Printer Friendly

Evolutionary dynamics of fitness recovery from the debilitating effects of Muller's ratchet.

Due to their high mutation rates, RNA viruses usually consist of highly heterogeneous populations termed quasispecies (Eigen and Schuster 1979; Eigen and Biebricher 1988). In a defined environment, the mutant spectrum of a quasispecies is generally centered around one or several master sequences that are more fit but nonetheless often represent only a minor proportion of the mutant distribution (Eigen and Schuster 1979; Domingo et al. 1985; Domingo and Holland 1988; Eigen and Biebricher 1988). Thus, during some infectious process, when a single genome of the quasispecies is randomly chosen to replicate and to generate a new quasispecies, there is a high probability that the new quasispecies may carry a slightly deleterious mutation relative to the most fit members of the parental quasispecies. Upon further replication, the debilitating mutation may revert or be compensated for by mutations at other sites. However, when the effective population size is small, compensatory mutations might not arise. Succesive sampling or genetic bottleneck events followed by low-fidelity genome replication may lead to substantial fitness losses. Muller (1964) predicted this situation by stating that when mutation rates are high and population sizes are small, a kind of irreversible ratchet mechanism will lead to the gradual decrease in fitness of populations, particularly in asexual organisms. Experimental evidence for the operation of Muller's ratchet in viral populations has been reported by Chao (1990) for the bacteriophage [Phi]6, and by our group (Duarte et al. 1992, 1993, 1994; Clarke et al. 1993) for the animal pathogen vesicular stomatitis virus (VSV), a negative-sense, unsegmented RNA virus.

Variable but significant fitness losses occurred during repeated plaque-to-plaque transfers of VSV starting from clones with higher than, equal to, or slightly lower than wild-type VSV fitness (Duarte et al. 1992; Clarke et al. 1993). Fitness losses of VSV could be countered by increasing the size of the serially transfer populations (Clarke et al. 1993). However, using different VSV clones, it was shown that the effect of a bottleneck on fitness is greater when the fitness of the parental population is higher (Novella et al. 1995a; Elena et al. 1996) and that two consecutive population expansions intercalated between bottlenecks were not enough to avoid the Muller's ratchet effect (Duarte et al. 1993).

In contrast to these previously published experiments in which viruses with high to moderately low fitness were used, the present work observes the evolution of different clones derived from an extremely debilitated population under different demographic conditions. We found that the fitness of debilitated VSV clones increased in all cases, regardless of the demographic regime, to values that approximate that of the wild-type VSV. The implications of this finding are of both theoretical and applied importance. The theoretical implications shed light on the dynamics of adaptation and evolution of viral populations. Of applied interest are possible implications on the safety of attenuated vaccines. Several different procedures have been proposed for the severe debilitation of viral clones to be used as vaccines (Hurlbut 1956; Peleg 1971; Singh 1971; Taylor and Marshall 1975; Novella et al. 1995c). This research demonstrates that recuperation of such debilitated viruses is evident in vitro, even in spite of severe recurrent genetic bottlenecks.

MATERIALS AND METHODS

MARM F Clone and Subclones

The low fitness clone F of VSV, an [I.sub.1]-monoclonal, antibody-resistant virus (MARM), was obtained after 20 plaque-to-plaque transfers by Clarke et al. (1993) from the high fitness clone MARM X. This original F clone was diluted and plated on a monolayer of [BHK.sub.21] cells, and four well-isolated plaques were collected (labeled as I, II, III, and IV). The fitnesses of each F subclone were 0.00009 [+ or -] 0.00001 ([+ or -] SE), 0.00036 [+ or -] 0.00001, 0.00012 [+ or -] 0.00001, and 0.00004 [+ or -] 0.00001, respectively.
TABLE 1. Summary of the demographic conditions characterizing each
experimental regime. [N.sub.min] and [N.sub.max] are, approximately,
the minimum and maximum population sizes in each cycle; k is the
number of population expansions between two consecutive bottlenecks,
and g is the approximate number of genome replications between
bottlenecks (based on a simple doubling process and taking into
consideration the dilutions between serial population transfers).
For more details, see text.

Regime          [N.sub.min]     [N.sub.max]       k         g

A                 1             [10.sup.5]         0        17
B                 1             [10.sup.10]        1        33
C                 1             [10.sup.10]        2        47
D                 1             [10.sup.10]        5        86
[E.sub.20]      [10.sup.6]      [10.sup.10]       20       286
[E.sub.40]      [10.sup.6]      [10.sup.10]       40       552


Other Biological Materials and Experimental Procedures

The biological materials and experimental protocols used to estimate fitness have been previously described (Holland et al. 1991; Duarte et al. 1992, 1993; Clarke et al. 1993). In short, [BHK.sub.21] cells were grown as monolayers under Dulbecco modified Eagle's minimum essential medium (DMEM) containing 5% newborn bovine calf serum. Cell monolayers were infected with a multiplicity of infection of 0.1 virus per cell (to avoid the possible appearance of defective interfering particles). Virus density was quantified by plaque assays using confluent [BHK.sub.21] cell monolayers under DMEM solidified with 0.7% agarose. The mouse monoclonal antibody, employed to distinguish wild-type from MARM viruses, was the [I.sub.1] ([I.sub.1] MAb) produced and characterized by VandePol et al. (1986).

Demographic Regimes

The six demographic regimes were designed to study the effect of an increasing number of large population transfers and replication rounds between two consecutive bottleneck events. Table 1 summarizes the important demographic features of each regime (minimun and maximum population sizes, number of population expansions, and number of genome replications between consecutive bottlenecks).

Regime A. - Virus from a well-defined single plaque was isolated, diluted, and transferred in a daily plaque-to-plaque passage regime a total of 20 times. In parallel, an aliquot of the virus from the initial previous plaque was amplified in an infection involving a large number of cells ([approximately]2 x [10.sup.6] cells). From the virus obtained in this large-scale infection the following regimes were derived.

Regime B. - Virus from this first amplification was diluted 104 times and plated on a new monolayer. After 24 h of incubation, we isolated a new single plaque and used it to initiate a new large population passage. This process of bottlenecking and subsequent population expansion was repeated 20 times.

Regime C. - From the first population expansion, and through a [10.sup.4]-fold dilution, we initiated a second large-scale infection. When this second population passage was completed, we took a new sample, plated it, and isolated a new single plaque. With this plaque, we initiated a new cycle of two consecutive population expansions. This process of bottlenecking followed by two population expansions was repeated 20 times.

Regime D. - From the second population expansion of regime C, we took a sample that was used to initiate a third expansion. These expansions were interrupted after the fifth by a new plating and isolation of a new single colony. As in the previous three regimes, this regime alternated bottlenecking and five consecutive population passages was also repeated 20 times.

Regime E. - From the fifth large-scale infection of regime D, we continued the same protocol for up to 40 ([E.sub.40]) large population passages without intervening plaque isolation.

Note that the five demographic regimes are not independent because each one is generated with viruses from the previous ones. To ensure that all the single clones picked were really random samples and to discard any bias in the selection process, we marked a dot on the back of the plates without looking and then selected the closest plaque to this dot, independent of its size or appearance. Infections in all dynamics were carried out at 37 [degrees] C. In all cases, [I.sub.1] MAb was added at passages 10 and 19 (and also in 39 in regime [E.sub.40]) to eliminate any possible wild-type revertants.

Relative Fitness Assays

At the end of each passage series, the derived viral populations were subjected to relative fitness assays (Holland et al. 1991). Each MARM population was mixed with a known amount of the wild-type clone. A differential quantitation of a genetically marked MARM clone, compared with total virus, was done by parallel plating of the virus with and without [I.sub.1] MAb in the agarose overlay. In most cases, triplicate platings were carried out for each virus plaque number determination. These virus mixtures were then used to initiate serial competition passages. After each competition passage, the resulting virus mixture yield was diluted [10.sup.4]-fold and used to initiate the next competition transfer by infection of a fresh cell monolayer. The same dilution was used to determine (as described above) the relative proportion of both competitors. The number of competition passages varied between two and a maximum of five, depending on the speed with which one competitor displaced the other. These determinations gave the proportion MARM ([p.sub.t]) to wild type (1 - [p.sub.t]) at passage number t. The antilogarithm of the slope of the regression [Mathematical Expression Omitted] is taken as an estimate of the mean fitness of the MARM population relative to the wild-type population (Hartl and Clark 1989). (In previously published experiments [Duarte et al. 1992, 1993, 1994; Clarke et al. 1993], we used an approximation of this expression [Dykhuizen and Hartl 1980] that is only valid for small values of selection coefficients: [Mathematical Expression Omitted]. All statistical analyses were carried out using the SPSS package (Norusis 1992).

[TABULAR DATA FOR TABLE 2 OMITTED]

RESULTS

Fitness Recovery under Each Type of Demographic Regime

Table 2 shows the estimates of fitness of MARM populations at the end of each experimental regime. Fitnesses of the derived clones from all demographic regimes and replicates were always several orders of magnitude higher than the fitness of their parental F subclones. In fact, the relative fitness value of the five regimes ranges between 0.11 (in replicate IV under regime A) to 1.2 (in replicates II and III under regime E). The derived clones are 2600- to 31,000-fold more fit than the original F, but the fitness of the original MARM X (3.05 [+ or -] 0.03) was never reached. Similar results were found by Clarke et al. (1993) when they subjected the F clone to a regime similar to the E regime for 22 cycles; the maximum fitness value reached was 1.8 [+ or -] 0.3 (with the present fitness definition). This value was 44,000-fold higher than the original F clone.

As seen in Table 2, there is significant heterogeneity in the mean fitness within the different regimes. A Levene's test showed violation of homogeneity of variances ([F.sub.5,18] = 4.001, P = 0.013), so nonparametric tests were necessary for a proper statistical analysis of fitness heterogeneity (Sokal and Rohlf 1981). To test the null hypothesis of an equal fitness across regimes, we used the Kruskal-Wallis test, which showed a significant heterogeneity among all six different regimes (H: 12.577, 5 df, P = 0.028). Then we asked if there was a correlation between the magnitude of the fitness improvement and the number of large population transfers. A Pearson's correlation coefficient was employed and showed a significant correlation between the resulting fitness and the number of large population passages between bottlenecks (r = 0.925, 4 df, P = 0.008). To identify groups with equal effects on fitness, we used Tukey's multiple range tests. Three groups were obtained: the first group included regimes A, B, and C, all of which had mean fitness estimates smaller than the fitness of wild type; the second group formed by regimes D and [E.sub.20] had larger fitness gains, reaching the fitness of the wild type (1.0); finally, the third group was formed by regime [E.sub.40] alone, where the fitness values were higher than the wild type. Thus, despite the heterogeneity among replicates, the type of demographic regime strongly influenced the extent of fitness change; the greater the number of large population passages between bottlenecks, the greater the gains in fitness.

Parallel Evolution among Replicates

The four replicate populations showed similar fitness results under all regimes tested. This kind of parallel evolution is surprising, especially if one considers that their evolution depended on the independent accurence and random sampling of mutations in each population. A critical factor promoting parallel evolution could be the fact that populations have evolved in identical environments. A wider fitness range is seen in those demographic regimes with lower effective population sizes (Table 2). An index of the relative extent of divergence versus parallelism in evolution is given by the ratio [Mathematical Expression Omitted] (Vasi et al. 1994), where [[Sigma].sub.G] is the genetic standard deviation for fitness and [Mathematical Expression Omitted] is the average change from the ancestors. The index gives us a measure of the average difference among the independently derived genotypes relative to the average evolutionary change from the ancestral state. The indices corresponding to the first equal fitness group defined above (0.534 for A, 0.984 for B, and 1.712 for c) were larger than the indices corresponding to the other two groups (0.312 for D and 0.150 for [E.sub.20]; 0.377 for [E.sub.40]) such negative correlation between the index and the final fitness suggest that the genetic differences among independently derived genotypes are greater relative to the average change in fitness from the ancestral state when bottlenecks are tighter. This is a reflection of the fact that genetic drift increases genetic differences due to the random fixation of mutations.

Fitness Trajectories

To study the fitness recovery trajectory, we analyzed fitness changes from regime E, which showed the biggest fitness gains. Fitness values were determined at different times for each replicate. Figure 1 shows the trajectory of fitness during the 40 population passages of evolution. From Figure 1 it is evident that some maximum value was reached. For this reason [TABULAR DATA FOR TABLE 3 OMITTED] we decided to fit a model showing this characteristic. The model choosen was a logarithmic-hyperbolic model similar to that previously used by Lenski and Travisano (1994) to account for the fitness changes in E. coli populations: [Mathematical Expression Omitted], where t is the passage number, [Mathematical Expression Omitted] a is the final asymptotic log-fitness value, and b is the passage number at which the log-fitness is equal to half the maximum asymptotic value. This model has two main features: mean fitness evolves rapidly during the first passages and asymptotically approaches a maximum value. The estimated trajectories for mean fitness as well as the estimated parameters for each replicate are shown in Fig. 1 and Table 3, respectively. The existence of a maximum fitness value attainable by F clone was also observed by Clarke et al. (1993). They did not find significant differences (t = 1.000, 12 df, P = 0.169) between fitness estimates of viruses yielded after either [Mathematical Expression Omitted] or 61 [Mathematical Expression Omitted] large population passages (as shown by fitness vectors in fig. 3 of Clarke et al. 1993).

By contrast, in the experiments carried out by Novella et al. (1995b) using VSV clones of higher fitnesses than clone F, they did not observe any maximum fitness after their evolutionary process and for this reason they fitted a double-exponential model, based on a more complex mutation-selection-drift balance (Gabriel et al. 1993; Lynch et al. 1993): [Mathematical Expression Omitted], where [Alpha], [Beta], [Gamma], and [Delta] are combinations of population parameters such as selection coefficients, mutation rates, and carrying capacities. Although both the logarithmic-hyperbolic and the double-exponential models have as common features an initial period of rapid evolution followed [TABULAR DATA FOR TABLE 4 OMITTED] by a slower period, they differ in that the double-exponential model does not predict any maximum value for fitness gain.

To determine whether the more complex double-exponential model fits the experimental observations better than the logarithmic-hyperbolic, a partial F-test (Kleinbaum and Kupper 1978) was performed. In Table 4 we present the results of this test when applied to all the available experimental data (F clone and the D and N clones studied by Novella et al. 1995b as shown in their figs. 2 and 3, respectively). For F clone and for one of the replicates of N clone, no significant difference between the models was obtained. However, in the remainder of cases the use of the double-exponential model provides a significant increase in the goodness of fit despite the use of one extra parameter.

DISCUSSION

Previous experiments with VSV illustrated the profound fitness losses associated with plaque-to-plaque transfers of viral clones (Duarte et al. 1992) and rapid fitness regains as result of large population passages (Clarke et al. 1993; Novella et al. 1995b). In the present report, we have explored fitness variation in VSV subclones derived from an extremely unfit clone obtained after 20 successive plaque transfers of a high fitness VSV clone termed MARM X (Clarke et al. 1993). We showed that these highly debilitated clones gained fitness in all demographic regimes tested. Contrasting with this result, Duarte et al. (1993) observed that the intercalation of one large population passage between two consecutive bottlenecks failed to overcome the ratchet effect when the experiment was done with neutral or high fitness viruses. Why does this discrepancy exist? A possible explanation could be that, although genomic mutation rates are the same for both high and low fitness clones, the proportion of beneficial mutations is greater in a clone with lower fitness by the simple fact that there is more room for improvement. In a fully deterministic world where Muller's ratchet plays no role, this would mean that the rate with which beneficial mutations are fixed is greater for clones with lower fitness. Hence a clone of low fitness will experience a more rapid fitness increase than a clone of high fitness.

The most striking result is that of regime A. In spite of frequent and severe genetic bottlenecks, precisely the same conditions as that in which the fitness was lost before, fitness increased dramatically. A partial explanation may be that competition is taking place between variants within a single plaque. Due to the quasispecies nature of this virus, mutants abound in any given population. In this scenario, a beneficial mutation that confers faster replication may occur during the growth of a single plaque; the probability of picking this fitter clone from the plaque is disproportionaly large due to its elevated frequency. Thus, in some sense, we have an involuntary bias during the picking proccess: we randomly picked virus from a sample of plaques biased toward fitter individuals due to their increased chance to form visible plaques after 24 h. This bias would be more important at the beginning of the experiment, when the majority of clones were close to the initial low fitness, than at the end, when all of them have increased their fitness. To better understand what happened under regime A, it should be desirable to know both the fitness trajectory followed by F clone when it was generated from MARM X and the fitness trajectory followed by the F subclones during this experiment. The first kind of information will allow us to know if it was a single drop in fitness due to a single mutation of major effect (perhaps easier to compensate) or it was a continuous decline in fitness due to the accumulation of many deleterious mutations of small effect. The second kind of information will give us information about how many steps were necessary to recover fitness: just one compensating a major deleterious mutation or many more compensating small-effect mutations.

The fitness trajectories under regime E show that the populations evolved until reaching similar adaptative peaks. The existence of a maximum for fitness during evolution can be explained by many different reasons. Perhaps the most parsimonious reason is based on the possibility that all the deleterious mutations accumulated during the isolation of clone F had been compensated for or reverted during our experiment. It is interesting to note that the maximum fitness value reached under regime E is equal to the fitness showed by the wild type (1.0), yet lower than that observed in MARM X clone (3.05) from which clone F was derived. It is easy to argue that some long-standing level of fitness may be quickly attained by successive beneficial mutations of small effect. An argument for nonasymptotic increases in fitness, however, must necessarily invoke some sort of macromutation, or beneficial mutation of large effect, that bridges an adaptive valley. Such mutations are rare when compared with the small-effect, beneficial mutations that climb adaptive slopes step by step. It seems likely that these more common, small-effect, beneficial mutations are the type responsible for rescuing F clone from its adaptive pit. However, these small-effect, beneficial mutations are not reverting F clone to the original fitness peak from which it dropped (that occupied by MARM X). They are compensating some unknown lost function in a different way, driving the resulting populations to a different and more accesible fitness peak.

The observations here presented have important implications for the further development of vaccines by attenuation: the use of highly debilitated viral clones as live vaccines is risky business in light of the evidence presented here that such clones are able to recover fitness under several, and widely variable, demographic conditions. The predictions of this work are in concordance with the observations of Novella et al. (1995c). They obtained several attenuated viruses by continuous replication of VSV in sandfly cells and performed in vivo vaccination experiments in mice. Although initially these attenuated viruses were effective as vaccines, protecting the mice against further infections with lethal doses of virus, the attenuated virus eventually recovered fitness and neurovirulence after several passages on mammal cells. Certainly, a virus that has been engineered to lack some important gene for its replication and/or transmission should be less likely to regain fitness than those that had became deleterious by random accumulation of deleterious mutations.

ACKNOWLEDGMENTS

We are thankful for the valuable comments and suggestions of F. Gonzalez-Candelas, R. Lenski, J. Mongold, P. Gerrish, P. Moore, D. Rozen, and V. Cooper. Work in Valencia was supported by DGICYT grant PB94-0034-C02-02; work in Madrid by grants from DGICYT (PB94-0034-C02-01), FIS (n93/0003-01), and Fundacion Ramon Areces; work in San Diego by grant AI4627 from National Institutes of Health. SFE had a predoctoral fellowship from Conselleria d'Educacio i Ciencia of Generalitat Valenciana.

LITERATURE CITED

CHAO, L. 1990. Fitness of RNA virus decreased by Muller's ratchet. Nature 348:454-455.

CLARKE, D. K., E. A. DUARTE, S. E ELENA, A. MOYA, E. DOMINGO, AND J. J. HOLLAND. 1993. Genetic bottlenecks and population passages cause profound fitness differences in RNA viruses. J. Virol. 67:222-228.

DOMINGO, E., AND J. J. HOLLAND. 1988. High error rates, population equilibrium, and evolution of RNA replication systems. Pp. 3-36 in E. Domingo, J. J. Holland, and P. Ahlquist, eds. RNA genetics. Vol. 3. CRC Press, Boca Raton, FL.

DOMINGO, E., E. MARTINEZ-SALAS, F. SOBRINO, J. C. DE LA TORRE, A. PORTELA, J. ORTIN, C. LOPEZ-GALINDEZ, P. PEREZ-BRENA, N. VILLANUEVA, R. NAJERA, S. VANDEPOL, D. STEINHAUER, N. DEPOLO, AND J. J. HOLLAND. 1985. The quasispecies (extremely heterogeneous) nature of viral RNA genome populations: biological relevance - a review. Gene 40:1-8.

DUARTE, E. A., D. K. CLARKE, A. MOYA, E. DOMINGO, AND J. J. HOLLAND. 1992. Rapid fitness losses in mammalian RNA virus clones due to Muller's ratchet, Proc. Nat. Acad. Sci. USA 89: 6015-6019.

DUARTE, E. A., D. K. CLARKE, S. F. ELENA, A. MOYA, E. DOMINGO, AND J. J. HOLLAND. 1993. Many-trillionfold amplification of single RNA virus particles fails to overcome the Muller's ratchet effect. J. Virol. 67:3620-3623.

DUARTE, E. A., I. S. NOVELLA, S. LEDESMA, D. K. CLARKE, A. MOYA, S. E ELENA, E. DOMINGO, AND J. J. HOLLAND. 1994. Subclonal components of consensus fitness in an RNA virus clone. J. Virol. 68:1325-1330.

DYKHUIZEN, D., AND D. L. HARTL. 1980. Selective neutrality of 6PGD allozymes in E. coli and the effects of genetic background. Genetics 96:801-817.

EIGEN, M., AND C. BIEBRICHER. 1988. Sequence space and quasispecies distribution. Pp. 211-245 in E. Domingo, J. J. Holland, and P. Ahlquist, eds. RNA genetics. Vol. 3. CRC Press, Boca Raton, FL.

EIGEN, M., AND P. SCHUSTER. 1979. The hypercycle: a principle of natural self-organization. Springer-Verlag, Berlin.

ELENA, S. F., F. GONZALEZ-CANDELAS, I. S. NOVELLA, E. A. DUARTE, D. K. CLARKE, E. DOMINGO, J. J. HOLLAND, AND A. MOYA. 1996. Evolution of fitness in experimental populations of vesicular stomatitis virus. Genetics 142:673-679.

GABRIEL, W., M. LYNCH, AND R. BURGER. 1993. Muller's ratchet and mutational meltdown. Evolution 47:1744-1757.

HARTL, D. L., AND A. G. CLARK. 1989. Principles of population genetics. Sinauer, Sunderland, MA.

HOLLAND, J. J., J. C. DE LA TORRE, D. K. CLARKE, AND E. A. DUARTE. 1991. Quantitation of relative fitness and great adaptability of clonal populations of RNA viruses. J. Virol. 65:2960-2967.

HURLBUT, H. S. 1956. West Nile infection in arthropods. Am. J. Trop. Med. Hyg. 5:76-85.

KLEINBAUM, D.G., AND L. L. KUPPER. 1978. Applied regresion analysis and other multivariate methods. Duxbury Press, North Scituate, MA.

LENSKI, R. E., AND M. TRAVISANO. 1994. Dynamics of adaptation and diversification: a 10000-generation experiment with bacterial populations. Proc. Nat. Acad. Sci. USA 91:6808-6814.

LYNCH, M., R. BURGER, D. BUTCHER, AND W. GABRIEL. 1993. The mutational meltdown in asexual populations. Evolution 44: 1725-1737.

MULLER, H. J. 1964. The relation of recombination to mutational advance. Mut. Res. 1:2-9.

NORUSIS, M. J. 1992. SPSS[R] for Windows[TM]: advanced statistics. Rel. 5. SPSS, Inc., Chicago.

NOVELLA, I. S., S. F. ELENA, A. MOYA, E. DOMINGO, AND J. J. HOLLAND. 1995a. Size of genetic bottleneck leading to virus fitness loss is determined by mean initial population fitness. J. Virol. 69:2869-2872.

NOVELLA, I. S., E. A. DUARTE, S. F. ELENA, A. MOYA, E. DOMINGO, AND J. J. HOLLAND. 1995b. Exponential increases of RNA virus fitness during repeated transmission. Proc. Nat. Acad. Sci. USA 92:5841-5844.

NOVELLA, I. S., D. K. CLARKE, J. QUER, E. A. DUARTE, C. H. LEE, S. C. WEAVER, S. F. ELENA, A. MOYA, E. DOMINGO, AND J. J. HOLLAND. 1995C. Extreme fitness differences in mammalian and insect hosts after continuous replication of vesicular stomatitis virus in sandfly cells. J. Virol. 69:6805-6809.

PELEG, J. 1971. Growth of viruses in arthropod cell cultures: applications. Attenuation of Semliki Forest (SF) virus in continuously cultured Aedes aegypti mosquito cells (Peleg) as a step in production of vaccines. Curr. Top. Microbiol. Inmunol. 55: 155-161.

SINGH, K. R. P. 1971. Propagation of arboviruses in Singh's Aedes cells. Curr. Top. Microbiol. Immunol. 55:127-133.

SOKAL, R. R., AND F. J. ROHLF. 1981. Biometry. 2d ed. W. H. Freeman and Company, New York.

TAYLOR, W. P., AND I. D. MARSHALL. 1975. Adaptation studies with Ross River virus: laboratory mice and cell cultures. J. Gen. Virol. 28:59-72.

VANDEPOL, S. B., L. LEFRANCOIS, AND J. J. HOLLAND. 1986. Sequence of the major antibody binding epitopes of the Indiana serotype of vesicular stomatitis virus. Virology 148:312-325.

VASI, F., M. TRAVISANO, AND R. E. LENSKI. 1994. Long-term experimental evolution in Escherichia coli. II. Changes in life-history traits during adaptation to a seasonal environment. Am. Nat. 144:432-456.
COPYRIGHT 1998 Society for the Study of Evolution
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1998 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Elena, Santiago F.; Davila, Mercedes; Novella, Isabel S.; Holland, John J.; Domingo, Esteban; Moya,
Publication:Evolution
Date:Apr 1, 1998
Words:4565
Previous Article:The coadaptation of parental and offspring characters.
Next Article:Tradeoffs between horizontal and vertical modes of transmission in bacterial plasmids.
Topics:

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters