вход по аккаунту


Estimating mutation rates using abnormal human hemoglobins.

код для вставкиСкачать
Estimating Mutation Rates Using Abnormal Human
Departments ofAnthropology (RE.N.1 and Medicine (Diuision of Medical
Genetics) (G.S., IIE.N.), University of Washington, Seattle, Washington
Problems in estimating mutation rates, Mutation rates per
nucleotide, Unstable hemoglobin disease, Hemoglobin M disease
Estimates of mutation rates in human populations vary
widely. Considering the numerous potential sources of error, which include
illegitimacy, the existence of genocopies andor phenocopies, variation in
expressivity, and incomplete penetrance, the extent of variation among estimates is not surprising. When mutational events are defined in detail at the
level of DNA sequence, one can estimate rates at which nucleotide substitutions give rise to the abnormal phenotypes under consideration. The vast
body of data now available on globin-gene sequence and organization, as well
as on abnormalities in amino acid sequences of globin chains, renders certain
dominantly inherited hemoglobinopathies especially well suited to analyses
of mutation rates. Using cases of unstable fi-chain disease and Hb M disease
that appeared among children of unaffected parents, rates at which these
conditions arise by new mutation were estimated, a s were rates of substitution per a- or @-genenucleotide and per a gene (423 nucleotides) or @ gene
(438 nucleotides). These estimates, and the differences between them and
others calculated using sporadic cases of various phenotypically defined conditions, are discussed in relation to the biases to which they are subject.
Much of the genetic novelty necessary to evolutionary change arises through
mutations, both cytogenetic (structural rearrangements of chromosomes) and locusspecific (e.g., nucleotide substitutions in exons of protein-encoding genes). Thus,
efforts to measure rates at which mutations occur in various organisms have long
occupied students of evolutionary processes. Studies of mutation in human subjects
have focused primarily upon the extent to which newly mutated genes contribute to
the incidence of genetic disease and, hence, to the genetic burdens carried by human
populations (Trimble and Smith, 1977). Since the 1950s, increasing public concern
over the genetic effects (on both somatic and gametal cells) of radiation and chemical
mutagens released into the environment by human activity has given voice to fears
that increased levels of such agents constitute a collective threat to the genetic wellbeing of any or all forms of life (Trimble and Smith, 1977; Denniston, 1982). Crow
(1981:13) clearly expressed these fears in the following passage:
“I am not sure what the optimum mutation rate for the human species
is, when viewed from a very long perspective of millions of years of past
evolution and (if we are lucky) for similar periods in the future. However,
it is clear to me that from the standpoint of immediate human welfare
the mutation rate is too high. There is already a large store of genetic
variation, enough to satisfy even the most wild-eyed eugenicist. If the
mutation rate were to drop to zero, we would not notice any decreased
0 1984 Alan R. Liss. Inc.
IVol. 27, 1984
variability for hundreds or thousands of generations, except for the disappearance of some of the worst dominant diseases. If we ever find out in
the future that we need more mutations, that is one thing we do know
how to produce.”
That the incidence of hereditary disease in the human population is high is
without question. For example, in a study of children born with nonchromosomal
hereditary disease in British Columbia, Canada, Trimble and Doughty (1974) estimated that 2.3 per 1,000 liveborn infants suffered from an autosomal dominant,
autosomal recessive, or X-linked disorder, while 47.3 per 1,000 suffered from irregularly inherited abnormalities. This represented a total of 4.96% of all liveborn
children. An additional 42.8 per 1,000 children were severely afflicted with congenital malformations. Presumably, some unknown proportion of these cases of malformation was of genetic etiology.
Trimble and Doughty (1974) provided data of value in estimating the incidence of
nonchromosomal genetic disease in human populations, while Trimble and Smith
(1977) presented guidelines to follow in assessing the burden of genetic ill-health
upon society at large. Yet, the rates a t which mutations arise in maintaining this
burden remain subjects of considerable speculation.
If an estimate of a mutation rate is to be judged reliable, there must be grounds
for assuming that, in the population under consideration, all cases of the abnormality in question have been detected. Autosomal dominant abnormalities are most
frequently used in estimating mutation rates. Calculation of the rates at which
many such conditions arise are generally straightforward, the mutation rate being
the number of affected children born during a given interval to unaffected parents,
divided by twice the number of births during the same interval. Application of this
direct method is subject to several kinds of error. Vogel and Rathenberg (1975) list
as potential sources of error illegitimacy, the existence of phenocopies (nonhereditary traits that closely mimic those associated with genetic abnormalities), genocopies (phenotypically similar, but genetically different, traits), and incomplete
penetrance (failure of a n abnormal allele to produce detectable phenotypic effects in
all carriers). To this list can be added variation in expressivity and a special case of
variable expressivity, variable age of onset. Neurofibromatosis, a n autosomal dominant disease, ranges in severity from the presence of a few pigmented spots in the
skin to the malignant degeneration of multiple, subcutaneous neurofibromas (McKusick, 1983). In attempting to identify cases of this disease among children of
unaffected parents, care must be exercised to ensure that some of the least severely
afflicted subjects do not escape detection. Huntington chorea is a classic example of
an autosomal dominant disease whose signs first appear in different subjects a t ages
ranging from less than 10 to over 75 years (Vogel and Motulsky, 1979). Since some
heterozygotes will die from other causes before the choreic movements andor progressive dementia characteristic of the disease appear, cases arising by new mutation may well be missed.
Of the potential sources of error listed above, the inclusion of genocopies andor
phenocopies among the sporadic cases of new mutation will lead to overestimation
of mutation rates, as will the inclusion of illegitimate children whose biological
fathers, unbeknownst to the investigator, transmitted copies of the abnormal allele
to the probands. Variation in expressivity could, in principle, lead to either underor overestimation of mutation rates, depending upon whether or not expression was
minimal (and thereby overlooked) in heterozygotes born to genetically normal parents, or minimal (and likewise overlooked) in a parent of an obviously affected child.
On balance, we judge the former problem to be of greater significance than the
latter, simply because identification of a proband with a severe genetic disease of
variable expressivity commonly encourages the careful conduct of familial studies.
Nute and Stamatoyannopoulos]
Clearly, the direct method of estimating mutation rates has shortcomings, but so
do the indirect methods. One indirect approach, which is commonly used to estimate
rates a t which phenotypes of clinical significance arise by new mutation, requires
the assumption that the frequency of a n abnormal allele is maintained at a nearly
constant level from generation to generation in the population. Thus, the rate a t
which selection acts to remove the allele from the gene pool must be balanced by the
rate a t which new copies are introduced by mutation. An accurate measurement of
the fitness of affected subjects relative to that of normal members of the population
is necessary to a reliable calculation of the mutation rate. With few exceptions,
however, estimates of fitness are subject to considerable error. Genocopies andor
phenocopies, variable expressivity, incomplete penetrance, modern therapeutic procedures, and genetic counseling can introduce biases that become more critical as
the fertility of affected subjects more closely approaches the average for the population. Recognizing the difficulties imposed by the variables noted above, and because
there is considerable variation in the reproductive patterns of “normal” individuals,
Vogel and Rathenberg (1975) concluded that this indirect approach can provide but
a general notion of the order of magnitude of a mutation rate even when the fertility
of affected individuals equals or is close to zero.
A second indirect approach relies upon detection of “private” alleles (those limited
to one or a few families in a population) that encode the amino acid sequences of
electrophoretically variant proteins (Neel and Rothman, 1981). In applying this
procedure, the potential effects of genetic drift must be taken into account, especially
when dealing with small populations and alleles judged selectively neutral or nearly
so. Further, one must assume that the population in question has long been free of
immigration. In the case of a variant allele associated with decreased fitness, loss of
that allele through selection might be countered by gain through immigration;
under these conditions, the two most apparent sources of bias, being opposite in
direction, might produce little net effect (Neel and Rothman, 1981).
If one’s goal is the estimation of a mutation rate by application of the direct
method, the kind of mutation and nature of its phenotypic manifestations are of
primary importance. The abnormal phenotype should be traceable to mutation at a
particular locus, and inherited as a simple, autosomal dominant condition. Penetrance should be complete, and the phenotype should be expressed throughout
postnatal life. (If expressed prenatally, it should constitute little threat to fetal
survival.) Under these conditions, it should be possible to closely approximate the
rate at which a particular abnormal phenotype arises as a consequence of mutation
a t a particular locus. Still, the meeting of these criteria is no guarantee that the
estimate will reflect the overall rate at which a gene mutates, since mutations
leading to quite different phenotypic consequences may not be traced to the locus in
question. If we add the condition that each case of the abnormality should be
attributable to a specific nucleotide substitution in a gene of known sequence, the
overall rate of nucleotide substitution per gene per generation can be calculated.
The conditions outlined above are stringent indeed, and it is hardly remarkable
that few estimates of rates of substitution per nucleotide have been calculated for
any human gene. By 1979, the wealth of data accumulated on normal and abnormal
human hemoglobins and the genes encoding the amino acid sequences of their
constituent polypeptide chains suggested to us that analyses of this type could
generate fairly reliable results.
The human globin genes
The human globin genes are arranged in two clusters: the 0-like genes on chromosome 11(Deisseroth et al., 1978) and the d i k e genes on chromosome 16 (Deisser0th et al., 1977). The @-likegenes appear in the order 5’-$@2-~-Gy-Ay-$@1-6-@-3’,
series that includes five active globin genes k,Gy, *Y, 6, P) and two unexpressed
pseudogenes ($@l,$02) (Fritsch et al., 1980).The array of {-like genes, 5’-{2-+{1-$al-
[Vol. 27, 1984
a2-al-3' (Lauer at al., 1980; Little, 1982),contains three active genes ( { Z , a l , a2) and
two pseudogenes (${l,Gal),The 5' to 3' orientation of the functional genes in both
clusters reflects the sequences in which the genes are expressed during ontogenesis.
2 1 , Portland ({2y2),
The embryonic hemoglobins, Hbs Gower 1({Ztz), Gower 2 ( ~ ~ 2 ~ and
are produced exclusively or nearly so, in the yolk sac (Clegg and Gagnon, 1981). The
and y chains are produced later in life as well, but synthesis of { and E chains
predominates in the earliest stages of erythropoiesis. The Ay and Gy chains combine
with a chains to form fetal homoglobin (Hb F: a2y2), the major component in
erythrocytes of the fetus. The p and 6 chains appear a t low levels in fetal erythrocytes
(wherein they constitute about 5% of the 0-like chains). The switch from production
of fetal to that of adult hemoglobins takes place during the perinatal period, and by
the sixth postnatal month, Hb A (a2&) and Hb A 2 (~~262)
constitute about 97% and
2.5%, respectively, of the hemoglobin in normal subjects (Stamatoyannopoulos and
Nute, 1974). Fetal hemoglobins, containing Ay and Gy chains, are found throughout
adult life in a small proportion of erythrocytes (F-cells) (Wood et al., 1975). In
addition, secondarily modified hemoglobins, e.g., Hb AI,, which contains p chains
with glycosylated amino-terminal residues (Bunn et al., 19751, are also found in red
cells from normal adults.
Analyses of the primary structures of the normal human globin chains have
established that a-like chains are 141 amino acid residues and the p-like chains are
146 residues in length. The coding sequences of the a-like (423 nucleotides) and plike (438 nucleotides) genes have also been analyzed in full (Liebhaber et al., 1980,
1981; Slightom et al., 1980; Spritz et al., 1980; Lawn et al., 1980; Baralle et al., 1980;
Cohen-Solal et al., 1982).
Specification of the amino acid substitution appearing in a globin chain as a
consequence of a nucleotide substitution permits, with few exceptions, specification
of that nucleotide substitution. Not all globin-chain loci are, however, appropriate
objects of analysis. Embryonic material is not readily acquired, and it is likely that
E - and {-chain variants with deleterious effects will so compromise development that
few will be detected. Moreover, confirmation of a suspected case of new mutation
requires that a proband's parents not carry the abnormal allele. Short of structural
analysis of the E or {genes carried by the parents, in whom embryonic chains are no
longer synthesized, confirmation of a new mutation would be impossible. Detection
of mutant y chains is subject to similar, if less severe, constraints. Of the globin
chains found in adults, the 6 chain is found at such low levels that structural
anomalies are not apt to produce outward signs of clinical significance. This leaves
only the a and fi chains, several variants of which are known to produce persistent
symptoms of sufficient severity to prompt clinical examination (Table 1).
Abnormal human hemoglobins
Since Ingram (1959)identified the structural abnormality in the p chains of Hb S,
the list of structurally different variants of human hemoglobins has lengthened
steadily, reaching 318 (Table 1) by the time we initiated our efforts to assess
mutation rates. Significantly, over 50% of the variants contained abnormal chains
and over 25% abnormal a chains that had sustained single amino acid substitutions.
With very few exceptions, each of these substitutions is unambigously attributable
to the replacement of one particular nucleotide by another in a n a- or p-globin gene.
TABLE 1. Human globin-chain variants'
Abn or ma 1
or 012)
01 (011
(*r or
or €
Total No.
of variants
Hbs M and
unstable Hbs
Other functionally
abnormal variants
'From International Hemoglobin Information Center (1979a-c)
Nute and Stamatoyannopoulos]
Although the majority of these variants (174) show no signs of functional abnormality, there remain several whose structural anomalies lead, in heterozygotes, to
physiological dysfunction of clinical significance. Most of the latter variants have
been characterized as unstable hemoglobins, methemoglobins, and hemoglobins
with unusually high or low affinities for oxygen.
Innocuous variants of no clinical import are typically detected while electrophoretically screening the hemoglobins from large numbers of people. Given the rarity
of carriers of such innocuous variants, only a small proportion of the total number
present in a large population are likely to be detected unless the screening program
is of truly heroic proportions. Conversely, abnormal hemoglobins of clinical significance, although similarly rare, are quite likely to be detected because affected
subjects seek medical aid.
Most of the unstable hemoglobins observed to date result from single amino acid
substitutions in their CY or p chains. The unstable chain precipitates within and
damages the erythrocyte, leading ultimately to lysis of the cell. Heterozygotes for
a n unstable chain suffer chronic hemolytic anemia that is readily attributable,
through tests of stability in vitro, to a n abnormal hemoglobin.
Any of seven, specific, amino acid substitutions, two in each of the structurally
identical LY chains encoded at the a1 and 012 loci and three in the /3 chain, can produce
a methemoglobin, the chief manifestation of which is cyanosis. Definitive diagnosis
requires identification of the structural defect in the chain (Stamatoyannopoulos et
al., 1976).
Hemoglobins with high affinities for oxygen do not release adequate amounts of
oxygen to the tissues, the ultimate result being development of a compensatory
erythrocytosis. Detection of the abnormal hemoglobin can be difficult because many
variants of this type are normal in electrophoretic and chromatographic behavior.
Moreover, definitive diagnosis requires analysis of oxygen-binding properties and
determination of the structural abnormality (Stamatoyannopoulos et al., 1973; Nute
et al., 1974). Finally, familial erythrocytosis is an important criterion for ascertainment of other hematological anomalies (McKusick, 1983). Hence, cases of hemoglobins with high affinities for oxygen are at risk of misclassification. To date, only one
case of such a variant (Hb Bethesda) appears to have stemmed from a new mutation
(Bunn et al., 1972).
Hemoglobins with low affinities for oxygen are typically associated with a mild,
normocytic anemia. In some instances, cyanosis is the most obvious manifestation.
Owing to the mildness of the anemia, many affected individuals are likely to escape
detection. Very few cases of such hemoglobins have been described in the literature,
and we are aware of only two that appear, on the basis of limited evidence, to be
products of new mutations. Both are cases of Hb Beth Israel, a variant associated
with clinically apparent cyanosis (Nagel et al., 1976; Efremov et al., 1978).
The four types of conditions described above are inherited as autosomal dominant
traits but, given the problems involved in establishing the genetic etiologies of
individual cases of normocytic anemia and erythrocytosis, only cases of unstable
hemoglobin disease and hemoglobinopathic methemoglobinemia were judged suited
to studies of mutation rates.
Cases of unstable hemoglobin disease and methemoglobinemic cyanosis were
ascertained by screening the relevant literature published from 1950 through 1982.
When there were grounds for suspecting that the parents of a proband were unaffected, the case was considered further. In many publications, it was merely stated
that no signs of hemoglobinopathy were evident upon examination of the proband's
parents andor siblings. Authors of these papers were contacted and requisite familial data, including dates of probands' births and, where available, results of tests of
paternity were acquired.
By the end of 1982, data bearing on 66 cases of new mutation were in hand. Of
these, 48 cases of unstable hemoglobin disease and 16 of Hb M disease were attrib-
[Vol. 27, 1984
utable to substitutions of single nucleotides in a-or P-globin genes. The cases of Hb
Freiburg (Jones et al., 1966) and Hb Gun Hill (Murari et al., 1977) were not included
in subsequent calculations of mutation rates because these unstable variants derive,
not from nucleotide substitutions, but from deletions of one or more of the codons
specifying the primary structures of their 0 chains. A case of Hb Istanbul (Aksoy et
al., 1977) and one of Hb Perth (= Abraham Lincoln) (GrovB et al., 1977) were also
omitted from consideration because data on the numbers of births in the probands'
countries of origin (Turkey and Namibia, respectively) were unavailable. Finally,
single cases of the unstable variants Hb Bristol (Y. Ohba and M. Sakuragawa, pers.
commun.) and Hb Hammersmith (Schroeder et al., 1980; Schroeder, pers. commun.),
and the methemoglobins Hb M Saskatoon (D. Labie, pers. commun.) and Hb M
Boston (Hayashi and Yamamura, 1964) were excluded because the probands were
born prior to 1925 or after 1974 (see below).
Of the 58 cases judged appropriate to the present analysis, 51 were compiled from
the literature. Authors who were asked to provide information in addition to that
presented in published reports also furnished unpublished data on the seven other
cases. Paternity had been tested in 20 of the 58 cases, and in no instance was
paternity excluded. The six cases unknown to us earlier (Stamatoyannopoulos and
Nute, 1982) are listed in Table 2, while the 58 cases are summarized in Table 3.
Since there are twice as many a as 0 genes in the normal human genome, one
might be tempted to assume that the scarcity of probands with newly mutated a
genes (five) relative to the number with newly mutated 0 genes (53) reflects far
lower rates of mutation at the a loci. There is, however, a more plausible explanation
for the discrepancy. As originally proposed by Ingram (1963),mutations affecting a
chains (which are produced at high levels during fetal development) are more likely
to imperil the fetus than are mutations affecting /3 chains (which attain high levels
after parturition). This suggestion is indirectly supported by the figures in Table 1.
Of the 318 variants known in 1979, 192 contained abnormal 0 chains, as opposed to
99 with abnormal a chains. While 60% of the 0-chain variants were functionally
abnormal, only 28%of the a-chain variants were so. Our failure to find any cases of
unstable a-chain disease that arose by new mutation may simply indicate that
unstable hemoglobin disease, when expressed prenatally, jeopardizes fetal development to the extent that relatively few fetuses survive to term.
The discrepancy between the numbers of probands with aM disease (five) and OM
disease (nine) provides a lesser measure of support for Ingram's (1963)suggestion. It
is lesser because of the small number of cases discovered and because the number of
nucleotide substitutions per gamete that can produce a n aM gene (four, two in each
a gene) differs little from the number that can produce a PM gene (three in one gene).
Thus, in a large population, one would expect to find not twice a s many cases of aM
TABLE 2. Cases of new mutations not included in previous studies'
Case NO.^
Year of birth3
Unstable Hbs
ca. 1945
ca. 1972
Data acquired from:
Adams et al. (1979)
Johnson et al. (1980),
Schroeder (pers. commun.)
Rousseaux et al. (1980)
Iuchi et al. (1979),
Iuchi (pers. commun.)
Y. Ohba (pers. commun.)
Maggio et al. (19811,
Maggio (pers. commun.)
'Stamatoyannopoulos et al. (1981);Nute and Stamatoyannopoulos (1981);Stamatoyannopoulos
and Nute (1982).
'Each case is designated by the country of origin and a number; the numbers indicate the order
in which cases originating in a given nation became known to us.
3Years of birth for cases USA-11and France-8 were estimated by subtracting the probands' ages
(given in the publications cited) from the dates of publication.
TABLE 3. Cases of unstable Hb and H b M diseases used in estimating mutation rates'
Year of
Unstable hemoglobins
ca. 1945
New Zealand-l
Hemoglobins M
ca. 1972
W. Germany-101
New Zealand-101
Santa Ana
Abraham Lincoln
Bryn Mawr
(= Bucuresti)
I Toulouse
St. Etienne
(= Istanbul)
St. Louis
Hammer smith
Abraham Lincoln
Santa Ana
Abraham Lincoln
M Milwaukee-1
M Kankakee
( = M Iwate)
M Hyde Park
M Saskatoon
(= M Chicago)
M Saskatoon
M Boston
M Boston
M Iwate
M Saskatoon
M Boston
M Saskatoon
M Hyde Park
M Saskatoon
(= M Kurume)
M Saskatoon
Amino acid
088 Leu-Pro
0141 Leu-Arg
091 Leu-Pro
898 Val-Met
p24 Gly+Val
0106 Leu-Pro
0106 Leu-Pro
032 Leu-Pro
081 Leu-Arg
085 Phe-Ser
0112 Cys-Arg
042 Phe-Leu
Change in
DNA codon
028 Leu-Pro
042 Phe Ser
042 Phe-Ser
098 Val-Met
067 Val-Asp
p106 Leu-Pro
p98 Val+Gly
066 Lys-Glu
092 His-Gln
or A A ~
GTG- G T ~
028 Leu-Gln
028 Leu-Pro
063 His-Pro
042 Phe-Ser
098 Val-Met
032 Leu-Pro
032 Leu-Arg
0115 Ala-Pro
088 Leu-Pro
6106 Leu-Pro
027 Ala-Asp
071 Phe-Ser
027 Ala-Asp
098 Val-Met
098 Val-Met
068 Leu-Pro
042 Phe-Ser
0142 Ala-Pro
632 Leu-Pro
028 Leu-Pro
898 Val-Met
8117 His-Pro
~ ~
067 Val-Glu
a87 His-Tyr
092 His-Tyr
063 His-Tyr
063 His-Tyr
or58 His-Tyr
or58 His-Tyr
a87 His-Tyr
063 His-Tyr
a58 His-Tyr
063 His-Tyr
092 His-Tyr
063 His-Tyr
663 His-Tyr
'Sources of data on all cases not listed in Table 2 are cited in Stamatoyannopoulos et al. (1981) and Stamatoyannopoulos and Nute
(1982). Three cases (Japan-4 and 102, and Portugal-101) listed in these references do not appear above because the probands were
born outside the interval from 1925 through 1974.
'ABO, Rh only.
(Val. 27, 1984
as OM mutation, but roughly equal numbers of the two types. This expectation is
not, however, met. Calculations based upon the five cases of aM disease yield
mutation rates per gene and per nucleotide per generation that are approximately
half those calculated using the nine cases of OM disease (see below), results suggestive of a n adverse effect of aMdisease on fetal development.
Under the assumption of complete ascertainment, mutation rates can be directly
estimated as the number of heterozygotes for new mutations born during a specified
interval in a country, divided by twice (for rates of P-gene mutation) or four times
(for rates of a-gene mutation) the number of births that occurred in the country
during the same interval. We selected two intervals, each roughly equivalent to a
generation, the first spanning the years 1925-1949 and the second 1950-1974. Thus,
if a single proband were found in a country, the total number of births that occurred
in that country during the generation of the proband’s birth would be used in
estimating the mutation rate. Similarly, the numbers of births that occurred over
all generations and countries in which probands with unstable hemoglobin disease,
aM disease, or PM disease were born are summed to yield the sizes of the total
populations in which mutations of each type arose (Table 4).
Rate of appearance of unstable hemoglobin disease by mutation
The 44 probands with unstable hemoglobin disease appeared in ten countries
among a total of 520.4 x lo6 births (Table 4). Because all cases are attributable to
substitutions of single nucleotides in P-globin genes, the estimated rate a t which
this disease arises by nucleotide substitution in P genes is 4442 x 520.4 x lo6), or
4.2 x 10~slgenelgeneration.
Rates of appearance of OM and aM diseases by mutation
Nine cases of PM mutants arose among 255.9 x lo6 children born in six nations
(Table 4). Thus, the rate a t which OM disease arises through substitution of single
nucleotides is expressed as 942 x 255.9 x lo6), or 1.8 x 10-s/P genelgeneration.
TABLE 4. Sizes ofpopulations, by nation and generation, ofprobands’ births’
No. of
Cases of unstable hemoglobin
New Zealand
(All cases)
Cases of OM disease
New Zealand
(All cases)
Cases of aM disease
W. Germany
(All cases)
Range of probands’
years of birth
Total births
1925 through
1950 through
1925 through
1925 through
1925 through
1950 through
1950 through
1950 through
1950 through
1925 through
1925 through 1974
1950 through 1974
1950 through 1974
1950 through 1974
1950 through 1974
1950 through 1974
1950 through 1974
‘Source of data on birthsigeneration are cited in Stamatoyannopoulos and Nute (1982).
Nute and Stamatoyannopoulos]
Five probands with aM chains were found among 156.8 x lo6 individuals born in
five countries (Table 4). There being four a genes per person, the rate at which Hb
M disease arises as a consequence of nucleotide substitution in a n a gene is estimated a s 544 x 156.8 x lo6), or 0.8 x 10-slgene/generation.
Mutation rates per nucleotide per generation
In all but two of the 58 cases, it was possible to specify the nucleotide substituted.
The case of Hb Louisville (Canada-l), a n unstable variant, could have arisen by
replacement of either the first or third nucleotide in the codon that specifies insertion
of phenylalanine in position 42 of the normal 0 chain (i,e., AAA could be altered to
AAT, AAC, or GAA in effecting the substitution of phenylalanine by leucine at @42
in Hb Louisville). The case of Hb Bristol (UK-4) derives from substitution in a codon
that is itself variable in some human populations (Forget, 1977).Thus, substitution
of either one or, less likely, two nucleotides, could have produced the change from a
codon for valine (CAA or CAG), the residue occupying position 67 in the normal @
chain, to a codon for aspartic acid (CTA or CTG). There being no evidence to the
contrary, we assume that this case stems from a single nucleotide substitution.
When a single case of a particular @-chainvariant appears among a total of 520.4
x lo6 individuals, only one of 1,040.8 x lo6 copies of a given nucleotide has been
replaced in producing the underlying mutant gene. Because one nucleotide can be
substituted, presumably with equal likelihood, by any one of three others, the
overall rate of substitution of the nucleotide in question would be 31(1040.8 x lo6),
or 2.9 x 10-g/generation. Thus, each unstable variant appearing but once in Table
3 signifies that a particular nucleotide is substituted a t this rate. When considering
variants that appear twice or more (the six cases of Hb Koln, four each of Hbs Casper
and Hammersmith, three each of Hbs Genova and Abraham Lincoln, and two each
of Hbs Santa Ana and Volga), the mutation rates per nucleotide would be 6 , 4 , 3, or
2 x 2.9 x lO-’/generation, respectively. Summing the rates calculated over each of
the 27 structurally different unstable hemoglobins found among the 44 cases of
unstable hemoglobin disease, and dividing the sum by the number of different sites
of nucleotide replacement (25) yields the rate (an average over the 25 sites) of 5.1 x
lo-’ substitutionslnucleotidelgeneration. More simply, this result can be obtained
by solving the equation p = 3n/2Nx, where p = rate of nucleotide substitution, 3 =
the number of ways one nucleotide can be replaced by another, n = 44 (cases of
unstable @-chaindisease), 2N = 1,040.8 x lo6 (twice the population size), and x =
25 (number of different sites a t which nucleotide substitutions occurred in producing
the 44 cases).
Applying the same approach to the nine cases of PM disease arising among 255.9
x 106 individuals, where x = 3, the estimated rate is 17.6 x lO-’l@-gene nucleotidel
generation. It is not especially noteworthy that this rate exceeds that calculated
using cases of unstable hemoglobin disease by a factor of nearly 3.5, given that one
rate is the average over 25 nucleotide-bearing sites and the other is a n average over
only three sites. Nonetheless, the two estimates differ by far less than a n order of
magnitude. A more reliable estimate of 5.9 x lo-’ is calculated when all cases of
unstable @-chaindisease and PM disease are combined (n = 53,2N = 1,044.2 x lo6,
x = 26).
Finally, the rate of substitution per a-gene nucleotide is given as p = 3n/4Nx,
where n = 5,4N = 627.2 x lo6, and x = 2. The result is 12.0 x lo-’ substitutions1
a-gene nucleotidelgeneration.
Nucleotide substitutions per gene per generation
One need only multiply the rate of substitution per nucleotide by the number of
nucleotides in the coding portion of a 0 gene (438) or a n a gene (423) to derive a n
estimated rate of substitution per gene. Thus, the rate per (3 gene, based upon cases
of unstable hemoglobin disease is 438(5.1 x lo-’), or 2.2 x lop6 substitutions/
generation. Based on cases of bd disease, the rate is 438(17.6 x lo-’), or 7.7 x lop6
substitutions/@genelgeneration. The rate estimated using all cases involving abnor-
[Val. 27, 1984
ma1 /3 chains is 438(5.9 x lo-’), or 2.6 x lop6 nucleotide substitutionslP gene/
generation. Similarly, the rate per a gene is 423(12.0 x lo-’), or 5.1 x lop6
nucleotide substitutionstgeneration.
The mutation rates presented above (and summarized in Table 5) are direct
estimates. That estimated rates per nucleotide per generation, as well as those per
gene per generation, differ by no more than a factor of 3.5 suggest that, if bias has
been introduced, it is similar in extent and direction over all estimates. Looking at
the issue from a slightly different perspective, the rates of appearance by mutation
of 0-chain instability in those nations in which the largest numbers of probands
were found (U.S.A., U.K., France, Japan) differ little, ranging from 3.6 x lo-’ to
10.4 x lO-’lfi gene/generation (in the U.S.A. and France, res ectively). The rate
over all four countries (n = 34, 2N = 692.2 x lo6) is 4.9 x 10- Ifi genetgeneration,
a figure that differs little from that calculated using all 44 cases of unstable hemoglobin disease (Table 5). Perhaps, then, a consistent bias has been introduced or,
alternatively, there has been a mutual cancellation of upward by downward biases.
Potential causes of overestimation of mutation rates
It is possible, but we think unlikely, that the mutation rates given in Table 5 are
unduly high because the numbers of individuals (N) used in the calculations are too
small. These numbers represent only those people born in countries and generations
of probands’ births. Failure to include populations in which no cases of new mutation
have been detected (and it may be argued that specific mutational events occur so
infrequently that none will be observed in many groups) could result in the calculation of unrealistically high mutation rates. Hook and Porter (19811, recognizing the
potential effects of such a bias in ascertainment, proposed that estimates of mutation
rates for unstable hemoglobin disease and Hb M disease be founded upon numbers
of births in 17 developed nations, nations in which ascertainment of these conditions
is most likely to be complete. In these countries, roughly 427 x lo6 births occurred
between the beginning of 1925 and the end of 1974. In six of the countries (U.S.A.,
Canada, U.K., France, New Zealand, Japan), 38 probands with unstable hemoglobin
disease have been found; in the remaining 11 nations (Austria, Belgium, West
Germany, Ireland, Switzerland, Luxembourg, The Netherlands, Denmark, Sweden,
Finland, Norway), no cases of this disease attributable to new mutation have been
described. Using these figures, the rate at which unstable hemoglobin disease arose
is estimated as 38/(2 x 427 x lo6), or 4.4 x 10-8/P genetgeneration. Based upon the
same set of data, the estimated rate per P-gene nucleotide (x = 22) is 6.1 x lo-’/
generation, while that over all 438 nucleotides encoding the P-chain sequence is 2.7
x 10-6//3 genetgeneration. Similarly, the estimates based upon the eight cases of PM
disease found in these 17 countries are 0.9 x lO-’/P genetgeneration (rate of appearance of the disease), 9.4 x 10-9/@-genenucleotidelgeneration (x = 31, and 4.1 x
TABLE 5. Mutation rates calculated using cases of unstable H b and H b M diseases
Rate of mutation producing:
Unstable Hb disease
DM disease
olM disease
Mutation rate per nucleotide, based upon cases of:
Unstable Hbs
pM variants
OM and unstable variants
Mutation rate per gene, based upon cases of:
Unstable Hb disease
bM variants
BM and unstable variants
4.2 x 10-8/p geneigeneration
1.8 x lO-’/fi gene/generation
0.8 x lO-’/ol geneigeneration
5.1 x
17.6 x
12.0 x
5.9 x
2.2 x
7.7 x
5.1 x
2.6 x
10-6/438 nucleotides/generation
10-6/438 nucleotides/generation
10-6/423 nucleotides/generation
10-6/438 nucleotidesiaeneration
Nute and Stamatoyannopoulos]
10-6/438 P-gene nucleotides/generation. Finally, the rate at which Hb M disease
arose, as a consequence of a-gene mutation, in these 17 nations is 0.2 x 10-a/a gene/
generation, while the rate per a-gene nucleotide is 3.5 x 10-g/generation (x = 2)
and that per 423 a-gene nucleotides is 1.5 x 10-6/generation. Of the rates presented
in Table 5, only those based upon cases of aM and PM disease are higher than those
calculated as suggested by Hook and Porter (1981). The differences (1.9- to fourfold)
are, however, slight. Furthermore, the three sets of rates based upon cases of
unstable hemoglobin disease are virtually identical. On these grounds, our failure
to include populations in which no probands appeared does not seem to have unduly
inflated our estimates.
Clearly, the approach recommended by Hook and Porter (1981) is founded upon
the long-established epidemiological principle that groups of subjects at risk be
selected without foreknowledge that the condition under investigation is present in
or absent from one or more of them. In the present study, this approach would be
mandated if potential probands from populations selected in computing N were
subject to analysis to the extent required for inclusion in our set of cases. Unfortunately, cases of Hb M and unstable hemoglobin disease are far less likely to be
brought to light in some areas of the world than in others (e.g., we have found no
cases from South or Central America, China, India, or most of Africa), not because
these conditions are less likely to arise in these areas, but because the requisite
clinical and laboratory analyses have not been conducted. We feel that the risk of
underestimation of mutation rates, if we included populations in which no probands
have been found, could easily exceed the risk of overestimation that arises from the
manner in which we have determined values of N. Obviously, neither approach is
wholly satisfactory.
A second factor that might have led to overestimation of mutation rates is illegitimacy. It is certainly possible that any of the 38 probands whose paternity remains
untested inherited a copy of a n abnormal allele from a n unidentified father. In fact,
we have encountered one family in which analyses of blood groups demonstrated
that two children with Hb Koln were illegitimate (Stamatoyannopoulos et al., 1981).
An illegitimate child with Hb M Boston has also been found (Hollan et al., 1966;
Szelbnyi, pers. commun.). The mothers and nominal fathers of these three children
were unaffected. Conversely, the very severe, chronic hemolytic anemias associated
with the heterozygosity for many of the unstable hemoglobins, and the cyanosis
exhibited by heterozygotes for a n Hb M, lessen the likelihood of transmission. It is
in cases of unstable variants associated with relatively mild clinical signs that false
paternity must be considered with the greatest care. Further analyses of blood
groups, enzymes, and HLA types will be instrumental in identifying those probands
(if any) whose illegitimacy has led to improper designation of a case a product of new
Potential causes of underestimation of mutation rates
Our estimates of the rates at which ciM,OM, and unstable 0-chain diseases arise by
new mutation may well fall short of the actual mutation rates. There are several
reasons for harboring this suspicion.
First, inclusion of a case in our calculations requires that a long series of steps be
taken, starting with ascertainment by a primary physician and continuing through
referral to a hematologist, demonstration of a n abnormal hemoglobin (by electrophoresis, chromatography, spectrophotometry and/or tests of stability), and finally,
demonstration that the abnormality involves substitution of a single amino acid
residue in either the a or P chains. At some point, the hematological status of the
proband's parents must also be evaluated. In some instances of unstable hemoglobin
or Hb M disease, one or both parents were unavailable (e.g., the case of Hb M
described by Betke et al., 1966),and the cases were not considered further. Inasmuch
as some of these could have arisen by new mutation, their exclusion might have
contributed to a lowering of estimated mutation rates.
[Vol. 27, 1984
Second, Hook and Porter (1981) suggested that ascertainment of new mutations
arising during the later generation (1950-1974) might have been more effective than
that of new mutations arising in the earlier generation (1925-1949). Given that the
criteria for inclusion of a case could not be fully satisfied until techniques for
structural analysis of abnormal hemoglobins became comnlonplace in the 1960%
such suspicions are justified. To the extent that some affected individuals failed to
survive for from 15 to 40 years, the cases of new mutation identified among people
born from 1925 through 1949 will be fewer than the number of new mutants actually
produced during the same interval. If calculations of the rates a t which unstable 0chain, OM, and aM diseases arise are based only on cases arising in the later
generation, the results are 5.2 x lO-'lfi genelgeneration (n = 36 cases of unstable
0-chain disease, 2N = 686 x lo6), 2.1 x lO-'/P genelgeneration (n = 8 cases of OM
disease, 2N = 372.2 x lo6), and 1.1 x 10-8/a genelgeneration (n = 4 cases of aM
disease, 4N = 348 x lo6). These estimates are all higher than those presented in
Table 5, but none is markedly so. Because several years generally pass between
birth of a proband and determination of the structural abnormality in a globin
chain, it is quite likely that the unstable hemoglobins or Hbs M of some subjects
born in the later generation have not yet been fully characterized. Thus, the lack of
striking discrepancies between rates calculated over both generations and those
calculated over the later generation alone may simply reflect a failure to recognize
some cases that arose near either end of the 50-year interval.
Third, an additional set of factors, relating to conditions that encourage publication of descriptions of abnormal hemoglobins, has likely led to underestimates of
mutation rates. Of the 58 cases of new mutation listed in Table 3, 51 came to our
attention as subjects of published reports. Roughly 75% of these reports contained
either the results of analysis of a particular amino acid substitution not previously
observed in a human a-or P-globin chain, or the results of functional analyses of the
abnormal hemoglobins. With time, second, third, or fourth cases of a specific abnormal hemoglobin will be subject to analysis, yet will remain unpublished, there being
little or nothing to add to existing knowledge of the biochemical nature or physiological effects of the defect. Inasmuch as very few of the 58 cases were reported because
they arose as new mutations, documentation of some cases of new mutation may
well remain buried in laboratory notes.
Of the several potential sources of error described above, we judge the effects of
those leading to underestimation to outweigh the effects of errors that would, by
themselves, produce overestimates of mutation rates. In this light, the rates based
upon cases of Hb M and unstable hemoglobin diseases should be considered minimal.
Ours is not the first attempt to estimate mutation rates from data on cases of
unstable 0-chain disease or Hb M disease. Lehmann and Carrell (1969) found two
cases of Hb Hammersmith, a variant with unstable /3 chains, in a population of
approximately 2 x lo7 individuals. Like all cases of Hb Hammersmith observed to
date, these were products of new mutation that resulted in replacement of the
phenylalanyl residue in position 42 of the normal 0 chains by a leucyl residue.
Lehmann and Carrell thus estimated that the rate a t which Hb Hammersmith
arises by new mutation is 1 x 1Op7Iperson(or 5 x 10-'10 genelgeneration). Given
that Hb Hammersmith can arise only through the alterations of one AAA codon to
yield an AGA codon (Table 31, this estimate is very high. In fact, it is roughly
equivalent to that we have calculated for the rate of appearance of unstable 0-chain
disease, cases of which have stemmed from point mutations affecting 25 different
nucleotides. The high estimate of Lehmann and Carrell (1969) most likely reflects
the problems encountered when calculations are based upon cases found in relatively
small groups of subjects.
Motulsky (1968) estimated rates of nucleotide substitution in globin genes from
the frequency of rare, globin-chain variants among northern Europeans. First,
Motulsky derived the relationship pn = xql2, where x = the frequency of a variant
Nute and Stamatoyannopoulos]
in a population, q = the proportion of carriers with unaffected parents, and pn = the
rate at which a particular nucleotide is replaced in effecting the amino acid substitution in question. Assuming that the likelihood of substitution is uniform over all
nucleotides that encode a- and 0-chain sequences, the frequency in a population of a
particular abnormal allele was given by Motulsky as x = (2z/2)(l/n) (l/c), where z =
the combined frequencies of all electrophoretically detectable variants in the population (about 0.0005), n = the number of amino acid residues in a globin chain
(about 140), and c = the number of potential amino acid substitutions that could be
produced by substitution of single nucleotides in a codon (about seven). Assuming
that approximately 50%of all possible amino acid substitutions will be electrophoretically detectable (this is probably a n overestimate), z is multiplied by 2; division of
22 by 2 is predicated upon Motulsky’s assumption that the a and 0 chains are
encoded by genes a t single loci, and that both types of genes are at equal risk of
mutation. Under these assumptions, x = 5 x lop7.
Estimating that about 1%of all subjects with a n aM or OM allele are born to
unaffected parents, Motulsky calculated the rate a t which any one of the five forms
of Hb M arises as pn = xq/2 = (5 x lop7) (1/2) (0.01), or 2.5 x 10-g/nucleotide/
generation. If 10% of affected subjects have normal parents, pn becomes 2.5 x lo-’/
nucleotide/generation. These figures are estimated rates at which any given nucleotide in a n exon of a n a- or 0-globin gene is substituted by a particular one of three
other nucleotides, since it is assumed that the rate is the same over all sites. If these
results are amended to reflect the existence of duplicate a-chain loci and the fact
that any nucleotide can be replaced b any one of three others, x becomes (2z/3)
(l/n) (l/c) = 3.4 x
pn = 1.7 x 10- 3 or 1.7 x lo-’ (depending upon one’s choice
ofq = 0.1 or q = 0.011, and the overall rate of substitution per a- or @-genenucleotide
becomes 5.1 x lo-’ or 5.1 x lO-’/generation. The latter estimate agrees closely
with estimates presented in Table 5 .
In applying Motulsky’s (1968) approach to data on variant hemoglobins collected
in Japan by Kimura and Ohta (19731, Vogel and Rathenberg (1975) estimated pn as
5 x lo-’. Application of the corrections noted above yields a n estimated rate of
substitution per a- or 0-gene nucleotide of 10 x lO-’/generation. Again, the discrepancies between this and the rates given in Table 5 are not especially noteworthy.
Neel et al. (1980) directly estimated that mutations producing electrophoretically
discernible variants of a wide array of proteins arise a t a rate of 3.4 x lOP6/gene/
generation. Their estimate was based upon the discovery of a single new mutant
among 18,946 individuals (a total of 289,868 locus tests were performed), one or both
of whose parents were within 2,000m of the hypocenters during the atomic bombings
of Hiroshima and Nagasaki. A total of 208,196 additional locus tests were conducted
on a cohort of 16,516 children, one or both of whose parents were in either of these
cities during the bombings, but were more than 2,500m from the hypocenters (and,
hence, judged to have received essentially no radiation). No new mutations were
detected in this group. Calculated over both groups, the rate at which mutation
produces electromorphs is 2.0 x lO-‘/gene/generation.
Lewontin (1974) estimated that only 27% of all possible amino acid substitutions
in a polypeptide chain of average length (about 300 residues) and composition will
alter net charge and, hence, electrophoretic mobility. Moreover, Motulsky (1968)
noted that approximately 25% of all nucleotide substitutions in the protein-encoding
portion of a gene will produce no change in the amino acid encoded. Adjusting the
rate at which electromorphs arise for the occurrence of “silent” mutations, we obtain
nucleotide substitutions/
a n overall rate of (2.0 x 1OP6)/(0.27x 0.75) = 9.9 x
genetgeneration. While this rate is somewhat higher than those estimated over all
nucleotides in the a- or 0-globin genes (Table 5), it is not markedly so. Furthermore,
many of the polypeptides screened by Neel et al. (1980) are considerably longer than
globin chains. Hence, their corresponding genes are a t greater risk of sustaining a
nucleotide substitution in any given generation. If we combine the estimated mutation rates for the 423 a-gene and 438 @-genenucleotides (Table 5), we obtain a rate
of 7.7 x 1OP6/861nucleotides/generation. Extrapolation of this rate to a n “average”
[Vol. 27, 1984
gene of 900 nucleotides yields 8.0 x lop6 nucleotide substitutionslgenelgeneration,
a figure that compares quite favorably with that produced upon adjustment of the
estimate of Neel et al. (1980).
At this juncture, it is worth noting that attempts to directly estimate rates of
mutation from anlyses of variant proteins have generally s s e r e d from the small
numbers of subjects (or loci) tested. While 500,000 locus tests is, by most standards,
a n impressive number, it is small when one is attempting to measure rates of 1 x
lop5 or less. Neel et al. (1980)remark that their “data scarcely lend themselves to
statistical manipulation.” In the same context, it must be acknowledged that adjustments of results based upon such data fail to enhance the reliability of the data
themselves. When the results of several analyses are, however, in general agreement, the credibility of each is enhanced.
Most efforts to directly estimate mutation rates in human subjects have relied
upon ascertainment of well-defined, clinically significant, dominant phenotypes.
With few exceptions, neither the nature nor site of the underlying abnormality in
DNA sequence has been demonstrated. Direct estimates of mutation rates for autosoma1 dominant diseases have ranged from a high of 1.4 x lO-*Igamete down to 1.8
x 10-7/gamete. The former, the highest found for a human condition (neurofibromatosis), may reflect inclusion of cases of similar phenotypes produced by somaticcell mutation. The latter (for von Hippel-Landau syndrome) is the lowest for any
dominant condition defined by a specific phenotype (references in Vogel and Rathenberg, 1975). The difference between the two extremes, nearly three orders of
magnitude, justifies the uncertainty experienced by those who estimate mutation
rates for conditions that have not been traced to well-defined genetic alterations.
Commonly, estimates have exceeded those a t which Hb M and unstable hemoglobin diseases are judged to arise (top set in Table 5) by between two and three orders
of magnitude (Vogel and Motulsky, 1979). While the sources of these differences
remain ill-defined, there are several possibilities. Of the 934 abnormalities known
to be inherited as autosomal dominant traits (McKusick, 19831, the majority are
much lower in frequency than those for which mutation rates have been estimated.
Many have, in fact, been found in but one or a few isolated families. It appears that
the clinically significant traits for which mutation rates have been estimated are
among those most frequently encountered in human populations. Although such
traits are best suited to calculations of mutation rates since there are considerable
amounts of information about them at hand, many of the resultant estimates are
likely to far exceed the average for all autosomal dominant mutations. A reasonable
approximation of the average rate for autosomal dominant diseases is 1 x lop6/
gamete (Stevenson and Kerr, 19671, a figure that is 24-125 times higher than the
rates estimated for a M ,OM, and unstable 0-chain diseases (Table 5).
As noted previously, excessively high estimates can result from the inclusion of
genocopies among cases of a particular phenotypic abnormality. For example, Mdrch’s
(1941) direct estimate of the rate at which achondroplasia arises by new mutation
was elevated by inclusion of chondrodystrophic syndromes of varied genetic etiologies. In later analyses (cited in Vogel and Rathenberg, 1975; Vogel and Motulsky,
19791, cases of achondrogenesis and thanatophoric dwarfism were excluded, and
resultant estimates (6-13 x 10p6/gamete) were considerably lower than that presented by Mdrch (about 4 x 10-5/gamete). There remains a nagging suspicion that
even the more recent estimates are based upon cases of dwarfism stemming from
abnormalities a t two or more loci. Similar sources of upward bias may prove partially responsible for the high mutation rates estimated for a number of autosomal
dominant disorders (e.g., aniridia, retinoblastoma, and myotonic dystrophy).
Brief consideration of Hb M disease should serve to underscore the nature of the
problems encountered in directly estimating mutation rates for clinically defined
syndromes. Hemoglobin M disease produced by mutation of a n a-globin gene differs
little in expression from that produced by 0-gene mutation. If, in ignorance of the
genetic basis of globin-chain structure, one attributed the origin of Hb M disease to
mutation at a single locus, the rate estimated would be that per three genes (two a
Nute and Stamatoyannopoulos]
genes and one 0 gene per normal haploid set of human chromosomes). Obviously,
attempts to measure rates of mutation per gene are subject to error should, as we
think likely, many of the phenotypes ascertained be produced by mutations at any
of several loci. (For this reason, rates of mutations producing abnormal, clinically
defined conditions are generally stated as per gamete, not per gene.) Finally, some
conditions are known to arise through any of a small, select set of nucleotide
substitutions out of the many that can occur in one or more genes. Of the 3[2(423) +
4381 = 3,852 nucleotide replacements that could affect the combined sequences of
the two a genes and one 0 gene in each human gamete, only seven (two in each a
and one in the (3 gene) will produce Hb M disease. Thus, the rate a t which aM or OM
disease arises by new mutation should be far closer to the rate of substitution per Qor 0-gene nucleotide than it is to the overall rate of nucleotide substitution per a or
(3 gene. Inspection of Table 5 shows this to be the case.
The ability to unambiguously attribute a phenotypic abnormality to mutation of a
specific gene fosters confidence in a n estimate of the rate per gene at which the
abnormality arises. Only when the cases in question are attributable to specific
nucleotide substitutions can the rate of replacement per nucleotide (or over all
nucleotides in the coding portion of the gene) be reliably estimated. While the
estimates based upon cases of newly mutated globin-chain genes are subject to
biases of unknown (but, we suspect, of reIatively slight) net effect, they are based
upon cases of fully characterized mutations of genes whose number and structures
are known in detail. These are the only cases of new mutation in man that have
been fully documented a t the molecular level.
Collection and analysis of data on cases of Hb M and unstable hemoglobin diseases
were supported in part by contract N01-ES-90002 from the National Institute of
Environmental Health Sciences and grant GM 15253 from the National Institutes
of Health, U.S. Public Health Service.
Adams, JG 111, Boxer, LA, Baehner, RL, Forget,
BG, Tsistrakis, GA, and Steinberg, MH (1979)
Hemoglobin Indianapolis (8112[G14] Arginine)an unstable &chain variant producing the phenotype of severe 8-thalassemia. J. Clin. Invest.
Aksoy, M, Erdem, S, Efremov, GD, Wilson, JB,
Huisman, THJ, Schroeder, WA, Shelton, JR,Shelton, JB, Ulitin, ON, and Muftiioglu, A (1972)
Hemoglobin Istanbul: Substitution of glutamine
for histidine in a proximal histidine (F8(92)0).J.
Clin. Invest. 51::2380-2387.
Baralle, FE, Shoulders, CC, and Proudfoot, N J
(1980)The primary structure of the human t-globin gene. Cell 21:621-626.
Betke, K, Kleihauer, EF, Gehring-Muller, R,
Braunitzer, G, Jacobi, J, and Schmidt, I (1966)
HhM Hamburg, eine 8-Ketten-Anomalie:
a2j’2635r(= HbM Saskatoon). Klin. Wochenschr.
Bunn, HF, Bradley, TB, Davis, WE, Drysdale, JW,
Burke, JF,Beck, WS, and Laver, MB (1972)Structural and functional studies on hemoglobin Bethesda (a2p2145H’s),a variant associated with
compensatory erythorocytosis. J. Clin. Invest.
Bunn, HF, Haney, DN, Gahbay, KH, and Gallop,
PM (1975)Further identification of the nature and
linkage of the carbohydrate in hemoglobin AI,.
Biochem. Biophys. Res. Commun. 67:103-109.
Clegg, Jl3, and Gagnon, J (1981) Structure of the {
chain of human embryonic hemoglobin. Proc. Natl.
Acad. Sci. USA 78t6076-6080.
Cohen-Solal, MM, Authier, B, deRiel, JK, Murnane, MJ, and Forget, BG (1982) Cloning and nucleotide sequence analysis of human embryonic {globin cDNA. DNA 1:255-363.
Crow, JF (1981) The neutralist-selectionist controversy: An overview. In EB Hook and IH Porter
(eds):Population and Biological Aspects of Human
Mutation. New York: Academic Press, pp. 3-14.
Deisseroth, A, Nienhuis, A, Lawrence, J, Giles, R,
Turner, P, and Ruddle, FH (1978) Chromosomal
localization of human fl globin gene on human
chromosome 11in somatic cell hybrids. Proc. Natl.
Acad. Sci. USA 751456-1460.
Deisseroth, A, Neinhuis, A, Turner, P, Velez, R,
Anderson, WF, Ruddle, F, Lawrence, J, Creagan,
R, and Kucherlapati, R (1977) Localization of the
human a-globin structural gene to chromosome
16 in somatic cell hybrids by molecular hybridization assay. Cell 12:205-218.
Denniston, C (1982) Low level radiation and genetic risk estimation in man. Annu. Rev. Genet.
Efremov, GD, Stojmirovic, E, Lam, HL, Wilson, JB,
and Huisman THJ (1978) Hh Beth Israel (8102
[G4] Asn-Ser) observed in a Yugoslavian teenager. Hemoglobin 275-77.
Forget, BG (1977) Nucleotide sequence of human 0
globin messenger RNA. Hemoglobin 1:878-881.
Fritsch, EF, Lawn, RM, and Maniatis, T (1980)
Molecular cloning and characterization of the human 8-like globin gene cluster. Cell 19:959-972.
GrovB, SS, Jenkins, T, Kamuzora, HL, and Lehmann, H (1977)Congenital Heinz body haemolytic
anaemia due to haemoglobin Perth in a Nama
child seemingly aggravated by the high nitrate
content of the water supply. Acta Haematol.
Hayashi, A, and Yamamura, Y (1964) Hemoglobin
Mosaka, a new variant of hemoglobin M. Jpn. J.
Human Genet. 9t87-94.
Hollan, SR, Szelenyi, JG, Lehmann, H, and Beale,
D (1966) A Boston-type haemoglobin M in Hungary: Haemoglobin M Kiskunhalas. Haematologia 1:ll-18.
Hook, EB, and Porter, IH (1981) Addendum. In EB
Hook and IH Porter (eds): Population and Biological Aspects of Human Mutation. New York: Academic Press, pp. 348-350.
Ingram, VM (1959) Haemoglobin S. Abnormal human haemoglobins 111. The chemical difference
between normal and sickle-cell haemoglobins.
Biochim. Biophys. Acta 36r402-411.
Ingram, VM (1963) Hemoglobins in Genetics and
Evolution. New York: Columbia University Press.
International Hemoglobin Information Center
(1979a)List of variants. Hemoglobin 3:104-115.
International Hemoglobin Information Center
(1979b) Variants of the alpha chain. Hemoglobin
International Hemoglobin Information Center
(1979~)Variants of the beta chain. Hemoglobin
luchi, I, Hidaka, K, Harano, T, Shibata, S, Ueda, S,
Shimasaki, S, Ohfuchi, S , and Kurokawa, I(1979)
Identification of abnormal hemoglobin Koln (p98
(FG-5) Val-Met): The fourth Hb Koln variant
found in Japan. Kawasaki Med. J. 5t61-69.
Johnson, CS, Moyes, D, Schroeder, WA, Shelton,
JB, Shelton, JR,and Beutler E (1980) Hemoglobin Pasadena, cu&75(E1Y)’’eu-(1Arg- identification
by high performance liquid chromatography of a
new unstable variant with increased oxygen affnity. Biochim. Biophys. Acta 623:360-367.
Jones, RT, Brimhall, B, Huisman, THJ, Kleihauer,
E, and Betke, K (1966) Hemoglobin Freiburg: Abnormal hemoglobin due to deletion of a single
amino acid residue. Science I54:1024-1027.
Kimura, M, and Ohta, T (1973) Mutation and evolution at the molecular level. Genetics [Suppl.]
Lauer, J, Chen, C-KJ, and Maniatis, T (1980) The
chromosomal arrangement of human d i k e globin
genes: Sequence homology and a-globin gene deletions. Cell 20:119-130.
Lawn, RM, Efstratiadis, A, O’Connell, C, and Maniatis, T (1980)The nucleotide sequence of the human P-globin gene. Cell 21t647-651.
Lehmann, H, and Carrell, RW (1969)Variations in
the structure of human haemoglobin with particular reference to the unstable haemoglobins. Br.
Med. Bull 2514-23.
Lewontin, RC (1974) The Genetic Basis of Evolutionary Change. New York: Columbia University
Liebhaber, SA, Goossens, M, and Kan, YW (1980)
Cloning and complete nucleotide sequence of human 5’-wglobin gene. Proc. Natl. Acad. Sci. USA
Liebhaber, SA, Goossens, M, and Kan, YW (1981)
Homology and concerted evolution at the a1 and
a2 loci of human a-globin. Nature 290:26-29.
Little, PFR (1982) Globin pseudogenes. Cell 28:683684.
Maggio, A, Massa A, Giampaolo, A, Mavilio, F, and
Tentori, L (1981)Occurrence of Hb M Iwate (a287
His-Tyr &) in an Italian carrier. Hemoglobin
(Vol. 27, 1984
McKusick, VA (1983)Mendelian Inheritance in Man
(6th ed.). Baltimore: Johns Hopkins University
Mdrch, ET (1941) Chondrodystrophic Dwarfs in
Denmark. Opera ex Domo Biologiae Hereditariae
Humanae Universitatis Hafniensis, Vol. 3. Copenhagen: E. Munksgaard.
Motulsky, AG (1968) Some evolutionary implications of biochemical variants in man. Proc. VIIIth
Int. Congr. Anthropol. Ethnol. Sci. 1:364-365.
Murari, J, Smith, LL, Wilson, JB,Schneider, RG,
and Huisman, THJ (1977) Some properties of
hemoglobin Gun Hill. Hemoglobin 1:267-282.
Nagel, RL, Lydield, J, Johnson, J, Landau, L,
Bookchin, RM, and Harris, MB (1976) Hemoglobin Beth Israel: A mutant causing clinically apparent cyanosis. N. Engl. J. Med. 295t125-130.
Ned, JV,and Rothman, E (1981) Is there a difference among human populations in the rate with
which mutation produces electrophoretic variants? Proc. Natl. Acad. Sci. USA 78t3108-3112.
Neel, JV,Satoh, C, Hamilton, HB, Otake, M, Goriki, K, Kageoka, T, Fujita, M, Neriishi, S, and
Asakawa, J (1980) Search for mutations affecting
protein structure in children of atomic bonb survivors: Preliminary report. Proc. Natl. Acad. Sci.
USA 77:4221-4225.
Nute, PE, and Stamatoyannopoulos, G (1981)Estimates of mutation rates per nucleotide in man,
based on observations of de nouo hemoglobin mutants. In EB hook and IH Porter (eds): Population
and Biological Aspects of Human Mutation. New
York Academic Press, pp. 337-347.
Nute, PE, Stamatoyannopoulos, G, Hermodson,
MA, and Roth, D (1974) Hemoglobinopathic erythrocytosis due to a new electrophoretically silent
variant, hemoglobin San Diego (fl109(Gll)Val
+Met). J. Clin. Invest. 53t320-328.
Rousseaux, J, Nuyts, Jp,Demouveau, G, and Dautrevaux, M (1980) A severe hemolytic anemia related to a new case of hemoglobin Perth (Abraham
Lincoln) in a French patient. Hemoglobin 4:8993.
Schroeder, WA, Shelton, JB,Shelton, JR,Powars,
D, Friedman, S, Baker, J, Finklestein, JZ, Miller,
B, Johnson, CS, Sharpsteen, JR, Sieger, L, and
Kawaoka, E (1982)Identification of eleven human
hemoglobin variants by high-performance liquid
chromatography: Additional data on functional
properties and clinical expression. Biochem. Genet. 20t133-152.
Slightom, JL, Blechl, AE, and Smithies, 0 (1980)
Human fetal Gy- and *-,-globin genes: Complete
nucleotide sequences suggest that DNA can be
exchanged between these duplicated genes. Cell
Spritz, RA, DeRiel, JK,Forget, BG, and Weissman,
SM (1980) Complete nucleotide sequence of the
human &globin gene. Cell 21:636-646.
Stamatoyannopoulos, G, and Nute, PE 11974) Genetic control of haemoglobins. Clin. Haematol.
Stamatoyannopoulos, G, and Nute, PE (1982) De
novo mutations producing unstable Hbs or Hbs
M. 11. Direct estimates of minimum nucleotide
mutation rates in man. Hum. Genet. 60:181-188.
Stamatoyannopoulos, G, Nute, PE, Adamson, JW,
Bellingham, AJ, and Funk, D (1973) Hemoglobin
Olympia (020 valine-methionine): An electrophoretically silent variant associated with high
oxygen affinity and erythrocytosis. J. Clin. Invest.
Stamatoyannopoulos, G , Nute. PE. Giblett, E. Detter, J, and Chard, R (1976) Haemoglobin M Hyde
Nute and Stamatoyannopoulos]
Park occurring as a fresh mutation: Diagnostic,
structural, and genetic considerations. J. Med. Genet. 13t142-147.
Stamatoyannopoulos, G, Nute, PE, and Miller, M
(1981) De novo mutations producing unstable
hemoglobins or hemoglobins M. I. Establishment
of a depository and use of data to test for a n
association of de novo mutation with advanced
parental age. Hum. Genet. 58r396-404.
Stevenson, AC, and Kerr, CB (1967) On the distribution of frequencies of mutation to genes determining harmful traits in man. Mutat. Res. 4t339352.
Trimble, BK, and Doughty, JH (1974) The amount
of hereditary disease in human populations. Ann.
Human Genet. 38:199-223.
Trimble, BK, and Smith, ME (1977) The incidence
of genetic disease and the impact on man of a n
altered mutation rate. Can J. Genet. Cytol.
Vogel, F, and Motulsky, AG (1979) Human Genetics. New York: Springer-Verlag.
Vogel, F, and Rathenburg, R (1975) Spontaneous
mutation i n man. Adv. Hum. Genet. 5223-318.
Wood, WG, Stamatoyannopoulos, G, Lim, G, and
Nute, PE (1975) F-cells in the adult: Normal values and levels in individuals with hereditary and
acquired elevations of Hb F. Blood 4633'71-682.
Без категории
Размер файла
1 463 Кб
using, mutation, abnormal, rate, estimating, hemoglobin, human
Пожаловаться на содержимое документа