 1 Arg -+ Pro Takefu, etc. TlCGA + CCA G+C Arai etal. (1990), Takahashi et al. (1987b)
- 1 Arg+ Leu Jaffna TlCGA + C I A G-T Galliano et of. (1989)
1 Asp+ Val Iowa City-2 AlGAT -+ G P A-T Brennan et al. (1989)
Blenheim
& 3 His+Gln Nagasaki-3 CAGIA + C A a or CJ C + AlG Arai et 01. (1989b),
r4
(exonic (CIGT + AGMCIA) Takahashi et al. ( 1 9 8 7 ~ )

-
borders = I )
60 Glu + Lys Torino TIGAA + -MA G-A Galliano et al. ( 1990)
63 Asp + Asn Malmo (#95) TISAC -MC G+A Carlson ef af. (1992)
82 Glu -+ Lys Vibo Valentia TlGAA + -MA G+A Galliano et al. ( 1990)
114 Arg + Gly Yanomama ClGGA + -aA C+G Takahashi et al. ( 1 9 8 7 ~ )
119 Glu -+ Lys Nagoya AIGAG + -MG G+A Arai et al. (1990)
128 His + Arg Komagome-2 TICAT + CGT A+G Madison et al. ( 1991)
240 Lys + Glu Herborn Cl-MA -+ GAA A-+G Minchiotti el al. (1993)

--
268 Gln --* Arg Malmo (# 10) T1C-A + CGA A-G Carlson et al. ( 1 992)

-
269 Asp + Gly Nagasaki- 1 AlGAT + GGT A+G Arai el al. (1989b). Sugita et al. (1987), Takahashi et al. ( 1 9 8 7 ~ )

-
313 Lys+ Asn Tagliacozzo, etc. TlAAG A A U or C) G .--, TIC Galliano ef al. (1986, 1990)
318 Asn- Lys Malmo (#47) A A s l T A A U or CJ C AlG Carlson ef al. (1992)

-
320 Ala + T h r Redhill TlWT ACT G+A Brand et al. (1984). Hutchinson and Matejtschuk (1985),
Brennan et al. (1990b)
321 Glu + Lys Roma TIGAG AAG G-A Galliano et al. (1988, 1990)
---
333 Glu Lys
354 Glu Lys
Sondrio
Hiroshima-I
T&AA -
-
-MA
TIGAA- AAA
G- A
G-A
Minchiotti et al. (1992)
Arai et al. (1989b), Takahashi et al. ( 1 9 8 7 ~ )

- ---
358 Glu Lys Coari-I, etc. AIGAG -AG G- A Arai et al. (1989a)
365 Asp + His Parklands AIGAT + CAT G-C Brennan (1985)
365 Asp Val Iowa City-I AIGAT G T T A-T Madison el al. (1991)

-- --
372 Lys + Glu NaskapilMersin, etc. Cl-AA GAA LA- G Takahashi et al. (1987a)
375 Asp- Asn Nagasaki-2 CISAT -AT G- A Arai et al. (1989b), Takahashi et al. ( 1 9 8 7 ~ )

--
376 Glu Lys Tochigi, etc. TIGAA -MA G- A Arai et al. (1989b)
376 Glu Gln Malmo (#5, etc.) TIGAA CAA G+C Carlson et al. ( 1992)

--
382 Glu + Lys Hiroshima-2 GIGAA -MA G-A Arai et al. (1989b), Takahashi el al. ( 1 9 8 7 ~ )
479 Glu + Lys Dublin, etc. AlGAA -MA G- A Sakamoto et al. (1991)

- ---
494 Asp + Asn Casebrook CISAT -AT G- A Peach and Brennan (1991)
501 Glu --* Lys Vancouver, etc. AIGAG -MG G- A Huss et al. (1988)
505 Glu Lys Ortonovo TIGAA -MA G+ A Galliano et al. ( 1993)

- -
536 Lys + Glu Castel di Sangro CI-MG GAG A-G Minchiotti et 01. (1990)
541 Lys + Glu Maku A/-MA GAA A+G Takahashi et al. ( 1 9 8 7 ~ )

-
550 Asp Gly Mexico TIGAT GGT A+G Franklin et al. (1980)

---
550 Asp- Ala Malmo (#61 and 96) TIGAT + C&T A+C Carlson et al. ( 1992)
563 Asp Asn Fukuoka- 1, etc. CISAT + -&4T G-A Arai et al. ( 1990)
565 Glu + Lys Osaka-1 GIGAG -&4G G-A Arai el al. ( 1990)
570 Glu + Lys B-tYpe CIGAG -AG G-A Winter et al. (1972)
573 Lys + Glu Gent (MilFg) TI-MA GAA A+G ladarola et al. (1985)
574 Lys-, Asn Vanves AIAAB + AA(T or C
J A + TIC Minchiotti et al. (1987)

-
Positions of currently known single-site substitutions in human serum albumin. Provided in advance of publication as a courtesy by F. W. Putnam.
Except for proalbumin Malmo I, none of the above represents a C + T change (even in the absence of a CG), i.e., CG TG. Chain termination mutants
not listed above: Catania, Rugby Park, and Venezia.
’ Diagonal lines on either side of a codon are used to separate bases from a preceding or following codon. This facilitates identification of CpG
dinucleotides overlapping two codons.
164 DANIEL C. CARTER A N D JOSEPH X. HO

TABLE IV
Physical Meczsuremenls of Serum Albumin

Parameter Value Source Ref.

Diffusion coefficient &,,w 6.0 x lo-' dcm2/sec HSA Oncley el al. (1947)
Partial specific volume 0.733cmS/g HSA Charlwood (1961)
Sedimentation coefficient 4.5 x 10-13s HSA Oncley et al. (1947)

Intrinsic Viscosity 0.042 (N), 0.16 (F) HSA Charlwood (1961)


Specific absorbance 0.51 (HSA), 0.667 (BSA) Peters (1975)
(liters g-l cm-')
Molecular dimensions"
Equilater triangle
side 80 A
depth 30 A
Molecular volume" 88248.9 A3
Molecular surface area 28202.1 A2
Isoelectric point, pH -5.3 -4.7 (1-2 fatty acids)
(defatted)
Radius of gyration" 26.7 A

Calculated from the atomic coordinates of human serum albumin (He and Ca;ter.
1992).

with a height of 60 A and a depth of 45 8, (Slayter, 1965). Stokes radii of


BSA determined by diffusion data were approximately 37 8, (Longs-
worth, 1954). Perhaps the most interesting of these data are the low-
resolution dark-field electron micrographs of the homologous human
and bovine AFP (Fig. 3) (Luft and Lorscheider, 1983), which revealed
AFP as a U-shaped molecule having dimensions of approximately 80 8,.
However, at that time, any of its similarity in shape to albumin was
dismissed because AFP, unlike albumin, lacked a pair of disulfide bridges
at L6, which it was believed would allow the AFP molecule to fold in this

Serum Albumin
041-k 4 - A

T
'Oi
FIG. 2. Classical preception of the structure of serum albumin. Reproduced with per-
mission from Peters (1985) and Academic Press.
SERUM ALBUMIN STRUCTURE 165

A B

FIG. 3. Digitized matrices of mass-scattered electrons from single-average molecular


images of human a-fetoprotein (A) and bovine a-fetoprotein (B) with respective contoured
mass maps (C and D). Reproduced with permission from Luft and Lorsheider (1983).
OCopyright 1983 American Chemical Society.

manner. Other interesting conflicting data were presented by Hagag and


colleagues ( 1983).They predicted nearly equal distancesbetween Cys-34,
Trp-214, and Tyr-411 (-25 A) using a covalently attached fluorescent
probe on Cys-34 and fluorescence energy transfer methods, and con-
cluded that albumin must be folded more in accordance with a U-shaped
model, hereafter referred to as “heart-shaped.” Subsequent studies on
the N-B transition (see Section II,C,2,d) of HSA using ‘HNMR indi-
166 DANIEL C. CARTER AND JOSEPH X. HO

cated that an oblate elipsoid structure of HSA was unlikely; rather, it was
proposed that HSA was folded into the heart-shaped structure similar to
AFP (Bos et al., 1989). Overall, discrepancies with the cigar-shaped model
have largely been dismissed in previous albumin reviews (Peters, 1985,
1992; Kragh-Hansen, 1990; Brown and Shockley, 1982).

C . Three-Dimensional Structure

I . Crystalliuztion
Detailed structural knowledge of a protein molecule can only be pre-
cisely determined by X-ray crystallographic methods. A descriptive his-
tory of albumin crystallization is therefore presented, owing to its under-
lying importance to both the determination of the three-dimensional
structure and to the bulk purification of albumin. Crystallized preparations
of albumin were known as early as 1894 (Giirber, 1894), long before the
advent of X-ray crystallography. Much later McMeekin (1939) gave a
more detailed description of the preparation of the ESA crystals used by
Giirber, showing that preparations of albumin with improved homoge-
neity and low carbohydrate content could be produced. These hexagonal
crystals are shown in Fig. 4D (see color insert). Today, the majority of the
commercial albumin preparations owe their origin to the fractionation
methods developed by Cohn et al. (1947). These methods were aimed at
producing stable human blood substitutes for battlefield applications
using BSA purified by cold ethanol fractionation. Crystals of human
serum albumin could also be produced by this method by the addition of
cofactors such as decanol o r caprylic acid. Expounding on the same
techniques of Cohn, Lewin (195 1) published an extensive list of deri-
vatized human and bovine albumin crystals. Similarly, although origi-
nally noted in 1947, Hughes and Dinitz (1964) reported details of the
crystallization of a mercury dimer of HSA and discussed this technique as
a method for producing mercaptalbumin.
Selected forms of human albumin were first studied by Low and Wei-
chel (195 l ) , followed by more detailed studies using X-ray diffraction
methods (Low and Richards, 1954). Their work primarily focused on the
hydration of protein crystals and determination of various protein
properties such as molecular dimensions and weight. Although these
studies were conducted prior to the first successful structure determina-
tions for proteins (Perutz et al., 1960; Kendrew et al., 1960), they included
some of the first observations that protein crystals contain large percent-
ages of aqueous solvent. Modern crystallographic studies of albumins
were begun in the early 1970s and resulted in the preliminary character-
ization of several additional crystal forms, including the ESA crystals
SERUM ALBUMIN STRUCTURE 167

originally reported by Giirber, the HSA crystals of Kendall (Kendall,


1941; McClure and Craven, 1974), and one additional new crystal form
of HSA (Rao et al., 1976). These attempts to provide structural informa-
tion were unsuccessful, presumably due to the lack of reproducibility or
other undesirable properties of the crystals (Peters, 1985). Progress was
no doubt further impeded by the available technology of the time. It is
presumed that many of the crystal forms that can be produced from
various albumin preparations have gone unreported because of the very
poor diffraction qualities of these crystals. For example, ESA crystals
grown from solutions of polyethylene glycol, although quite large and
optically beautiful, exhibit very poor diffraction quality (Fig. 4L,see color
insert). In 1989 we reported a new crystal form of human serum albumin
that could be grown reproducibly and that eventually proved suitable for
structure determination (Carter et al., 1989). These crystals (Fig. 4A, see
color insert) are unusual in possessing large, continuous solvent channels
that have cross-sectional dimensions of approximately 100 A and an
overall solvent content of 78% (Fig. 5). A markedly better crystal form
was later produced from rHSA (D. C. Carter, B. Chang, K. Keeling, Z.
Krishnasami, and J. X. Ho, unpublished results; He and Carter, 1992)
and subsequently also by using wild-type HSA. Currently, work in our
laboratory has produced several new crystal forms and the observation of
improved quality for crystals grown in the microgravity environment
(Miller et al., 1992). Indeed, HSA was among the very first proteins
crystallized in space (DeLucas et al., 1989). Several additional high-quality
crystal forms are now grown routinely in our laboratory for structural
studies (Fig. 4, see color insert). These crystals include canine serum
albumin (CSA) (Fig. 4C and E), which exhibits diffraction to 2.0 A, and a
crystal produced from turkey albumin (TSA) (Fig. 4J), which diffracts to
1.7 A. In addition, several diffraction-quality crystal forms of HSA and
canine serum albumin (CSA) containing homogeneous preparations of
various long-chain fatty acids have now been realized. Crystallographic
data have been collected on each of the crystal forms shown in Fig. 4.
Most of these crystal structures have now been solved by the molecular
replacement method (Rossman and Blow, 1962) and are in the
refinement stage.

2. X-Ray Structure of Albumin


A discovery is like falling in love and reaching the top of a mountain after a hard
climb all in one, an ecstasy induced not by drugs but by the revelation of a face of
nature that no one has seen before and often turns out to be more subtle and
wonderful than anyone had imagined.

Max F. Perutz (1989), in “Is Science Necessary”


168 DANIEL C. CARTER AND JOSEPH X. HO

FIG.5. The arrangement of the electron density in a tetragonal crystal of human serum
albumin. Prominent features of the molecular packing arrangement are large (90 x 90 A)
solvent channels (shown in white) that pass through the crystal parallel to the crystallo-
graphic c axis. The unit cell and symmetry operations parallel to the c axis are illustrated.
Reproduced with permission from Carter et al. (1989); 0 American Association for the
Advancement of Science (AAAS).

Although the structure determination of albumin was far from the


thrilling experience shared by Max Perutz and John Kendrew when
observing for the first time the atomic configuration of a protein mole-
cule so eloquently described above by Perutz (Perutz, 1989), the albumin
structure was quite different both in outline and detail, and clearly “more
subtle and wonderful than anyone had imagined.” It is of interest to note
that the first glimpses of albumin structure (Fig. 6, see color insert)
marked exactly 150 years of albumin research. To date there have been
several albumin structures determined by crystallographic methods in
our laboratory, including human wild-type (HSA), recombinant human
(rHSA), equine, and canine albumins. Currently, the most detailed infor-
SERUM ALBUMIN STRUCTURE 169

mation is available for human and equine serum albumins, therefore


much of the following discussion will be limited to these.
Rather than the prolate ellipsoid predicted by so many studies, as
discussed previously (Section II,B), the crystal structure of albumin re-
veals a heart-shaped molecule that can be approximated to an equilat-
eral triangle with sides of -80 8, and a depth of -30 8, (Fig.7, see color
insert). This observation is virtually identical with the low-resolution
dark-field micrographs of the 39% homologous human and bovine
a-fetoprotein (Fig. 3) (Luft and Lorschieder, 1983), and is in agreement
with the spectroscopically determined distances between Cys-34, Trp-
214, and Tyr-411 (Hagag et al., 1983). Thus albumin, under neutral
conditions of pH, has an axial ratio of approximately 2.66, which agrees
well with the value of 3.0 based on observations of dielectric and bire-
fringence relation times (Moser et al., 1966). The solvent-accessible sur-
face and molecular volume of HSA are 28,202 A' and 88,249 AS,respec-
tively, based on the calculations using the molecular coordinates and the
algorithm of Richards (1985). Similarly, the radius of gyration is 26.7 A,
which compares well with the hydrodynamic radius of HSA of 26.4 8,
(rotational) as measured by light scattering and electron spin resonance
(ESR) (Cannistraro and Sacchetti, 1986). As proposed by spectroscopic
(Jacobsen, 1972; Sjoholm and Ljunstedt, 1973; Chen and Lord, 1976)
and predictive methods (McLachlan and Walker, 1977), the albumin
structure is predominantly a-helical. Approximately, 67% of HSA is
helical, with the remaining polypeptides occurring in turns and extended
or flexible regions between subdomains (Fig. 7). The overall agreement
between the percentages of observed and predicted secondary structure
is quite remarkable. For example, Pearson (1990) predicted a 65% a-
helix content for serum albumin based on sequence comparisons of several
albumins and further estimated a P-sheet and p-turn percentages of 10%
and 19%,respectively. Although there is no /3 sheet in the structure of
serum albumin, -23% of the structure is in an extended chain con-
formation, which would be predicted as p strand, and -10% of the
remaining structure exists in turns. This adds further supporting evi-
dence for the greater accuracy of secondary structure predictions that
utilize reliably aligned multiple sequence data (Zvelebil et al., 1987).

a. Domain Structure. As noted in the discussion of the primary struc-


ture of serum albumin, the three-dimensional configuration is composed
of three homologous domains (I, 11, and HI). Each domain in turn is
the product of two subdomains (IA, IB, etc.), which are predominantly
helical and extensively cross-linked by several disulfide bridges (Fig. 8,
170 DANIEL C. CARTER AND JOSEPH X. HO

see color insert). Furthermore, the six subdomains share a common


helical motif (Fig. 8). This motif primarily corresponds to the amino acids
encompassed within the double disulfide loops 1, 3, 4 , 6 , 7, and 9. Each
motif is related by a pseudo twofold axis (I: 168", 11: 163", 111: 171") shown
in Fig. 8. These data are consistent with the theory that the 190-amino
acid protoalbumin was the product of an early globin gene fusion event
that predates the occurrence of LSA. The association of protein domains
by pseudo twofold symmetry has been observed in the crystal structures
of many proteins that are products of tandem gene fusion. When the
a-carbons of the individual domains are superimposed (HSA), they give
an average root-mean-square (RMS) difference between a-carbons of
3.77,4.32, and 3.63 8, for domains 1-11, 1-111, and 11-111, respectively.
Cross-species comparison between HSA and ESA gives an average RMS
difference in a-carbons of 2.01 8, for the entire molecule, which indicates
the highly conserved nature of the three-dimensional structure. Ideally,
one can divide each of the domains into 10 helical segments, hl-h6 for
subdomain A and h7-hlO for subdomain B. The helical nomenclature
presented in Fig. 8 provides an important framework for subsequent
comparisons between the homologous domains, related structures, and
discussions involving con formational change.
Although all subdomains share a common four-helix motif, there are
distinct differences. T h e A subdomains supplement the three-helix bun-
dle on the C-terminal side with an additional but smaller disulfide double
loop (loops 2, 5 , and 8) to form a small globinlike structure that is
extensively cross-linked by four disulfide bridges (Fig. 9A, see color
insert). The B subdomains supplement the helical motif on the N-
terminal side with a conformationally extended polypeptide to create a
folding topology that closely resembles a simple up-down helical bundle
(Fig. 9B). This section of polypeptide represents the evolutionary loss of
homologous helix h 1, although a helical remnant remains as part of the
disulfide bridge. The A and B subdomains assemble through hydropho-
bic helix packing interactions primarily involving h2, h3, and h8. In
addition, the subdomains are linked together by a presumed flexible
extension of polypeptide encompassing residues Lys- 106 to Glu- 119,
Glu-292 to Val-315, and Glu-492 to Ala-511 in the three domains, I, 11,
and 111, respectively. These linkages do not appear crucial for the confor-
mational stability at neutral pH and show the greatest differences in
conformation when comparing related three-dimensional structures.
Domains I and I1 and domains I1 and 111 in turn are connected through
helical extensions of hlO(1)-hl(I1) and hlO(I1)-hl(III), creating the two
longest helices in HSA. Consequently, the actual number of helices in the
structure is 28 rather than 30. The positional occurrence of these helices
SERUM ALBUMIN STRUCTURE 171

in the primary structure of HSA is presented in Table 11. Because this


helical extension restricts the potential packing arrangement between
domains, they do not associate by simple twofold rotation. Instead, quite
surprisingly, they appear to be related by a pseudo twofold screw-axis
symmetry with a rotation angle of 165" for 1-11 and 167" for 11-111.

b. Nature of DzSulfides in Albumin. Disulfide pairings in albumin occur


as predicted by Brown (1975) based on amino acid sequences of proteo-
lytic fragments and steric considerations. In further agreement with
Raman spectroscopic studies (Akoi et al., 1973), the conformations of the
disulfides are primarily gauche-gauche-gauche and C/~~-SI-SP-CPO,
with torsion angles clustering around +80". The disulfide pairings in
HSA are located almost exclusively between helical segments. Such disul-
fide pairings are rarely observed in protein structure and usually involve
isolated occurrences, e.g., crambin (Teeter and Hendrickson, 1979). Al-
bumin is the first observed case wherein an entire protein folding topol-
ogy is based on this theme. Given this interesting feature, one is better
able to appreciate the stability of albumin under a variety of experimen-
tally harsh conditions. For example, albumin can be heated to 60°C in the
presence of caprylic acid, for 10 hr without deleterious effects (Shrake et
al., 1984). The helical motif of albumin undoubtedly provides a natural
framework that restricts mixed disulfide pairing.
Katchalski et al. (1957) concluded that the disulfides in albumin were
protected at neutral pH from reducing agents. This is also apparent in
the structure, which shows that the majority of disulfides are well pro-
tected and most are not readily accessible to the solvent.
It is well known that blocking of the free sulfhydryl, Cys-34, with
iodoacetamide, cysteine, or glutathione prevents the occurrence of
mixed disulfides in aged albumin, as well as the occurrence of the al-
bumin dimer (Peters, 1985). In the structure, Cys-34 is not in close
proximity to any disulfide, consequently, it is difficult to imagine its
intramolecular participation without substantial conformational change
of the albumin molecule. However, it could perhaps participate in the
formation of mixed disulfides by catalyzing the reaction via intermolecu-
lar interaction, because Cys-34 is known to have an unusually low p K s ~
(Pedersen and Jacobsen, 1980) (see Section I K C , 1).

c. Surface Charge Distribution. As a highly soluble protein that can be


prepared at concentrations as high as 30% (w/v),albumin is ideally suited
for its role as the major plasma protein. This amazing property is no
doubt related to its high negative charge at neutral pH. In this regard,
albumin is known to have an interesting asymmetric charge distribution
172 DANIEL C. CARTER A N D JOSEPH X. HO

within the primary structure. At neutral pH, Peters (1985) calculated a


net charge of - 10, -8, and 0 for domains I, 11, and 111 of HSA, respec-
tively, and similar distributions are present in other albumins. The sur-
face charge distribution of human serum albumin is shown in Fig. 10 (see
color insert). Unlike the asymmetric charge distribution expected based
on the primary structure, the distribution seems fairly uniform. There
does not appear to be any significant pattern of basic versus acidic resi-
dues, except for a group of invariant basic amino acids on the surface of
domain I (see Section IV). There are, however, distinct areas of amino
acids with neutral charge that could potentially be important for interac-
tions with long-chain fatty acids.
There have been 43 point mutations identified for human serum
albumin, as shown in Table 111. These variants were usually discovered
because of their anomalous electrophoretic migration (see, for example,
Tarnoky, 1980). The specific point mutations have been determined
primarily through the efforts of F. W. Putnam and co-workers. With the
exception of the point mutation Ala-320 to Thr, which is partially inter-
nalized in the structure to pack with other lipophilic residues, all of the
remaining substitutions in Table 111 are located on the surface of HSA
and are exposed to solvent.

d. Conformational Flexibility. The ability of albumin to undergo a ma-


jor reversible conformational isomerization with changes in pH was first
demonstrated by Luetscher in 1939. Foster, noting these early studies
involving electrophoretic heterogeneity, reinvestigated this fascinating
phenomenon in much greater detail (Foster, 1960, 1977). The only other
known major conformational changes that occur in albumin are induced
by the interaction of albumin with fatty acids (Peters, 1985).
Foster classified the pH-dependent forms as “N,” for normal form,
which is predominant at neutral pH; “B,” for the basic form occurring
above pH 8.0; “F,” for fast migrating form produced abruptly at pH
values less than 4.0; “E,” for expanded form at pH less than 3.5; and “A,”
for aged form occurring with time at pH values greater than 8.0. Perhaps
the most interesting of these isomers are the B and F pH transitions, and
fatty acid-induced, conformational isomers.
i. N-F transition. As early as 1960, Foster clearly recognized that the
formation of the F form involved an abrupt opening of the molecule.
Much later Geisow and Beaven (1977) proposed that the N-F transition
involved the unfolding of domain 111 from the rest of the molecule, and
this was later verified by Khan (1986) using proteolytic fragments of BSA
encompassing residues 377-582, the so-called “T-A fragment” (Peters
and Feldhoff, 1975). It had been previously shown that no conforma-
SERUM ALBUMIN STRUCTURE 173

tional changes occur during the N-F transition for fragments containing
domains I and I1 (Khan and Salahuddin, 1984). Recent experiments in
our laboratory using proteolytic fragments of BSA and turkey serum
albumin (TSA) confirm the conclusions of previous studies and further
illuminate the nature or “structure” of the N-F transition. Using limited
peptic cleavage at residue 306 to produce the two fragments, A and B,
which constitute the two halves of albumin (King, 1973),we undertook to
test our hypothesis that during the N-F transition the albumin molecule
expands by the dissociation of the C-terminal half, or “tail,” from the
“head” of albumin previously described in Section II,C,2. Figure 11
illustrates the retention times on fast protein liquid chromatography
(FPLC)columns of pepsin-treated BSA (BSAp) and TSA (TSAp) under a
variety of pH conditions. As can be seen, BSAp elutes with a retention
time identical to that of BSA until pH 4.0, where both TSA and BSA
abruptly split into two fragments of equal size. This coincides exactly
with the pH known to induce the F conformational state of albumin
(D. C. Carter, P. D. Twigg, and K. Keeling, 1993). Furthermore, as
demonstrated earlier by King (1973), these fragments recombine when
the pH is returned to neutral. These findings also provide additional
supporting evidence for the predominance of the heart-shaped con-
formation in solution in the pH range from 4.5 to 8.0, because the
fragments never dissociate during column chromatography at pH
greater than 4.0. Additionally, it is interesting to note that the F transition
is conserved even in distantly related species such as birds (Fig. 11)
(Feldhoff and Ledden, 1983).T h e change in hydrodynamic ratio indicat-
ing an increase in volume for BSA of 1 1% is consistent with this observa-
tion (Leob and Scheraga, 1956), as are the dimensions of roughly
40 X 129 8, predicted by Victor Bloomfield (1966) for BSA at low pH
(3.5) deduced from low-angle X-ray scattering experiments. The F form
is further characterized by a dramatic increase in viscosity, much lower
solubility, and a significant loss in helical content (Foster, 1960). T h e
proposed structures of the F and E conformational isomers are illus-
trated in Fig. 12.
Structurally, the interface between the two halves of the molecule is
held together by both hydrophobic and salt bridge interactions. Among
these residues there are potential salt bridges between Lys-205 and Glu-
465, Asp- 187 and both Lys-432 and Arg-52 1, Arg-2 18 and Asp-45 1, and
Lys-190 and Glu-425, which should be more clearly defined in a higher
resolution crystal structure. Hydrophobic interactions that associated IA,
IB, and IIA, with IIB, IIIA, and IIIB include a major interdomain
cluster involving Phe-206, Leu-48 1, Ala-482, Trp-2 14, Leu-347, Val-344,
Val-343, Leu-33 1, Ala-2 17, and Tyr-452.
174 DANIEL C. CARTER A N D JOSEPH X. HO

BSA tSA

IL,
Elution uutkn
Vdunn -vdunw,
1 13.34 ml 1 13.10ml
2 14.65 ml 2 14.6 ml

1( 13.15ml 1 12.95 ml

mlml
2 14.38ml 2 14.32 mi

1 13.01 ml
2 14.24ml

1 21.03ml 1 20.7ml
2 23.68ml 2 23.59ml

0 10 20 30 0 1 0 2 0 3 0
Elution Volume, mls Elution Volume, ml
FIG.11. Chart of retention times for bovine serum albumin (BSA) and turkey serum
albumin (TSA), which have been proteolitically treated with pepsin. For pH values 7.0-5.0,
peaks 1 and 2 represent the albumin dimer and the proteolitically nicked monomer,
respectively, at different pHs (N form). At pH 4.0 there is an abrupt dissociation of the two
halves of albumin to yield peaks 1 and 2, which represent the intact monomer (incomplete
digestion, F form) and the two equal molecular weight halves of albumin, respectively.
(D. C. Carter, P. D. Twigg, and K. Keeling, Gnpublished results).

It is interesting to speculate whether the F conformation is of physio-


logical significance given the lower pH measured on the membrane
surfaces of several tissues (Wilting et al., 1982).Additionally, it is possible
that this is the form that is bound to the surfaces of highly oriented
pyrolytic graphite as revealed by a scanning tunneling microscopy study
SERUM ALBUMIN STRUCTURE 175

FIG. 12. Ribbon diagrams of serum albumin in its N form, and in its proposed F and E
forms. Figure drawn using program RIBBONS (Carson, 1987).
176 DANIEL C. CARTER A N D JOSEPH X. HO

by Feng et al. (1988), although it was assumed in their paper to be the


normal conformation. Perhaps this conformation can also be induced by
the surface binding process. T h e conservation of the N-F transition and
the p H of the transition among various diverse species clearly suggest a
physiological role for this transformation. One could speculate that the
affinity for ligands is reduced in the F conformation, perhaps facilitating
ligand off-loading at various tissue interfaces. T h e observation that the
sulfhydryl becomes more solvent exposed under both N-F and N-B
transitions would also be consistent with this supposition.
ii. F-E transition.At pH values lower than 4.0 albumin undergoes
another expansion, and electron microscopy reveals the molecule as a
series of balls and strings with approximate dimensions of 21 x 250 8,
(Harrington et al., 1956). It seems quite reasonable to expect that this
isomeric form corresponds to the sequential subdomain-subdomain dis-
sociation concomitant with the loss of the intradomain helices h lO(1)-
hl(I1) and hlO(I1)-hl(II1) (Fig. 12).
iii. N-B transition.At pH 9.0 albumin undergoes another conforma-
tional change, albeit a less dramatic one, the basic form (B). This form
was originally identified by its slower anodic migration during electro-
phoresis. Structurally, the conformational change occurs more slowly
than the F transition and there is evidence that discrete steps are associ-
ated with this transition (Hart et al., 1986). There is a decrease in helical
content and increase in affinity for selected ligands (Zurakowski and
Foster, 1974). At the present time, little structural information is known
about this isomer.

111. NATURE BINDING


OF LIGAND

Th e most prolific area of albumin research has involved binding stud-


ies. An appreciation of the extraordinary binding chemistry of albumin
can be obtained by examining a small subset of the binding studies
performed to date (Table V). Most ligands are bound reversibly, and
typical association constants (K,)range from lo4 to lo6 M - ' . Because of
the incredible diversity of ligands bound by albumin, early researchers
saw ligand binding to serum albumin as nonspecific in nature and did not
recognize that there were discrete sites per se. Instead they envisaged the
ligands as randomly attached to the surface, somewhat like a sponge.
This rather uninteresting view of albumin has changed over the past
years, and now it is generally recognized that there are a small number of
distinct binding locations. These binding sites have been studied pri-
marily by equilibrium dialysis or spectroscopic methods. Data from such
studies are generally analyzed by the Scatchard method (1949) and this
Color Plates
FIG.4. Crystals of serum albumin. (A) Tetragonal crystals of HSA, (B) monoclinic crystals of
baboon serum albumin, (C) triclinic crystals of CSA, (D) hexagonal crystals of ESA, (E)
monoclinic crystals of rHSA complexed with lauric acid, (F) monoclinic crystals of HSA
complexed with cysteine, (C) monoclinic crystals of rHSA, form 2, ( H ) crystals of feline serum
albumin, (I) monoclinic crystals of rHSA, form 1, u) orthorhombic crystal of TSA, (K)
tetragonal crystals of BSA, and (L) hexagonal crystals of ESA.

FIG.6. Section of the 6.0-A electron density of subdomain IIIB. Helical rods of density 8.0 to
10 A in diameter, indicative of the a-helical structure of serum albumin, were the dominant
features of the electron density. Reproduced with permission from Carter el al. (1989) and the
American Association for the Advancement of Science (AAAS).

FIG.7. Stereo view of human serum albumin illustrating the overall topology and secondary
structure. The positions of the 17 disulfides and the side chain of Cys-34 are shown in red.
Structurally, HSA consists of 28 helices, which range in size from 5 to 31 amino acids in length
and which can be grouped into 10 principal helices within each domain. The positions of the
two major binding sites of HSA, located within subdomains IIA and IIIA, are illustrated with the
ligand 2,3,5triiodobenzoic acid, shown in white. Figure drawn using program RIBBONS
(Carson, 1987).

FIG.8. Cylindrical representation of a typical domain (11) structure. Note that helices h l ,
h2, h3, and h4 of subdomain A are related to helices h7, h8, h9, and h10 of subdomain B by
a pseudodiad (pointing toward the reader). Figure drawn using program RIBBONS (Carson,
1987).

FIG.9. Stereo view of the IIA (A) and IIB (B) subdomains. Figure drawn using program
RIBBONS (Carson, 1987).

FIG.10. Space-filling model of serum albumin molecule with basic residues colored in blue,
acidic residues in red, and neutral ones in yellow. (A) Front view, (B) back view, (C) left side,
and (D) right side. Figure drawn using program RIBBONS (Carson, 1987).
FIG.13. Stereo view showing the homology between the triiodobenzoic acid-binding sites
within subdomain IIA (A) and subdomain IIIA (B); the subdomains have been superimposed
by the method of least squares. The protein components are shown in orange and green for ESA
and HSA, respectively, and the positions of bound triiodobenzoic acid are shown in red and
yellow for ESA and HSA, respectively. It should be noted that many of the side-chain conforma-
tions illustrated in Fig. 6 can be determined with confidence only at higher resolution. In
addition, the complex has not been refined. Reproduced with permission from Ho el al. (1993).

FIG.14. Dotted surface diagram showing the major binding pocket inside subdomain IIIA
(yellow ribbons). Figure drawn using program RIBBONS (Carson, 1987).

Fiti. 15. Electron density of bound 2,3,5-triiodobenzoic acid (TIB) at 3.0 A produced from
(FI--FN)a,where F,and F, refer to measured diffraction data from the TIB cocrystal complex and
native data, respectively, and the phases, a,are from the native model. Atoms: red, oxygen;
yellow, carbon; orange, iodine. Yellow and blue contour levels at 2.0 ci and 4.0 0, respectively.
Reproduced with permission from Ho el aL (1993).

FIG.16. Conformational difference between native rHSA (cyan) and rHSA bound with fatty
acid-lauric acid (yellow). Note that the largest displacement occurs at the two terminus
subdomains, i.e., subdomains IA and IIIB. Figure drawn using program RIBBONS (Carson,
1987;J. X. Ho, D. Carter el al., unpublished results).

FIG 17. Stereo ball-stick model of serum albumin structure at the region around residue
Cys-34. Red, oxygen; yellow, carbon; blue, nitrogen; green, sulfur. Figure drawn using program
Turbo FRODO (Cambillaux and Horjales, 1987).

FIG.19. Ribbon diagram showing helices h l , h2, h3, and h4 of subdomain IIIA of rHSA
(cyan) superimposed on helices E, F, H, and G of myoglobin (yellow). Note that the polarity of
h3 and h4 is opposite to that of G and H, although the helices are packed similarly; N, the two
N termini. Figure drawn using program RIBBONS (Carson, 1987).
FIG.6
FIG.7

FIG.8
\ .,

FIG.9A

FIG.9B
FIG. 10A (top) & B (bottom!
FIG.1OC (top) 8c D (bottom)
FIG.13A

FIG.13B
FIG.14

FIG.15
FIG.16

FIG.17

FIG.19
TABLE V
Seleckd Binding Conrtuntsfor Endogenous and Exogenous Ligandr with Serum Albumina

Ligand K , , W-') K,, Of-') Albumid Ref.


~~ ~ ~

Endogenous Substances
Aldosterone 3.2 x 105 - HSAf Richardson et al. (1977)
Arachidonate 3x 107 - HSA Saw et al. (1981)
Bilirubin 1.4 X 10' 5 x 105 HSAf Jacobsen (1969)
Bilirubin 5.5 x 107 4.4 x 106 HSAd Brodersen (1979)
Bilirubin 5x 107 - HSA Jacobsen (1977)
Chenodeoxycholate 2x 105 - HSA Roda et al. (1982)
Cholate 3.2 x 104 2 x 103 HSA~ Burke etal. (1971)
Cholate 5.5 x 104 - HSA Roda et al. (1982)
Cortisol 5.0 x 103 - HSA~ Yates and Urguhart (1962)
Estradiol 1.0 x 105 - HSAf Daughaday (1959)
Hemin 5.0 x 107 - HSA~ Beaven et al. (1974)
Hemin 1.1 x 108 - HSA~ Adams and Berman (1980)
Hematin 1.1 x 108 - HSA Adam and Berman (1980)
Linoleate 1.3 x 107 2.5 X lo6 HSAd Goodman (1958)
Linoleate 7.9 x 107 - HSA Spector and Fletcher (1978)
Lithocholate 9x 107 - HSA Roda et al. (1982)
Lysolecithin 4.3 x 104 - BSAf Klopfenstein (1969a,b)
Oleate 1.1 x 108 4.0 X lo6 HSAd Goodman (1958)
Oleate 26.0 x 107 - HSA Spector and Fletcher (1978)
Oleate 9.4 x 107 - HSA Spector and Fletcher (1978)
Oleate 2.9 x 107 - HSA Spector and Fletcher (1978)
Oleate 2.1 x 107 - HSA Spector and Fletcher (1978)
Palmitate 6.0 x 107 3.0 X lo6 HSAd Goodman (1958)
Palmitate 6.2 x 107 - HSA Spector and Fletcher (1978)
Progesterone 3.6 x 105 6 X lo3 HSAd Westphal and Harding (1973)
Prostaglandin El 7x 104 - HSA Unger (1972)
(continued)
TABLE
V (continued)

Ligand K,, W-') Ka, w-') Albuminb Ref.


Stearate 15.0 x 107 - HSA Spector and Fletcher (1978)
Taurocholate 1.2 x 104 - BSAd Green et al. (1971)
Testosterone 2.38 x 104 - HSA~ Vermeulen and Verdonck
( 1968)
Testosterone 4.2 x 104 - HSA Pearlman and Crepy (1967)
L-Thyroxine 1.6 X lo6 6 X lo4 HSAd Steiner et al. (1966)
L-Tryptophan 1.6 x 104 - HSAf McMenamy and Oncley
( 1958)
L-Tryptophan 6.3 x 104 - HSA Sollene et al. (1981)
Negatively charged and electrostatic neutral drugs
Acenocoumarin 1.96 x 105 6.0 x 103 HSA~ Tillement et al. (1974)
Camptothecin 8.0 x lo6 - HSA~ Chignell(l973)
=; Chlorazepate 1.3 x 104 - HSAf Coassolo et al. (1978)
Chlorophenoxyisobutyrate 1.3 x 105 1.5 x 103 HSA~ Spector et al. (1973)
Chlorophenoxyisobutyrate 3.3 x 105 3.9 x 103 HSA~ Tillement et ~ l(1974)
.
Chlorothiizide 3.07 x 104 - HSAf Breckenndge and Rosen
(1971)
Chlorpropamide 4.5 x 104 1.7 x 10' HSA~ Crooks and Brown (1974)
Cinchophen 1.4 x 105 - HSAf Mudge et al. (1978b)
Clofibrate 2.5 x 104 4.7 x 102 HSAf Nazareth et ~ l (1974)
.
Dansylglycine 4.6 x 105 - HSAf Chignell(l969)
Diazepam 4.9 x 105 - HSAf Miiller and Wollert (1973)
Dicoumarol 2.9 X lo6 1.8 x 105 HSAf Perrin et ~ l (1975)
.
Dicoumarol 2.2 x lo6 1.3 x 104 HSA~ Garten and Wosilait (1972)
Dicoumarol 7.7 x 105 - HSAf Chignell(l970)
Dicoumarol 3.5 x 105 - HSA~ Choetal. (1971)
Digitoxin 4.3 x 104 - HSAf Brock (1975)
Digitoxin 6.9 x 104 - HSAf Lukas and De Martino (1969)
Furosemide 1.7 x 105 9.6 X lo3 HSAf Sebille et al. (1978)
Fusidic acid 7.8 x 104 Guttler et al. (1971)
Glibenclamide 7.7 x 105 - Crooks and Brown (1974)
Halofenate 1.6 x 105 7.0 x 103 Spector et al. (1973)
Ibuprofen 2.73 X lo6 1.95 x 104 Whitlam et al. (1979)
Indomethacin 1.0 x 106 1.0 x 105 Hultmark et al. (1975)
Indomethacin 3.0 x 105 1.4 x 104 Mason and McQueen (1974)
Iopanoate 6.7 X lo6 1.5 X lo6 Mudge et al. (1978a)
Iopanoate 7x 105 9x 102 Lang and Lasser (1967)
Iophenoxate 7.7 x 107 3.8 x 105 Mudge et al. (1978a)
Novobiocin 5.5 x 105 3.7 x 104 Brand and Toribara (1975)
Phenylbutazone 2.37 X 1V 4.56 x 104 Rosen (1970)
Phenylbutazone 1 x 105 4x 104 Chignell(l969)
Phenylbutazone 2.5 x 105 1.3 x 103 Brown and Crooks (1976)
Phen ylbutazone 2.3 x 105 5.6 x 103 Tillement et al. ( 1974)
Salicylate 7.1 x 104 3.3 x 103 Keresztes-Nagy et al. (1972)
Salicylate 2.2 x 105 1.6 x 103 Brown and Crooks (1976)
Salicylate 1.3 x 2.9 x
c(
4 105 103 Hultmark et al. (1975)
(D
Sulfaethidole 1.5 x 105 1.6 X los Janssen and Nelen (1979)
Sulfaphenazole 9.2 x 104 1.1 x 103 Brown and Crooks (1976)
Tolazamide 8.7 x 104 1.5 x 105 Crooks and Brown (1974)
Tolbutamide 2.2 x 105 1.7 X lo2 Brown and Crooks (1976).
Crooks and Brown (1974)
Warfarin 2.5 x 105 1.1 x 104 Sudlow et al. (1975)
Warfarin 8.9 x 104 6.7 x 105 Garten and Wosilait (1972)
Warfarin 1.5 x 105 1.5 x 103 Brown and Crooks (1976)
Warfarin 2.3 x 1P 5.9 x 105 Tillement et al. (1974)
Positively charged drugs
Chlorpromazine 4.2 x 104 HSA~ Gabay and Huang (1974),
Huang and Gabay (1974)
Chlorpromazine 1.9 x 105 Sharples (1975)
Desipramine 7.02 x 104 Sharples (1975)
Imipramine 2.39 x 104 Shardes (1975)
(continued)
V (continued)
TABLE

Ligand K,, W - ' ) Ka* of-') Albuminb Ref.


Lidocaine 1.3 x 105 - HSAf Sawinski and Rapp (1963)
Mepivacaine 2.5 x 105 - HSA~ Sawinski and Rapp (1963)
Pamaquine 6.4 x 107 - BSAf Naik et al. (1975)
Procaine 3.1 x 103 - HSAf Sawinski and Rapp (1963)
Promazine 8.5 x 104 - HSA~ Sharples (1975)
Quinidine 1.4 x 103 - HSAf Nilsen and Jacobsen (1976)
Quinine 7.5 x 103 - HSAf Paubel and Niviere (1974)
Tnflupromazine 5.5 x 104 - HSAd Huang and Gabay (1974)
Inorganic ions
c1- 7.2 X lo2 6.1 x 10' HSAI Scatchard and Yap (1964)
1- 6.15 x 103 6.7 X 10' HSAf Scatchard and Yap (1964)
SCN- 3.35 x 104 7.8 x lo2 HSAf Scatchard and Yap (1964)
Zn2+ 5.7 x lo2 - HSAf Waldmann-Meyer (1960)
Cd2+ 1.3 x 103 - BSA~ Waldmann-Meyer (1960)
Mn2+ 2.4 x 104 5.0 x 10' HSA~ Nandedkar el al. (1973)
cu2+ 9x 106 - BSAf Peters (1975)
cu2+ 1.6 X 10l6 - HSA Lau, el al. (1974)
Ni2+ 3x 105 - HSA~ Callan and Sunderman (1973)
co2+ 6.5 x 103 1.6 x lo2 HSAf Nandedkar et al. (1972)
Ca2 + 1o2 - HSAI Pedersen (1971)
Ca2+ 9 x 102 - HSA Fogh-Andersen (1977)
Ca2+ 8.3 X 10' - HSA Fogh-Andersen (1977)
Mi?+ 1o2 - HSAf Pedersen (1971, 1972)

a Adapted and modified from Kragh-Hansen (1981).


f, Serum albumin with fatty acid; d, defatted serum albumin.
SERUM ALBUMIN STRUCTURE 181

field of research has been well reviewed by Lindup (1987). Although


controversy remains about the exact number of discrete binding locations
on albumin, the general consensus is as follows: there are two principal
binding areas for small heterocyclic or aromatic carboxylic acids; there
are at least two to three dominant long-chain fatty acid-binding sites
unique and separate from the binding sites for small anionic compounds
(at normal physiological concentrations); and there are two distinct
metal-binding sites, one involving Cys-34 and the other the N terminus.
Thus, with normal ligand/albumin concentrations, there are six domi-
nan: lreas of ligand association to albumin.
New perspectives on several of these binding sites were gained from
the crystal structures of albumin. The sites identified for the small hetero-
cyclic and aromatic carboxylic acids were found to reside within spe-
cialized cavities in subdomains IIA and IIIA. These sites are consistent
with Site I and Site 11 proposed by Sudlow et al. (1976). T h e two metal-
binding sites have been previously identified by various spectroscopic and
chemical means. The location of bound Cu(I1) and Ni(I1) involves the
first three residues of the N terminus, with a His at position 3 playing a
critical role (Peters and Blumenstock, 1967; Lau et al., 1974). T h e second
metal-binding site involves the free sulfhydryl of serum albumin, Cys-34.
Here, the amino acid cysteine and peptide glutathione are covalently
bound, as well as various metals such as Cd, Au, Hg, and Ag. The
locations of the two long-chain fatty acid sites remain more obscure,
although Brown and Shockley (1982) have proposed that long-chain fatty
acids occupy separate sites on the B subdomains. Crystallographic analy-
ses of a complement of fatty acid analogs bound to albumin are in
progress and answers to these questions should be forthcoming. The
bilirubin site is often considered distinct from the other ligands.
However, we believe that bilirubin is primarily bound to the site within
IIA, and this does not contradict the experimental data in the literature.
The following segments of this section describe in detail the current
understanding of each of the six binding sites.

A. Small Organic Compounds: Sites I and II


The vast majority of ligands in Table V are bound in one o r both sites
within specialized cavities of subdomains IIA and IIIA. At this time the
binding locations of several key compounds, historically used as markers
in drug or ligand displacement interactions, have been determined at
various resolutions (Table VI). Clearly, from Table VI it can be seen that
IIIA appears to possess the primary binding activity for albumin whereas
IIA is more specialized.
182 DANIEL C. CARTER AND JOSEPH X. HO

TABLE V1
Locations of Ligand Binding to HSA"

Data set Ligand D N Rf Observed location

HSAAlA Aspirin 4.0 7362 0.11 IIA,IIlA


HMRWN 1A Warfarin 5.0 2555 0.167 IIA
HSAV 1A Diazapam 6.8 2075 0.1 18 IIIA
HDGXlA Digitoxin 5.0 375 1 0.137 IIIA
HCHLOlA Chlofibrate 6.0 2175 0.138 IIIA
HMRIB 1A Ibuprofen 6.0 2402 0.215 IIIA
HAZTlA AZT 4.0 7548 0.124 llIA
HTIBlA TIB 4.0 543 1 0.12 IlA,lIIA
HMRIS4A 1s 4.0 6334 0.19 IIA,IIIA
HDlSlA DIS 4.0 4734 0.20 IlA,IIIA

a D,Resolution or d spacing in angstroms; N , number of unique reflections; R/, fractional


R-factor between native and derivative data sets. To prepare ligand-HSA complexes,
crystals of HSA were placed in a stabilizing solution of 45% polyethylene glycol 400 MW
(pH 6.8-7.0), which contained approximately 0.5 to 5 mM of the compound of interest,
and were allowed to stand for 24 to 72 hr. X-Ray diffraction data were collected on Siemens
multiwire area detector as described by He and Carter (1992). Reprinted with permission
from Nature (He and Carter, 1992); copyright 1992, Macmillan Magazines Limited.

Prior to the structure determination, solution studies by Sudlow et al.


(1975) led to the classification of binding sites on HSA in two categories,
which were denoted Site I and Site I1 from experiments based on molec-
ular interactions with fluorescent probes. The binding locus of large
heterocyclic compounds possessing a negative charge was identified as
Site I, with Site I1 showing a preference for small aromatic carboxylic
acids. Spectroscopic studies have confirmed the notion of two prin-
cipal binding sites, and identified Tyr-411 with Site 11, the principal
active binding region of HSA. Tyr-411 has also been identified with
the esterase-like activity of HSA (Sollene and Means, 1979; Ozeki et al.,
1980; Kurono et al., 1983). Based on competitive inhibition studies, Lys-
199 interacts with drugs of Sudlow's Site I (Sudlow et al., 1975). Addi-
tionally, in the same studies, fluorescence damping of Trp-2 14 on ligand
binding has also associated this residue with Site I. Consequently, it is
clear that Sudlow's Site I corresponds to the hydrophobic pocket in
subdomain HA, and his Site I1 to the pocket within IIIA. This realization,
together with the crystallographically determined major binding loca-
tions for several key markers (Table VI), can now be utilized to explain a
wealth of ligand-binding data without ambiguity. Hence, we have
adopted Sudlow's nomenclature and expanded it to include all six major
ligand-binding sites.
SERUM ALBUMIN STRUCTURE 183

The mystery associated with the diverse binding chemistry of albumin


can be explained in simple terms. One can, from efficiency consider-
ations, appreciate that nature has converged on relatively few binding
sites with multifunctional properties. The amino acid residues that line
the cavities are quite similar in charge distribution for both IIA and IIIA,
but still impart specialized selectivity. In each of the two subdomains
there is an asymmetric distribution leading to a hydrophobic surface on
one side and a basic or positively charged surface on the other (Fig. 13,
see color insert). This explains the discriminatory affinity of albumin for
small anionic compounds. The van der Waals surface of the binding
pocket in IIIA is shown in Fig. 14 (see color insert). The region is an
elongated sock-shaped pocket wherein the foot region is primarily hy-
drophobic and the leg is primarily hydrophilic. The opening to the
pocket is clearly accessible to the solvent.
Albumins of several species possess interesting enzymatic properties.
For example, Tyr-4 11 has been associated with weak esterase activity,
reacting specifically and rapidly with p-nitrophenyl acetate and many
other activated esters, including those of selected fatty acids (Koh and
Means, 1979). Human serum albumin is the most reactive of the species
studied thus far. In contrast, ESA exhibits little or no reactivity (Elkarim
and Means, 1988). In the crystal structure of HSA, the reactive hydroxyl
of Tyr-411, located in the IIIA binding pocket in close proximity to
bound ligands, is -2.7 8, from the nitrogens of Arg-410, which may
explain the unusual reactivity of this residue toward nucleophilic substi-
tution. However, in the recently determined structure of ESA, no obvi-
ous explanation can be offered for its lack of esterase activity. Previously,
this difference in activity was explained by the amino acid substitution of
Tyr-411 by Leu (Chincarini and Brown, 1976), based on partial amino
acid sequence determinations. However, the recently determined cDNA
sequence of ESA clearly indicates that Tyr-411 is conserved (Ho et al.,
1993).Perhaps the differences observed are due to polymorphism, which
has now been demonstrated for several hooved species. Other chemical
reactions of HSA associated with Site I and Site I1 involve the participa-
tion of Lys-199. This residue, which can be acetylated by aspirin (Walker,
1976), appears to be a site of nonenzymatic glycosylation of HSA (Day et
al., 1979), and is considered to be the major site of conjugation to ben-
zoylpenicillin groups (Yvon and Wal, 1988). Interestingly, the covalent
adduct produced by the reaction of Lys-199 with penicillin is the prin-
cipal antigenic determinant of the allergic reaction to this drug (Ahlstedt
and Kristofferson, 1982). The reactivity of Lys-199 may be attributed to
its unusually low pK, of 7.9 (Gerig and Reinheimer, 1975), which can be
rationalized based on the close interaction of Lys- 199 with His-242.
184 DANIEL C. CARTER AND JOSEPH X. H O

As one might predict, there is enantiomorphic selectivity with regard to


various chiral ligands. For example, albumins in general have a 100- -
fold higher affinity for L-tryptophan (King and Spencer, 1970; La-
gercrantz et d.,1981) and stereoselectivity is also known for many phar-
maceuticals such as d-oxazepam hemisuccinate (Muller and Wollert,
1975).
Much of the chemistry of albumin can be understood from detailed
observations of the 2,3,5-triiodobenzoic acid (TIB) complexes with al-
bumin. The crystalline complex of TIB with HSA and ESA has been
determined with resolution sufficient to position the molecule within the
binding pocket unambiguously and to identify the chemistry of interac-
tion (Fig. 15, see color insert). This ligand has a moderate and equal
affinity for IIA and IIIA in both HSA and ESA. Association constants
were estimated by Scatchard analysis (Scatchard, 1949) using direct linear
plots (Eisenthal and Cornish-Bowden, 1974) to be 2.2 x lo5 M-' and
8.3 X lo4 M - ' for HSA and ESA, respectively.
Overall, ESA and HSA exhibit a similar but distinct binding chemistry
with TIB (Fig. 13). However, it is the similarities, not the differences, that
are striking. Th e major binding differences occur when comparing the
two IIA subdomains. In IIA the carboxyl of ESA-bound TIB is oriented
primarily toward Arg-257, but also shares this interaction with His-242
(not Lys-199 as in HSA). This is a result of a shift of TIB in ESA by a few
angstroms toward helix h6(II) (Fig. 13A). T he binding of TIB in sub-
domain IIIA is quite similar for both ESA and HSA (Fig. 13B). Alto-
gether, they share 18 out of 19 equivalent residues in close proximity to
TIB in IIA and 15 out of 16 for IIIA (Table VII). When one considers
only the residues in immediate contact with ligand (approximately 10 to
11 residues in each subdomain), 8 or more are strictly conserved in all
mammalian albumins. Observations of the amino acid sequences of all
albumins in Table I1 reveal that several of the residues are invariant in
subdomain IIIA, whereas relatively few are conserved in IIA (Section
IV). This agrees with the observed similarity in binding chemistry with
TIB in subdomain IIIA of ESA and HSA. Knowledge of these residues
should provide important information in guiding future studies of the
underlying chemistry of albumin.
T h e absence of a similar binding chemistry in analogous subdomain IA
can be explained by the crystal structure. A nonhelical section of polypep-
tide, which includes Cys-62 of the first disulfide, allows helix h3 to
markedly increase its packing angle with h4 (normally near O"), effec-
tively eliminating the binding pocket within IA. This slight but signifi-
cant difference in folding toplogy among domains can be readily seen in
Fig. 7 (see color insert).
TABLE VII
Amino Acids Involved in Licand Binding'

11A

Y(149) K( 199) F(211) W(214) S(2 15) R(2 18) L(219) R(222) F(223)
Equine Y Y Y(V)W Y Y(A)w Y Y N(K) Y
Human Y Y Yw Y Yw Yf Y Y Y

L(234) L(238) H(242) R(257) L(260) A(261) l(264) l(290) A(291) E(292)
Equine Y(I) Y Y Y Yw Yw Y Y Y Yw
Human Y Y Y Y Yw Yw Y Y Y Yw

IIlA

P(384) L(387) I(388) N(391) C(392) F(395) R(4 10) Y(411) L(430)
Equine Y Y Y Y Yw Nw Y Y Y
Human Y Y Y Y Yw Yw Y Y Y

V(433) C(438) A(449) E(450) L(453) R(485) S(489)


Equine Y Yw Y(S) Y Y Y Y
Human Y Yw Y Y Y Y Y

Amino acids that are involved in binding in ESA and HSA are marked Y for yes, N for none. Boldface Y indicates that the amino acids are conserved
in both ESA and HSA. Amino acids that are more remote but still form important sides of the binding region are designated w for wall and f for far, as
additional identifiers. Reproduced with permission from Ho et al., 1993.
186 DANIEL C. CARTER A N D JOSEPH X. HO

B . Long-Chain Fatty Acids: Sates III and IV


Albumin has a well-known affinity for a variety of saturated and unsat-
urated fatty acids of varying chain lengths (Table V). It is recognized as
the principal transport protein for fatty acids and other lipids that would
otherwise be insoluble in the circulating plasma. The total fatty acid
capacity of albumin varies with fatty acid chain length, but averages
approximately six fatty acids per albumin molecule. Under normal physi-
ological conditions, albumin carries one or two fatty acids (Peters, 1985).
This seems to agree with a number of studies indicating that long-chain
fatty acids bind to separate and distinct sites, compared to ligands that
bind to Site I and Site I1 (Koh and Means, 1979; Sollene and Means,
1979). Reed (1986) and Parks et al. (1983) also find that there are two to
three dominant long-chain fatty acid sites. A method has been developed
by Brodersen et al. (1990) to compensate for the inherent solubility
problems associated with long-chain fatty acids in aqueous media. In this
method, which should have broad applicability, the fatty acids are trans-
ported from solutions of HSA containing fatty acids, across membranes
to albumin solutions devoid of fatty acids. Conclusions from these experi-
ments are in agreement with earlier classical studies and imply that there
are two high-affinity long-chain fatty acid-binding sites and four with
lesser affinity. Despite this consensus, one earlier study suggests that
there are no major sites that dominate; rather, it is proposed that all fatty
acids are distributed evenly over six to nine sites (Spector and Fletcher,
1978). Hamilton et al. (1991), using BSA and '%-enriched oleic acid,
determined one primary site in domain IB and two additional primary
sites in domain 111. Although little is known about the precise locations of
these bound long-chain fatty acids, Reed (1986) asserts that three lysine
residues are primarily associated with the carboxyls of the fatty acids,
namely Lys-116, Lys-349, and Lys-473. On the other hand, glucosylation
of albumin at Lys residues 525, 199, 281, and 439 greatly inhibits the
binding of cis-parinaric acid (Shaklai et al., 1984; Iberg and Fliickiger,
1986) and Shockley and Brown (1980) identify residues His-145, Lys-220,
Cys-34, Lys-412, and His-336 as reactants with N-danzylaziridine, which
inhibits binding of stearate, palmitate, and oleate. Further experiments
by Brown and Shockley (1982) revealed that stearate and palmitate, but
not bilirubin or octanoate, can inhibit the reaction with His-145. Hence,
there is no clear consensus within the current scientific literature on the
exact location of bound long-chain fatty acids, except that their locations
do not correspond to Site I and Site 11. Additionally, there appears to be
general agreement that the major sites reside in domains I and 111
(Peters, 1985), possibly in the B subdomains (Brown and Shockley, 1982).
SERUM ALBUMIN STRUCTURE 187

The binding chemistry for fatty acids shifts abruptly for fatty acids with
chain lengths of 10 carbon atoms or less. Smaller fatty acids such as these
compete for a common binding site with tryptophan and diazepam
(Kragh-Hansen, 1983). Thus, one of the primary binding sites for these
medium-length fatty acids can be assigned with confidence to the binding
pocket within IIIA.
Crystallographicstudies of fatty acid binding to serum albumin are in
progress in our laboratory as this manuscript finds its way to the pub-
lisher. However, some interesting preliminary results can be discussed
here. We have found from the study of cocrystals of rHSA and HSA with
long-chain fatty'acids of laureate or palmitate that significant conforma-
tional changes have taken place. The binding of three or more fatty acids
produces a slight opening of the interface between the two halves of the
molecule and a rotation of domain I (Fig. 16, see color insert). This is
contrary to previous fluorescence and absorption spectroscopic studies
that propose that albumin becomes more compact with the addition of
long-chain fatty acids (Honor6 and Pedersen, 1989). In addition, it ap-
pears that the environment surrounding Cys-34 becomes more exposed
to solvent. The latter findings are in excellent agreement with the ob-
served increase in oxidization of Cys-34 as a function of fatty acid binding
(Takabayashiet al., 1983).Conversely, the same study found that oxida-
tion of the free sulfhydryl increased the affinity of albumin for fatty acids.
Recent diffraction experiments wtih HSA cocrystallized separately with
lauric acid and a novel iodinated lauric acid analog have revealed three
major sites for the iodinated fatty acid. One located in IB, one in the
binding pocket of IIIA, and the other in IIIB. These findings, although
based on a derivatized fatty acid, agree well with previous studies that
indicate the presence on albumin of two to three high affinity fatty acid
binding sites (Koh and Means, 1979; Sollene and Means, 1979; Reed
1986; Parks et al., 1983). Furthermore, they are in agreement with the
number and location of fatty acid sites proposed by the "C cleic acid
studies of Hamilton et al. (1991). Based on the previous findings of Means
and co-workers one might assume that the two highest affinity long chain
fatty acid sites are those located on the two B subdomains. Additionally,
these two high affinity sites are located at or near the surface of the
molecule which may explain why, despite the high association con-
stants of fatty acids (lo' M-'),they can be readily exchanged between
albumin molecules in solution (Peters, 1985). In due course of the
structure refinement, with an improved model and diffraction data,
we hope a more detailed description of the binding process can be
made.
188 DANIEL C. CARTER AND JOSEPH X. HO

C . Metals: Sites V and VI


Albumin has a high affinity for Cu(II), Ni(II), Hg(II), Au(I), and
Ag(II), and weaker affinities for calcium and zinc. Many individuals have
spent significant portions of their professional careers studying the par-
ticular chemistry associated with each metal-binding site. For example,
Sarkar and co-workers have extensively examined the Cu and Zn binding
chemistry associated with the N terminus (Dixon and Sarkar, 1974), and
Shaw and co-workers studied the chemistry associated with Au com-
plexes with Cys-34 (Shaw, 1989). There have been fewer studies charac-
terizing the affinity for Zn. However, Lakusta et al. (1980) associate this
binding with the Cu(I1)-binding site. Consequently, albumin possesses
two major binding sites for metals, which we denote as Site V and Site VI
in the detailed discussion that follows.

1 . Cys-34: Site V
Albumin is responsible for the largest fraction of free sulfhydryl (Cys-
34) in blood serum, and studies have shown that it is also the most reactive
sulfhydryl. The chemical reactivity of Cys-34 is reported to have an
unusually low pKsH of 5 compared with 8.5 and 8.9 for cysteine and
glutathione, respectively (Pedersen and Jacobsen, 1980). In most prepa-
rations of albumin, 30-35% of Cys-34 is occupied by cysteine or gluta-
thione. Blocking of Cys-34 with cysteine, glutathione, or other chemicals
such as N-idosuccinimide stabilizes albumin against dimer formation
(Peters, 1985). Presumably, Cys-34 plays a direct or catalytic role in this
process.
Cys-34 also binds Au, Ag, Hg, Cd, and, to a lesser extent, Cu. Major
interests in the unusual chemistry at this site include understanding the
nature of bound pharmaceuticals, such as the antiarthritic auranofin, or
other gold(1)-containing thiolates (Shaw, 1989). Other pharmaceutical
interests involve reactions with non-metal-containing antibiotics and the
recently identified complexation with nitric acid (Stamler et al., 1992).
Examinations of the crystal structures of HSA and ESA in this vicinity
indicated that Cys-34 is located in a crevice on the surface of the protein
and that the reactive sulfur is somewhat protected by several residues
(Fig. 17, see color insert). In HSA, Cys-34 is in close proximity to Glu-82
and His-39. In ESA the Cys environment is also protected but, with the
exception of His-39, which (along with Cys-34) is conserved in mamma-
lian albumin sequences (Table 11),it involves contributions from differ-
ent amino acids. Thus, one may expect that His-39 plays a major role in
the enhanced reactivity of the free sulfhydryl. Interestingly, in the crystal
structure of albumin complexed with three or more long-chain fatty acids,
SERUM ALBUMIN STRUCTURE 189

the crevice containing Cys-34 has opened significantly, exposing the


sulfhydryl and increasing its distance from its nearest neighbors, includ-
ing His-39. It is noteworthy that in salmon albumin Cys-34 is replaced by
Ser, and both salmon and frog albumins have His-39 replaced by Leu. As
a result, one would predict the free sulfhydryl of Xenopur albumin to be
less reactive.

2. N Terminus: Site VI
Albumin possesses for Cu(I1) and Ni(I1) one distinct high-affinity site,
which has been well characterized by several research groups. Binding of
Cu by albumin was perhaps first described by Fiess and Klotz in 1952 and
was identified with the N terminus by Peters and Blumenstock in 1967.
They surmised that the copper atom was coordinated by the N-terminal
nitrogen, the next two peptide nitrogens, and the NE of His-3. Camer-
man et al. (1976) demonstrated similar copper binding chemistry with the
tripeptide crystal structure of glycylglycyl-L-histidine-N-methylamide, a
functional analog of the N-terminal tripeptide (Fig. 18). Albumins that
lack histidine at position 3, such as canine albumin, have a much lower
affinity for Cu(I1) and Ni(II), which presumably explains the greater
susceptibility of dogs to copper poisoning (Peters, 1984). In the crystal
structures of HSA, rHSA, and ESA, the first two to three residues are

FIG. 18. Structure of the copper-binding peptide, glycylglycyl-L-histidine-N-


rnethylarnideillustrating the coordination state and peptide conformation. Figure modified
from Carnerrnan et al. (1976).
190 DANIEL C. CARTER AND JOSEPH X. HO

disordered, implying a much greater degree of flexibility for the N


terminus. This is understandable because substantial conformational
changes would be required to accommodate the square-planar bipyrami-
dal coordination of the copper or nickel atom.

IV. EVOLUTION OF ALBUMIN STRUCTURE


A discussion of the structure of albumin would not be complete with-
out an examination of the conserved primary structure of albumin. All
the present lines of evidence point toward the evolution of albumin from
a precursor one-third the current size. This was originally suggested by
Brown (1976), and is now verified based on the internal sequence homol-
ogy and the shared topology of each of the three domains. T h e observa-
tion that, in the albumin structure, six subdomains exist that share a
common topology and are related by pseudo twofold symmetry within
each domain further indicates that the current domain of albumin
evolved from an incomplete gene duplication event, as originally sug-
gested by Brown (1976). That is, the B subdomains are missing the
additional smaller disulfide double loop. Consequently, it is the A (Fig.
13A) subdomains that represent the true protoalbumin, as well as the
principal binding properties of albumin. Whether the B subdomains
have evolved specialized functions with cell surface receptors or fatty acid
binding remains undetermined.
When considering the evolutionary significance of a series of the al-
bumin sequences in Table 11, it is interesting to examine the globin family
of proteins, for which a great deal of research has been amassed
(Dayhoff, 1978). Comparisons of the structures of hemoglobin and myo-
globin families revealed a number of residues that remained invariant
(14%). Many of these were directly related to function either by associa-
tion with the heme pocket or by participation in the interfaces responsible
for the allosteric effect. Still others seemed important for stabilization of
the protein (Dickerson and Geis, 1983). In albumin there are 332 (57%)
invariant residues in the sequences of mammalian albumins and 95 (16%)
invariant residues conserved, including frog and salmon albumins. In
both of the above comparisons, 34 of the conserved residues are cys-
teines. A useful matrix illustrating the sequence identities between the
known sequences of albumin is shown in Table VIII. T h e greatest level of
sequence identity occurs, not surprisingly, between bovine and ovine
albumins, which are 92% conserved. T h e evolution of albumin and the
construction of a phylogenetic tree has been revisited by Gray and Doo-
little (1992). Th e time for the gene duplication event leading to the
three-domain structure shared by all fish, tetrapods, and mammals is
SERUM ALBUMIN STRUCTURE 191

TABLE VIII
Matrix Indicating Sequence ldentily between Presently Known Segunces of Serum Albumin9

Source HSA BSA ESA OSA RSA FSA

BSA 44 1
ESA 442 430
OSA 435 539 438
RSA 426 409 422 404
FSA 22 1 218 222 216 225
SSA 161 170 153 165 159 154

a With the exception of lamprey serum albumin. See Table I for albumin abreviations.

placed at approximately 450 million years (MY)ago, much later in evolu-


tionary time than the 700 MY estimated by Brown and Shockley (1982).
In principle, the estimate by Gray and Doolittle is more in line with those
proposed for globin sequences (Dickerson and Geis, 1983).
As in hemoglobin, one would expect that the evolutionary pressure to
conserve residues in albumin would impart, together with the knowledge
of the atomic structure, some new insight into the chemistry and function
of albumin. This is clearly the case. Examination of the amino acids
known to participate in the binding process within IIA and IIIA reveals a
somewhat surprising result. Only two residues are invariant in the bind-
ing pocket of IIA, Arg-257 and Leu-260, whereas there are seven in IIIA
Asn-390, Cys-391, Arg-4 10, Tyr-4 11, Thr-4 12, Val-432, Cys-437, Glu-
450, Leu-452, and Arg-484. If one considers additionally the amino acids
that, with only one exception, are conserved, this brings the total to 10 in
IIIA and 3 in IIA. Clearly evolutionary pressure has operated to con-
serve the binding chemistry of the IIIA region of the serum albumin
molecule. Let us therefore consider another binding region of serum
albumin, Cys-34, which is conserved in all albumin sequences given in
Table 11, except salmon albumin. Examination of the residues that sur-
round the cleft containing the free sulfhydryl reveals no conserved pat-
tern. However, under the surface of the cleft is the highly conserved
His-39, which is in close proximity to Cys-34. One must then conclude
that it is this imidazole that imparts the unusually low p K s of
~ this residue
and further reflects its importance in sequestering various metals, cys-
teine, and glutathione. His-3, a requirement for the Cu- and Ni-binding
properties of albumin, is again conserved in all of the sequences given in
Table I1 with the exception of those from salmon and dog (Dixon and
Sarkar, 1974). As mentioned previously, this may explain the unusual
sensitivity of dogs to copper poisoning (Peters, 1984), and one wonders
192 DANIEL C. CARTER AND JOSEPH X. HO

whether the problem may be compounded in fish to impart an additional


sensitivity to exposure to Hg and Ag, for example.
Brown (1976) noted a similarity in the long disulfide loops of BSA and
the G-H region of sperm whale myoglobin, basing this observation on
the match of two prolines and a similar distribution of hydrophobic
residues. This led to his suggestion that albumin evolved from a pri-
mordial globin. Indeed, pori;ons of the structure of albumin strongly
resemble the myoglobin/hemoglobin fold. T h e helices h 1, h2, h3, and h4
superimpose on the last four helices E, F, G, and H of hemoglobin. The
striking similarity can be seen in Fig. 19, (see color insert). Although the
positions of the helices are conserved, the polarity is reversed in serum
albumin helices h3 and h4 compared to the corresponding G-H pair in
hemoglobin. Surprisingly, not only are the general helical features con-
sistent, the numbers of amino acids in each helical segment are within the
variation seen for myoglobin as well. This resemblance takes on added
meaning when one considers that the albumins of primates are known to
bind hematin with high affinity (Beaven et al., 1974). Moreover, apomyo-
globin is known to have a similar diverse affinity for a variety of small
molecules. Whether the albumin and globin families are actually related
by a common ancestor or represent yet another example of convergent
evolution is not clear. McLachlan and Walker (1977), who used an exten-
sive statistical analysis of three serum albumins to support the latter,
stated “our tests do not support the idea that serum albumin evolved
from myoglobin. It is more likely that parts of these proteins indepen-
dently acquired similar helical structures.” In this regard it is interesting
to note that the “myoglobin fold” has recently been identified in a num-
ber of functionally unrelated proteins that lack significant sequence ho-
mology (Holm and Sander, 1993).
In addition to the pseudosymmetry observed between subdomains, a
pseudo twofold screw axis relates two consecutive domains (see Section
11,C,2). The orientation of the axis relating domain I and I1 is slightly
different from that relating I1 and 111. This observation has interesting
implications regarding other multidomain structures, such as in lamprey
albumin (Gray and Doolittle, 1992). If the domain-domain packing
arrangements are conserved in albumins as distantly related as lamprey
albumin, then after applying the pseudosymmetry operator relating do-
mains I1 and 111 to the extra four domains of lamprey serum albumin
(LSA), the structure can be described as a straight helical coil with
a diameter of 80 X 40 A and a pitch of 55 (Fig. 20A). Otherwise, if
the pseudosymmetry operator relating domains I and 11, and the same
relating I1 and 111, are applied, the structure will be a curved helical coil,
or partial superhelical coil (Fig. 20B).It is reasonable to assume that the
FIG.20. Conceptual model of seven-domain structure of lamprey serum albumin. In
the HSA structure, there exist pseudo twofold screw axes between domains I and I1 and
between domains I 1 and I l l , whereas the symmetry axes are not quite parallel to each other.
In addition, the connections between two successive domains are made by the long, continu-
ing helix (consisting of the C-terminus helix h10 and the N-terminus helix h l of the
succeeding domain h 1). This arrangement greatly restricts the way two successive domains
can be packed together. If the lamprey serum albumin forms a structure similar to that of
HSA, with its extra four domains, the lamprey serum albumin structure should look like A,
assuming a single pseudo symmetry repeat, or like B, assuming two alternate-diad repeats.
The structure shown in A forms a straight zig-zag,or helical, shape; that shown in B forms a
curved zig-zag, or super-helical shape. Figures drawn using program RIBBONS (Carson,
1987).
194 DANIEL C. CARTER A N D JOSEPH X. HO

actual structure will be slightly more compact to minimize the molecular


surface. Furthermore, a cursory survey of the sequence of LSA suggests
that among the subdomains of LSA, its IA subdomain has the highest
homology to the IIIA subdomain of a “three-domain’’ albumin, based on
the distribution of previously determined invariant residues in Table 11.
One might postulate that it may also function likewise.

V. SUMMARY AND FUTUREDIRECTIONS

Albumin is clearly an extraordinary molecule of manifold functions


and applications. Although the exact function of albumin has been de-
bated, much of the present data support the notion that the principal role
of albumin in the circulatory system is to aid in the transport, metabolism,
and distribution of exogenous and endogenous ligands. T h e ability of
albumin to act as an important extracellular antioxidant (Halliwell, 1988)
or impart protection from free radicals, and other harmful chemical
agents (Emerson, 1989) agrees well with the increased susceptibility of
analbuminemic rats to cancer (Kakizoe and Sugimura, 1988). T h e ex-
pression and delivery of albumin to the circulatory system by the liver
therefore seem appropriate. An overview of the prolific ligand-binding
properties of albumin is summarized in Fig. 2 1. The positions of known
binding sites for important pharmaceutical markers such as diazepam,
ibuprofen, aspirin, and warfarin are illustrated. In addition, the impor-
tant endogenous markers tryptophan, octanoate, and bilirubin are also
shown. With the exception of the definitive positions of the long-chain
fatty acids, most albumin-ligand chemistry can now be explained by the
atomic coordinates derived from crystal structures. Knowledge of the
atomic structure coupled with the current applications of genetic engi-
neering, such as site-directed mutagenesis, promises to provide an even
greater understanding of albumin chemistry.
It is widely accepted in the pharmaceutical industry that the overall
distribution, metabolism, and efficacy of many drugs can be altered based
on their affinity to serum albumin. In addition, many promising new
drugs are rendered ineffective because of their unusually high affinity
for this abundant protein. Obviously, an understanding of the chemistry
of the various classes of pharmaceutical interactions with albumin can
suggest new approaches to drug therapy and design, placing albumin in
its rightful place as the “second step in rational drug design.” Application
of albumin in other therapeutic approaches is widely known. Some stud-
ies have suggested that modified serum albumin may be used as a se-
lective contrast agent for tumor detection and/or therapy (Sinn et al.,
1990). Other studies have demonstrated that albumin may be used to
SERUM ALBUMIN STRUCTURE 195

FIG. 2 1 . Illustration summarizing the various ligand-binding sites on serum albumin.


The asterisks denote binding sites that can be inferred; all others have been determined
crystallographically.

deliver toxic compounds for elimination of Mycobacterium tuberculosis via


receptor-mediated drug delivery (Majumdar and Basu, 1991). Recently,
chimeric albumin molecules such as HSA-CD4 (Yeh et al., 1992) and
HSA-Cu,Zn-superoxide dismutase (Ma0 and Poznansky, 1989) have
been utilized to increase the half-life and distribution, and reduce the
immunogenicity, of these potential protein therapeutics. Albumin has
now been cloned and expressed in several bacterial and fungal systems.
The primary motivation for many of these studies has been the potential
of recombinant albilmin to serve as a serum replacement product that is
free from unwanted viral contaminants, e.g., hepatitis, herpes, and hu-
man immunodeficiency virus (HIV). The most successful production has
been achieved by extracellular expression in yeast (Etcheverry et al.,
1986; Hinchcliffe and Kenney, 1986; Kalman et al., 1990; Okabayashi et
al., 1986; Quirk, et al., 1989; Sijmons et al., 1990; Sleep et al., 1991).
Clearly, future scientific and therapeutic applications of albumin appear
limitless.
196 DANIEL C. CARTER AND JOSEPH X. HO

In conclusion, albumin may be unique among proteins in that so many


scientists have spent the largest portion of their professional careers
studying very specific aspects of this protein. New appreciation for the
complexity and potential applications presented by the structure of al-
bumin promises to consume the careers of many more scientists.

w LEDGMENTS
ACKNO
The authors are indebted to several individuals, for their generous support during the
course of preparing this manuscript. In particular, we are grateful to Pam Twigg for help in
preparing many of the tables and references, Teresa Miller for proofreading the manu-
script, and Jewel1 Reynolds, Kim Keeling, Brenda Barnes, Tongi Shavers, and Mike Carson
for assisting with the preparation of the figures. We thank F. W. Putnam for providing the
tabulation of point mutations of albumin in advance of publication. Reproduction of the
color figures was provided as a courtesy of the National Aeronautics and Space Administra-
tion, Marshall Space Flight Center.

REFERENCES
Abdo, Y., Rousseaux, J.. and Dautrevaux, M. (1981). FEES Lett. 131, 286-288.
Adams, P. A., and Berman, M. C. (1980). Biochem. J. 191,95-102.
Adinolfi, A., Adinolfi, M., and Lessof, M. H. (1975).J. Med. Genet, 12, 138-151.
Ahlstedt, S., and Kristofferson, A. (1982). Prog. Allergy 30,67.
Akoi. K., Sato, K., Nagoaka, S., Kamada, M.,and Hiramatsu, K. (1973). Biochim. Biophys.
Acta 328,323-333.
Ancell, H. ( 1 839). Lancet 1,222-23 1 .
Arai, K., Huss, K., Madison, J., Putnam, F. W., Salzano, F. W., Franco, M. H. L. P., Santos,
S. E. B. and Freitas, M. J. M. (1989a). Proc. Natl. Acad. Sci. U.S.A., 86, 1821-1825.
Arai, K., Madison, J., Huss, K., and Putnam, F. W. (1989b).Proc. Natl. Acad. Sci. U.S.A. 86,
6092-6096.
Arai, K., Madison, J., Shimizu. A., and Putnam, F. W. (1990). Proc. Natl. Acad. Sci. U.S.A.,
87,497-501.
Beaven, G. H., Chen, S.-H., DAlbis, A., and Gratzer, W. D. (1974). Eur. J. Biochem. 41,
539-546.
Behrens, P. Q., Spiekerman, A. M.,and Brown, J. R. (1975). Fed. Proc., Fed. Am. SOC.Exp.
Biol. 34, 591.
Bendedouch, D., and Chen, S. H. (1983).J. Phys. Chem. 87,1473-1477.
Bloomfield, V. (1966).Biochemistty, 5,684-689.
Borst, G . C., Eil, C., and Burman, K. D. (1983).Ann. Intern.Med. 98,366-378.
Bos, 0.J. M., Labro, J. F. A., Fischer, M. J. E., Witling, J., and Janssen, L. H. M . (1989).
J . Biol. Chem. 264,953-959.
Brand, J. G., and Toribara, T. Y. (1975).Arch. Biochem. Biophys. 167,91-98.
Brand. S., Hutchinson, D. W., and Donaldson, D. (1984). Clin. Chim. Acta 136, 197-202.
Breckenridge, A., and Rosen, A. (1971).Biochim. Biophys. Acta 229,610-617.
Brennan, S. 0.(1985).Biochim. Biophys. Acta 830,320-324.
Brennan, S. O., and Carrell, R. W. (1978).Nature (London) 274,908-909.
Brennan, S. O., Peach, R. J., and Boswell, D. R. (1989).Biochim. Biophys. Acta 993,48-50.
Brennan, S . O., Arai, K., Madison, J., Laurell, C.-B., Galliano, M., Watkins, S., Peach, R.,
Myles, T., George, P., and Putnam, F. W. (1990a). Proc. Natl. Acod. Sci. U.S.A. 87,
3909-3913.
SERUM ALBUMIN STRUCTURE 197

Brennan, S. 0..Myles, T., Peach, R. J.. Donaldson, D., and George, P. M. (1990b).Proc.
Natl. Acad. Sci. U.S.A. 87,26-30.
Brock, A. (1975).Acta Phurmucol. Toxicol. 36, 13-24.
Brodersen, R. (1979).J.Eiol. Chem. 254,2364-2369.
Brodersen, R., Andersen, S., Vorum, H., Nielsen, S. U., and Pedersen, A. 0.(1990).Eur.J.
Biochem. 189,343-349.
Brown, J. R. (1975).Fed. Proc., Fed. Am. SOC.Exp. Eiol. 34,591.
Brown, J. R. (1976).Fed. Proc., Fed. Am. SOC.Exp. Biol. 35,2141-2144.
Brown, J. R. (1977).In “Albumin Structure, Function and Uses”(V. M. Rosenoer, M. Oratz,
and M. A. Rothschild, eds.), pp. 27-51. Pergamon, Oxford.
Brown, J. R., and Shockley, P. (1982).In “Lipid-Protein Interactions” (P. Jost and 0. H.
Griffith, eds.), Vol. 1, pp. 25-68. Wiley, New York.
Brown, K. F., and Crooks, M. J. (1976).Biochem. Phurmacol. 25, 1175-1178.
Brown, W. M., Dziegielewska, K. M., Foreman, R. C., and Saunders, N. R. (1989).Nwleic
Acids Res. 17, 10495.
Burke, C. W., Lewis, B., Panveliwalla, D., and Tabaqchali, S. (1971).Clin. Chim. A c h 32,
207-2 14.
Byrnes, L., and Cannon, F. (1990).DNA Cell Eiol. 9,647-655.
Callan, W. M., and Sunderman, F. W., Jr. (1973).Res. Commun. Chem. Pudol. Phurmacol. 5,
459-472.
Cambillau, C., and Horjales, E. (1987).J.Mol. Graphics 5, 174.
Camerman, N., Camerman, A., and Sarkar, B. (1976).C a n . ] . C h . 54,1309-1316.
Cannistraro, S.,and Sacchetti, F. (1986).Phys. Rev. 33,745-746.
Carlson, J. Sakamoto, Y., Laurell, C.-B. Madison,J., Watkins, S., and Putnam, F. W. (1992).
Proc. N a L Acad. Sci. U.S.A. 89,8225-8229.
Carson, M. (1987).J.Mol. Graphics, 5, 103-106.
Carter, D. C.,and He, X.-M. (1990).Science 249,302-303.
Carter, D. C., He, X.-M., Munson, S. H., Twigg, P. D., Gernert, K. M., Broom, M. B., and
Miller, T. Y. (1989).Science 244, 1195-1198.
Charlwood, P. A. (1961).Ei0chem.J. 78, 163-172.
Chen, M. C., and Lord, R. C. (1976).J.Am. Chem. SOC.98,990-992.
Chignell, C. F. (1969).Mol. Phurmucol. 5,244-252.
Chignell, C. F. (1970).Mol. Pharmucol. 6,1-12.
Chignell, C. F. (1973).Ann. N . Y . Acud. Sci. 226,44-59.
Chincarini, C. C., and Brown, J. B. (1976).Fed. Proc., Fed. Am. SOC.Exp. Eiol. 35, 1621.
Cho. M.J., Mitchell, A. G . , and Pernarowski, M. (1971).J.Phunn. Sci. 60, 196-200.
Coassolo, P., Briand, C., Bourdeaux, M., and Sari, J. C. (1978).Eiochim. Biophys. A c h 538,
5 12-520.
Cohn, E. J., Hughes, W. L., Jr., and Weare, J. H. (1947).J. Am. Chem. SOL. 69, 1753-
1761.
Cooke, N. E., and David, E. V. (1985).J . Clin. Invest. 76,2420-2424.
Crooks, M. J., and Brown, K. F. (1974).J.Phurm. Phurmacol. 26,304-31 1.
Daughaday, W. H. (1959).Physiol. Rev. 39,885-902.
Day, J. F., Thorpe, S. R., and Baynes, J. W. (1979).J.Eiol. C h m . 254,595.
Dayhoff, M. 0.(1978).“Atlas of Protein Sequence and Structure,” Vol. 5. Georgetown
University, Washington, DC.
DeLucas, L. J., Smith, C. D., Smith, W., Senadhi, V.-K., Senadhi, S. E., Ealick, S. E., Carter,
D. C., Snyder, R. S., Weber, P. C., Salemme, F. R., Ohlendorf, D. H., Einspahr, H. M.,
Clancy, L. L., Navia, M. A., McKeever, B. M., Nagabhushan,T. L., Nelson,G., McPher-
son, A., Koszelak, S., Taylor, G., Stammers, D., Powell, K., Darby, G., and Bugg, C. E.
(1989).Science 246,65 1-654.
198 DANIEL C. CARTER AND JOSEPH X. HO

Dickerson, R. E., and Geis, I. G. (1983).“Hemoglobin: Structure, Function, Evolution, and


Pathology.” BenjaminlCummings, Menlo Park, CA.
Dixon, J. W., and Sarkar, B. J. (1974).J.Biol. C h . 249,5872-5877.
Dugiaczyk, A., Law, S. W., and Dennison, 0. E. (1982).Proc. Natl. Acad. Sci. U.S.A. 79,
71-75.
Eisenthal, R., and Cornish-Bowden, A. (1974).Biochem. J. 139,715-720.
Elkarim, A. A., and Means, G. E. (1988).Comp. Biochem. Physiol. B. 91B,267-272.
Emerson, T. E., Jr. (1989).CRC Crit. Care Med. 17,690-694.
Etcheverry, T.,Forrester, W., and Hiueman, R. A. (1986).Bio. Technology 4, 720-726.
Fehske, K.J., Miiller, W. E., and Wollert, U. (1981).B i o c h . Phannocol. 30,687-692.
Feldhoff, R. C., and Ledden, D. J. (1983).Biochem. Biophys. Res. Commun. 114,20-27.
Feng, L.,Hu, C. 2..and Andrade, J. D. (1988).J.Colloid Inlerface Sci. 126,650-653.
Fiess, H. A.,and Klotz, 1. M. (1952).J.Am. Chem. SOC.74,887-891.
Figge, J., Rossing, T. H., and Fencl, V. (1991).J.Lab. Clin. Med. 117,453-467.
Fogh-Andersen, N. (1977).Clin. Chem. (Winston-Salem, N . C . ) 23,2122-2126.
Foster, J. F. (1960).In “The Plasma Proteins” (F. W. Putnam, ed.), 1st ed., Vol. 1, pp.
179-239. Academic Press, New York.
Foster, J. F. (1977).In “Albumin Structure, Function and Uses” (V. M.Rosenoer, M. Oratz,
and M.A. Rothschild, eds.), pp. 53-84. Pergamon, Oxford.
Franklin, S. G., Wolf, S. I., Ozdemir, Y., Yuregir, G. T., Isbir, T., and Blumberg, B. S.
(1980).Proc. Natl. Acad. Sci. U.S.A. 77,5480-5482.
Gabay, S., and Huang, P. C. (1974).Adv. Biochem. Psychophannocol. 9,175-189.
Galliano, M., Minchiotti, L., Iadarola, P., Stoppini, M., Ferri, G., and Castellani, A. A.
(1986).FEES Lett. 208,364-368.
Galliano, M., Minchiotti, L., Iadarola, P., Ferri, G., Zapponi, M. C., and Castellani, A. A.
(1988).FEES Lett. 233, 100-104.
Galliano, M., Minchiotti, L., Stoppini, M., and Tarnoky, A. L. (1989).FEES Lett. 255,
295-299.
Galliano, M., Minchiotti, L., Porta, F., Rossi, A., Ferri G., Madison, J., Watkins, S., and
Putnam, F. W. (1990).Proc. Natl. Acad. Sci. U.S.A. 87,8721-8725.
Galliano, M. et al. (1993).Biochim. Biophys. Actu, in press.
Garten, S . , and Wosilait, W. D. (1972).Comp. Ga.Pharmacol. 3,83-88.
Geisow, M.J., and Beaven, G. H. (1977).B i 0 c h . J . 163,477-484.
Gerig, J. T., and Reinheimer, J. D. (1975).J.Am. Chem. SOC.97, 168.
Goodman, D. S. (195q.J.Am. Chem. SOC.80,3892-3898.
Gorin, M.B., Cooper, D. L., Eiferman, F., van de Rijn, P., and Tilghman, S. M. (1981).
J. Biol. C h . 256, 1954-1959.
Gray, J. E., and Doolittle, R. F. (1992).Protein Sci. 1,289-302.
Green, H.O.,Moritz, J.. and Lack, L. (1971).Biochim. Biophys. Actu 231,550-552.
Giirber, A. (1894).Sitzungsber. Phys.-Med. Ges. Wurrburg, p. 143.
Guthrow, C. E., Morris, M. A., Day, J. F., Thorpe, S. R., and Baynes, J. W. (1979).Proc. Natl.
Atad. Sn’. U.S.A. 76,4258-4261.
Guttler, F., Tybring, L., and Engberg-Pedersen, H. (1971). B Y . ] . Phurmacol. 43, 151-160.
Haefliger, D. N.,Moskaitis,J. E., Schoenberg, D. R., and Wahli, W. (1989).J.Mol. Evol. 29,
344-354.
Hagag, N.,Birnbaum, E. R., and Darnall, D. W. (1983).Biochemistry, 42,2420-2427.
Halliwell, B. (1988).B i o c h . Pharmacol. 37,569-571.
Hamilton, J. A., Era, S., Bhamidipati, S. P., and Reed, R. G. (1991).Proc. Natl. Acad. Sci.
U.S.A. 88, 2051-2054.
Harrington, W. F., Johnson, P., and Ottewill, R. H. (1956).B i 0 c h . J . 62,569-582.
SERUM ALBUMIN STRUCTURE 199

Hart, B. J. T., Witling, J,, and De Gier, J. J. (1986).Biochem. Phanacol. 35, 1005-1009.
He, X.-M., and Carter, D. C. (1992).Noture (London)358,209-215.
Hinchcliffe, E.,and Kenney, E. (1986).Eur. Pat. Appl. 201-239.
Ho, J. X.,Holowachuk, E. W., Norton, E. J., Twigg, P. D., and Carter, D. C. (1993).Eu7.J.
Biochem. 415,205-212.
Holm, L., and Sander, S. (1993).Nature (London) 361,309.
Holowachuk, E.W. (1991).G m Bank D a l a h e , Accession No. M73993.
Honor6 B., and Pedersen, A. 0. (1989).Biochem. J. 258,199-204.
Huang, P. C., and Gabay, S. (1974).Biochem. Phanacol. 43,957-972.
Hughes, W. L. (1954).In “The Proteins” (H. Neurath and K. Biley, eds.), Vol. 2b, pp.
663-755.Academic Press, New York.
Hughes, W. L., and Dinitz, H. M. (1964).J.Biol.Chem. 439,845-849.
Hultrnark, D., Borg, K. 0.. Elofsson, R., and Palmer, L. (1975).Acta Pharm. Suec. 14,
259-276.
Huss, K., Madison,J., Ishioka, N., Takahashi, N., Arai, K., and Putnarn, F. W. (1988).Proc.
Natl. Acad. Sci. U.S.A. 85,6692-6696.
Hutchinson, D. W., and Matejtschuk, P. (1985).FEBS Lett. 193,211-212.
Iadarola, P., Minchiotti, L., and Galliano, M. (1985).FEES Lett. 180,85-88.
Iberg, N., and Fluckiger, R. (1986).J.Biol. Chem. 461, 13542-13545.
Jacobsen, C. (1972).Eur.J. Biochem. 47,513-519.
Jacobsen, J. (1969).FEES Lett. 5,112-1 14.
Jacobsen, J. (1977).Int. J.P e p . Protkn Res. 9,235-239.
Jagodzinski, L. L., Sargent, T. D., Yang, M., Glackin, C., and Bonner, J. (1981).Proc. Natl.
Acad. Sci. U.S.A. 78(6),3521-3525.
Janssen, L. H. M., and Nelen, T. H. A. (1979).J.Biol. Chem. 454,5300-5303.
Kakizoe, T., and Sugimura, T. (1988).Jpn.J. Cancer Res. 79,775-784.
Kalman, M., Cserpan, I., and Bajszar, G. (1990).Nuleic AcidcRes. 18,6075-6081.
Katchalski, E.,Benjamin, G. S., and Gross, V. (1957).J. Am. Chem. Soc. 79,4096-4099.
Kendall, F. E. (1941).J.Biol. Chem. 138,97-109.
Kendrew, J. C., Dickerson, R. E., Strandberg, B. E., Hart, R. G., Davies, D. R., Phillips,
D. C., and Shore, C. (1960).Nature (London) 185,422-427.
Keresztes-Nag, S., Mais, R. F., Oester, Y. T., and Zaroslinski,J. F. (1972).Anal. Biochem. 48,
80-89.
Khan, M. Y. (1986).Bi0chem.J. 236,307-310.
Khan, M. Y., and Salahuddin, A. (1984).Eu7.J. Biochem. 141,473-475.
King, T. P. (1973).Arch. Biochem. Biophys. 156,509-520.
King, T. P., and Spencer, M. (1970).J. Biol. Chem. 445,6134-6148.
Klopfenstein, W. E.(1969a).Biochim. Biophys. Acla 181,323-325.
Klopfenstein, W . E.(1969b).Biochim. Biophys. A c h 187,272-274.
Koh, S.-N., and Means, G. E. (1979).Arch. Biochem. Biophys. 194,73-79.
Kragh-Hansen, U.(1981).Phanacol. Rev. 33, 17-53.
Kragh-Hansen, U. (1983).Biochem.J. 209, 135-142.
Kragh-Hansen, U. (1990).Dan. Med. Bull. 37,57-84.
Kurono, Y.,Kondo, T., and Ikeda, K. (1983).Arch. Biochem. Biophys. 227,339-341.
Lagercrantz, C., Larsson, T., and Denfors, 1. (1981).Comp. Biochem.P h y s d . C . 69C,375-
378.
Lakusta, H., Deber, C. M., and Sarkar, B. (1980).Can.J. Chem. 58,757-766.
Lang, J. H., and Lasser, E. C. (1967).Invest. Radiol. 2,396-400.
Lau, S.-J., Kruck, T. P. A., and Sarkar, B. (1974).J.Biol. Chem. 449,5878-5884.
Law, S. W., and Dugaiczyk, A. (1981).Nuture (London)291,201-205.
200 DANIEL C. CARTER AND JOSEPH X. HO

Lawn, R. M., Adelman, J., Bock, S. C., Franke, A. E., Houck, C. M., Najarian, R. C..
Seeburg, P. H., and Wion, K. L. (1981).Nucleic AcidsRes. 9,6103-6114.
Leob, G. I., and Scheraga, H. A. (1956).J.Phys. Chem. 60, 1633-1644.
Lewin, J. (1951).J.Am. Chem. SOC.73,3906-3916.
Lindup, W. E.(1987).Prog. DrugMetab. 10, 141-185.
Longsworth, L. F. (1954).J.Phys. Chem. 58,770.
Low, B. W., and Richards, F. M. (1954).J.Am. Chem. SOC.76,2511-2518.
Low, B. W., and Weichel, E. J. (1951).J.Am. Chem. SOC.73,3911-3916.
Luetscher, J. (1939).J.Am. Chem. SOC.61,2888.
Luft, A. J., and Lorscheider, F. L. (1983).Biochemistry 22,5978-5981.
Lukas, D. S.,and De Martino, A. G . (1969).J.Clin. Invest. 48, 1041-1053.
Madison,J., Arai, K., Sakamoto, Y., Feld, R. D., Kyle, R. A,, Watkins, S., Davis, E., Matsuda,
Y.-l., Amaki, I., and Putnam, F. W. (1991).Proc. Natl. Acad. Sci. U.S.A. 88,9853-9857.
Majumdar, S . , and Basu, S. K. (1991).Antimicrob. Agents Chemother. 35, 135-140.
Mao, G. D., and Poznansky, M.J. (1989).Biomater. Artif. Cells Artif. Organs 17(3),229-244.
Mason, R. W.,and McQueen, E. G. (1974).P h m c o l o g y 12, 12- 19.
McClure, R. J., and Craven, B. M. (1974).J. Mol. Biol. 83,551-555.
McLachlan, A. D., and Walker, J. E. (1977).J.Mol. Biol. 112,543-558.
McMeekin, T. L. (1939).J. Am. Chem. SOC.61,2884-2890.
McMenamy, R. H., and Oncley, J. L. (1958).J.Biol. Chem. 233, 1436-1447.
Meloun, B., Moravek, L., and Kostka, V. (1975).FEES Lett. 58, 134-137.
Miller, T. Y.,He, X.-M., and Carter, D. C.(1992).J.Cryst. Growth 122,306-309.
Minchiotti, L., Galliano, M., Iadarola, P., Stoppini, M., Ferri, G., and Castellani. A. A.
(1987).Biochim. Biophys. Acta 916,411-418.
Minchiotti, L., Galliano, M., Iadarola, P., Zepponi, E., and Ferri, G. (1990).Biochim. Biophys.
Acta 1039,204-208.
Minchiotti, L., Galliano, M., Stoppini, M., Ferri, G., Crespeau, Rochu, D., and Porta, F.
(1992).Eichim. Biophys. Acta 1119,232-238.
Minchiotti, L., Galliano, M., Zapponi, M. C., and Tenni, R. (1993).Eur. J . Biochm. 214,
437-444.
Minghetti, P. P., Law, S . W., and Dugaiczyk, A. (1985).Mol. Eiol. Evol. 2,347-358.
Morinaga, T., Sakai, M., Wegmann, T. G., and Tamaoki, T. (1983).Proc. Nafl. Acad. Sci.
U.S.A. 80,4604-4608.
Moser, P., Squire, P. G., and OKonski, T. 0. (1966).J.Phys. Chem. 70,744-756.
Mudge, G . H., Desbiens, N., and Stibitz, G . R. (1978a).Drug Metab. Dispos. 6,432-439.
Mudge, G . H., Stibitz, G . R., Robinson, M. S., and Gemborys, M. W. (1978b).Drug Melab.
D~SPOS. 6,440-45 1.
Miiller, W. E., and Wollert, U. (1973).Naunyn-SchmieLberg's Arch. P h a m c o l . 280,229-237.
Miiller, W.E.,and Wollert, U. (1975).Mol. Phurmacol. 11, 52-60.
Nagase, S.,Shimamune, K., and Shumiya, S. (1979).Science 205,590-591.
Naik, D. V.,Paul, W. L., and Schulman, S. G. (1975).J.Pharm. Sci. 64, 1677-1680.
Nandedkar, A. K. N., Basu, P. K., and Friedberg, F. (1972).Bioinmg. Chem. 2, 149-157.
Nandedkar, A. K. N., Nurse, C. E.,and Friedberg, F. (1973).Int. J. Pept. Protein Res. 5,
279-28 1.
Nazareth, R.I., Sokoloski,T. D., Witiak, D. T., and Hopper, A. T. (1974).J.P h r m . Sci. 63,
203-211.
Nilsen, 0.G., and Jacobsen, S . (1976).Biochem. Phrmucol. 25, 1261-1266.
Oers, N. S. C., Cohen, B. L., and Murgita, R. A. (1989).J. Exp. Med. 170,811-825.
Okabayashi, K.,Nakagawa, Y., Hayasuke, N., et al. (1986).J.Biochem. (Tokyo) 100, 1533-
1542.
SERUM ALBUMIN STRUCTURE 20 1

Oncley, J. L., Scatchard, G., and Brown, A. (1947).J.Phys. ColloidChem. 51, 184-198.
Ozeki, Y., Kurono, Y., Yotsuyanagi, T., and Ikeda, K. (1980). C h . Pharm. Bull. 28,
535-540.
Parks, J . S., Cistola, D. P., Small, D. M., and Hamilton, J. A. (1983).J . Eiol. Chem. 258,
9262-9269.
Paubel, J.-P., and Niviere, P. (1974).Eur.J. Med. Chem. 9,508-512.
Peach, R. J., and Brennan, S. 0. (1991).Eiochim. Eiophys. Acta 1097,49-54.
Pearlman, W. H., and Crepy, 0. (1967).J.Eiol. Chem. 242, 182-189.
Pearson, W. R. (1990).In “Methods in Enzymology” (R. Doolittle, ed.), Vol. 183, pp. 63-98.
Academic Press, San Diego.
Pedersen, K. 0. (1971). Scand. J. Clin. Lab. Invest. 28,459-469.
Pedersen, K. 0. (1972). Scand. J. Clin. Lab. Invest. 29,427-432.
Pedersen, A. O., and Jacobsen, J. (1980). Eur. J. Eiochem. 106,291-295.
Perrin, J. H., Vallner, J. J., and Nelson, D. A. (1975).Eiochem. Phunnacol. 24,769-774.
Perutz, M. F. (1989). “Is Science Necessary?”Barrie and Jenkins, Ltd., London.
Perutz, M. F., Rossmann, M. G., Cullis, A. F., Muirhead, H., Will, G., and North, A. C. T.
(1960). Nature (London) 185,416-422.
Peters, T., Jr. (1975). In “The Plasma Proteins” (F. W. Putnam, ed.), 2nd ed., Vol. 1, pp.
133-181. Academic Press, New York.
Peters, T., Jr. (1980). “Serum Albumin-An Overview and Bibliography.” Miles Laborato-
ries, Elkhart, IN.
Peters, T., Jr. (1984). In “The Impact of Protein Chemistry on the Biomedical Sciences”
(A. N. Schecter and R. F. Goldberger, eds.), pp. 39-55. Academic Press, New York.
Peters, T., Jr. (1985).Adv. Protein Chem. 37, 161-245.
Peters, T., Jr. (1992). “Albumin: An Overview and Bibliography,” 2nd ed. Miles Inc.
Diagnostics Division, Kankakee, IL.
Peters, T., Jr., and Anfinsen, C. B. (1950).J.Eiochem. (Tokyo) 86,805-813.
Peters, T.,Jr., and Blumenstock, F. A. (1967).J.Eiol. Chem. 244, 1574-1578.
Peters, T., Jr., and Feldhoff, R. C. (1975).Biochemtsty 14,3384-3390.
Porta, F. A., Galliano, M., Rossi, A,, and Porta, F. (1990). Boll. Osp. Vurese 19, 197-210.
Quirk, A. V., Geisow, M. J., Woodrow, J. R., Burton, S. J., Wood, P. C., Sutton, A. D.,
Johnson, R. A., and Dodsworth, N. (1989). Eiotechnol. ApPl. E i o c h . 11, 273-287.
Rao, S. N., Basu, S. P., Sanny, C. G., Manely, R. V., and Hartsuck, J. A. (1976).J. Eiol. Chem.
251,3191-3193.
Reed, R. G . (I986).J.Eiol. Chem. 261, 15619-15624.
Richards, F. M. (1985).In “Methods in Enzymology” (H. Wyckoff el al., eds.), Vol. 115, pp.
440-464. Academic Press, Orlando, FL.
Richardson, K. S.C., Nowaczynski, W., and Genest, J. (1977).J.Steroid C h . 8,951-957.
Roda, A., Cappelleri, G., Aldini, R., Roda, E., and Barbara, L. (1982). J. Lipid Res. 23,
490-495.
Rosen, A. (1970).Eiochem. Phurmucol. 19,2075-2081.
Rossmann, M. G., and Blow, D. M. (1962). Acta Ctystallogr. 15,24-31.
Rothschild, M. A., Oratz, M., and Schreiber, S. S. (1988). Hepatology 8,385-401.
Russi, E.. and Weigand, K. (1983). Klin. Wochenschr. 61,541-545.
Sakamoto, Y., Davis, E., Madison, J., Watkins, S., McLaughlin, H., Leahy, D. T., and
Putnam, F. W. (1991). Clin. Chim. Acta 204, 179.
Sargent, T. D., Yang, M., and Bonner, J. (1981). Proc. Natl. Acad. Sci. U.S.A. 78, 243-
246.
Savu, L., Benassasyag, C., Vallette, G., Christeff, N., and Nuney, E. (1981).J.Eiol. C h .
256.9414-9418.
202 DANIEL C. CARTER AND JOSEPH X. HO

Sawinski, V. J., and Rapp, G. W. (1963).J.Dent. Res. 42, 1429-1438.


Scatchard, G. (1949). Ann. N . Y . Acad. Sci. 51,660-672.
Scatchard, G., and Yap, W. T. (1964).J.Am. Chem. SOC.86,3434-3438.
Schoentgen, F., Metz-Boutigue, M.-H.,Joll&s,J.. Constants, J., and Joll&s,P. (1986).Biochim.
Biophys. Acta 871, 189-198.
Sebille, B., Thuaud, N., and Tillement, J.-P. (1978).J. Chromatop. 167, 159-170.
Shaklai, N., Garlick, R. L., and Bunn, H. F. (1984).J.Biol. Chem. 259,3812-3817.
Sharpies, D. (1975).J.Phurm. Phurmocol. 27,379-381.
Shaw, C. F. (1989). Comments Inmg. Chem. 8(6), 233-267.
Shockley, P., and Brown, J. R. (1980).Fed. Proc., Fed. Am. SOC.Ex#. Biol. 39,2069.
Shrake, A., Finlayson,J. S., and Ross, P. D. (1984). VoxSung. 47,7-18.
Sijmons, P. C., Dekker, B. M. M., Schrammeijer, B., Venvoerd, T. C., van der Elzen,
P. J. M., and Hoekema, A. (1990). Biotechnology 8,217-221.
Sinn, H., Schrenk, H. H., Friedrich, E. A., Schilling, U., and Maier-Borst, W. (1990). Nucl.
Med. Biol. 17,8 19-827.
Sjoholm, I., and Ljungstedt, I. (1973).5.Biol. Chem. 248, 8434-8441.
Slayter, E M. (1965).J.Mol. Biol. 14,443-452.
Sleep, D., Belfield, G . P., and Goodey, A. R. (1991). BiolTechnology 8,42-46.
Sollene, N. P., and Means, G. E. (1979). Mol. Phurmocol. 14,754-757.
Sollene, N. P., Wu, H.-L., and Means, G. E. (1981). Arch. Biochem. Biophys. 207,264-269.
Spector, A. A., and Fletcher, J. E. (1978). In “Disturbances in Lipid and Lipoprotein
u.
Metabolism” M. Dietschy, A. M. Gotto, Jr., and J. A. Ontko, eds.), pp. 229-248. Am.
Physiol. Soc., Rockville, MD.
Spector, A. A., Santos, E. C., Ashbrook, J. D.,and Fletcher, J. E. (1973).Ann.N. Y. Acud. Sci.
226,247-258.
Squire, P. G., Moser, P., and OKonski, C. T. (1968).Biochemistty 7,4261-4272.
Stamler, J. S . , Singel, D. J., and Loscalzo,J. (1992). Science 258, 1898-1902.
Steiner, R. F., Roth. J.. and Robbins, J. (1966).J.Biol. Chem. 241,560-567.
Sudlow, G.,Birkett, D. J., and Wade, D. N. (1975).Clin. Exp. Phurmacol. Physiol. 2, 129-140.
Sudlow, G., Birkett, D. J.. and Wade, D. N. (1976). Mol. Pharmucol. 12, 1052-1061.
Sugita, O., Endo, N.. Yamada, T., Yakata, M., and Odani, S. (1987). Clin. Chim. A c h 164,
25 1-259.
Takabayashi, K., Imada, T., Saito, Y., and Inada. Y. (1983). Eur.J. Biochem. 136,291-295.
Takahashi, N., Takahashi, Y., Blumberg, B. S., and Putnam, F. W. (1987a).Proc. Nutl. Acud.
Sci. U.S.A. 84,4413-4417.
Takahashi, N., Takahashi, Y.,and Putnam, F. W. (1987b).Proc. Nutl. Acad. Sci. U.S.A. 84,
7403-7407.
Takahashi, N., Takahashi, Y.. Isobe, T., Putnam, F. W., Fujita, M., Chiyoko, S., and Neel, J.
( 1987~). Proc. Natl. Acad. Sci. U.S.A. 84,800 1-8005.
Tarnoky, A. L. (1980).Adu. Clin. Chem. 21, 101-146.
Teeter, M.M., and Hendrickson, W. A. (1979).]. Mol. Biol. 127,219.
Tillement, J.-P., Zini, R., DAthis, P., and Vassent, G. (1974). Eur. J. Clan. Phurmucol. 7,
307-313.
Tucker, E. M. (1968). Vox Sung. 15,306-308.
Unger, W. G. (1972).J.Phmm. Phurmocol. 24,470-477.
Vermeulen, A., and Verdonck, L. (1968). Steroidc 11,609-635.
Waldmann, T. A. (1977). In “Albumin Structure, Function and Uses” (V. M. Rosenoer, M.
Oratz, and M. A. Rothschild, eds.), pp. 255-273. Pergamon, Oxford.
Waldmann-Meyer, H. (1960).J.Biol. Chem. 235,3337-3345.
Walker, J. E. (1976). FEBSLett. 66, 173-175.
SERUM ALBUMIN STRUCTURE 203

Weinstock, J., and Baldwin, G. S. (1988).Nwleic Act& Res. 16(18),9045.


Westphal, U., and Harding, G. B. (1973).Eiochim. Ei0phys. Acta 310,518-527.
Whitlam, J. B.,Crooks, M. J., Brown, K. F., and Pedersen, P. V.(1979).BiOchem. Phurmacol.
28,675-678.
Wilting,J., Kremer, J. M. H., Ijzerman, A. P., and Schulman, S. G. (1982).Biochim. Eiophys.
Acta 706,96-104.
Winter, W. P., Weitkamp, L. R.,and Rucknagel, D. L. (1972).Eiochnnishy 11,889-896.
Wright, A. K., and Thompson, M. R. (1975).Ei0phys.J. 15, 137-141.
Yang, F.,Brune, J. L., Naylor, S. L., Cupples, R. L., Naberhaus, K. H.,and Bowman, B. H.
(1985).Proc. Natl. Acad. Sci. U.S.A. 82,7994-7998.
Yang, F., Bergeron, J. M., Linehan, L. A., Lally, P. A., Sakaguch, A. Y.,and Bowman, B. H.
(1990).Genomics 7(4),509-516.
Yates, F. E., and Urquhart, J. (1962).Physiol. Rev. 42,359-443.
Yeh, P., Landais, D., LeMaitre, M., Maury, I., Crenne, J.-Y., Becquart, J.. Murry-Brelier,
A., Boucher, F., Montay, G., Fleer, R.,Hirel, P.-H., Mayaux, J.-F., and Klatzmann, D.
(1992).PTOC.Natl. Acad. Sci. U.S.A. 89, 1904-1908.
Yvon, M., and Wal, J.-M. (1988).FEES Lett. 239,237.
Zurakowski, R.,Jr., and Foster, J. F. (1974).Eiochemitty 13,3465-3471.
Zvelebil, M.J. J. M., Barton, G. J., Taylor, W. R., and Stemberg, M. J. E. (1987).J.Mol. Biol.
195,957-961.

You might also like