Copyright 1971. All rights reserved

CURRENT STATUS OF THE GENE-FOB-GENE CONCEPT

H. H. FLOR Crops Research Division, Agricultural Research Service, USDA and Dept. of Plant

Pathology, North Dakota Agricultural E�periment Station, Fargo

INTRODUCTION

One of the most successful means of controlling plant diseases has been the development of varieties with major or vertical resistance genes. This type of resistance is easily manipulated in a breeding program and is efIec­tive until strains of the pathogen to which it does not confer resistance be­come established. Then, if another gene that conditions resistance to the new strains of the pathogen is available, this resistance gene may be incorporated into the variety by the plant breeder. In doing this, the breeder either con­sciously or unconsciously is applying the principle of the gene-far-gene hypothesis. Plants resistant to races that are virulent on old varieties possess the new resistance gene. With the diseases of some crops, this process has becn repeated at relatively frequent intervals (4D, 42, 82). However, in some instances a single gene has conferred adequate resistance for many years 80,82).

In plant diseases caused by living organisms, the same phenomena: in­fection type in rusts, percent of infected plants in smuts of cereals, fleck or lesion in apple scab, are criteria of both the reaction of the host and the pathogenicity of the parasite. They indicate the relative resistance or sus­ceptibility of the host and the relative avirulence or virulence of the para­site.

The gene-for-gene hypothesis was proposed (20,25) as the simplest ex­planation of the results of studies on the inheritance of pathogenicity in the .flax rust fungus, M elampsora lini. On varieties of flax, Linum usitatissimum that have one gene for resistance to the avirulent parent race, F 2 cultures of the fungus segregate into monofactorial ratios. On varieties having 2, 3, or 4 genes for resistance to the avirulent parent race, the F2 cultures segregate into bi-, trio, or tetra factorial ratios (20-22) respectively. This suggests that for each gene that conditions reaction in the host there is a correspond­ing gene in the parasite that conditions pathogenicity. Each gene in either member of a host-parasite system may be identified only by its counterpart in the other member of the system.

275

3531

Ann

u. R

ev. P

hyto

path

ol. 1

971.

9:27

5-29

6. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tanf

ord

Uni

vers

ity -

Mai

n C

ampu

s -

Lan

e M

edic

al L

ibra

ry o

n 05

/18/

13. F

or p

erso

nal u

se o

nly.

276 FLOR

Because of the lack of uniformity in the criteria by which plant patholo­gists identify gene-for-gene relationships, Person, Samborski & Rohringer (66) discuss its numerous facets and present the following definition:

A gene-for-gene relationship exists when the presence of a gene in one popula­tion is contingent on the continued presence of a gene in another population, and

where the interaction between the two genes leads to a single phenotypic expres­

sion by which the presence or absence of the relevant gene in either organism

may be recognized.

In this review I shall attempt to show how studies of gene-for-gene rela­tionships may bc utilized to establish the means of variation in pathogenic fungi, to identify major genes conditioning vertical resistance, to develop resistant varieties, to elucidate the physiology of resistance and susceptibil­ity, and to explain the co-evolution of host-parasite systems.

VARIATION IN THE PATHOGEN

Differential varieties monogenic for rust reaction.-Physiologic races of plant pathogens are identified by their interaction with a set of host vari­eties termed "differentials". Often the differentials have been selected by the trial and error method with no knowledge of the number or identity of the resistance genes in each. The original flax rust differentials were developed by this method (19). Information obtained from such a series of differen­tials is of limited value. For example, in 1940, the flax variety Koto was developed as a rust-resistant replacement for the variety Bison. While being increased, Koto was attacked by rust. Later it was discovered that North American races virulent on Koto had been collected prior to 1940. Although two of the differentials then in use possessed the Koto resistance gene, they also possessed additional resistance genes that obscured the virulence of the races on Koto.

A set of differentials in which each member possesses a single, unique resistance gene is most efficient and informative in studies on the classifica­tion and variation of plant pathogens. Numerous instances of variation in plant pathogens have been attributed to mutation, heterokaryosis, or para­sexualism. Most lack a satisfactory genetic explanation. Knowledge of the pathogenicity genotype of the parental culture and of the variant, presum­ably derived from it, should give the investigator a more convincing basis for explaining how the variant developed. A prediction of the gene-for-gene hypothesis is that the pathogenicity genotype of a fungus culture can be established by a selfing study on differentials having single resistance genes. A routine race identification test establishes the pathogenicity genotype of organisms having a haploid pathogenic phase. vVith organisms having a di­karyotic or diploid pathogenic phase, the routine race test identifies the re­cessive pathogenicity genes. The fungus must be selfed to determine the heterozygosity or homozygosity of the dominant genes.

By backcrossing to Bison, a variety that has been susceptible to all North American races of M.lini, and utilizing the selective pathogenicity of

Ann

u. R

ev. P

hyto

path

ol. 1

971.

9:27

5-29

6. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tanf

ord

Uni

vers

ity -

Mai

n C

ampu

s -

Lan

e M

edic

al L

ibra

ry o

n 05

/18/

13. F

or p

erso

nal u

se o

nly.

CURRENT STATUS OF THE GENE-FOR-GENE CONCEPT 277

various races, flax lines apparently monogenic for each gene conditioning rust reaction have been developed. Races are classified by a dichotomous key on 18 of these lines (24). Infection types 0, 1, and 2 are classified as resis­tant and 3 and 4 as susceptible. Additional monogenic lines may be used as supplemental differentials to identify sub -ra ces.

Resistance genes occur as multiple alleles in five loci. The symbols Land M, us ed by Myers (60) to designate resistance genes in two loci, have been

retained. Symbols K, N, and P designate genes lying in the other three loci. Of the 26 resistan(;e genes that have been identified, 1lies in K, 12 in L, 6 in M, 3 in N, and 4 in the P locus. Resistance genes in each locus are identified

by numerical superscripts (25). To show the specificity of the interaction between host and parasite, the symbol for the resistance gene is used as a subscript to the s ymbol s A and a, respectively, ind ica ti ng avirulence and virulence in the p ar asite (Table 1).

TABLE 1. PATHOGENICITY GENOTYPES OF RACES 22 AND 1 AND F, CULTURE A OF

RACE 22 X RACE 1 ON 16 FLAX RUST DIFFERENTIALS MONOGENIC

FOR RUST RESISTANCE

Pathogenicity genotype of

Differential Resistance Fi culture A Reaction"

to Fl variety genotype

Race 22 Race 1 Culture A Race 22 Race 1

nucleus nucleus

Clay KK aKaK AKAK aK AK R Ottawa 770B LL aLaL ArAL aL AL R

Stewart L2L2 AL,Au AL.AL, AL2 AL2 R

Kenya UL4 aL,aL' AV.AL' aL' AL• R

Wi ld en L6L6 auau AuAu a L6 AJ,6 R

B ir io L6L6 auau AuAu au ALe R

Dakota MM aMaM AloIaM ali AM R

Williston Brown MiMi aloI!a�[1 a�!1a�!1 aMl aM! S

Cass MaMa aM3a�[! AmAM' aM' AM' R

Victory A M'M' aM.aM' aM.AM' aM' aM' S

Bombay NN ANAN ANAN AN AN R

Polk NlNl aN!aNl ANlANl aNl AN' R

Koto PP apap Apap ap Ap R

Akmolinsk plp' aplapl ap,APl apl apl S

Abyssinian p2p2 ap2apZ Ap,Ap, ap, ApI R Leona pap. aplIapa Ap,APB a p, Ap, R

" R=resistant; S=susceptible.

Although Bison has been susceptible to all North American races of M. lini and has served as the universally susceptible differential in my studies, it is resistant to certain races that occur in Australia (81) and India (69). Kerr (45) found that the gene that conditions resistance in Bison to Aus-

Ann

u. R

ev. P

hyto

path

ol. 1

971.

9:27

5-29

6. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tanf

ord

Uni

vers

ity -

Mai

n C

ampu

s -

Lan

e M

edic

al L

ibra

ry o

n 05

/18/

13. F

or p

erso

nal u

se o

nly.

278 FLOR

tralian race A lies in the L locus, and it has been designated L9. Because the monogenic differentials were derived from Bison backcrosses, those having resistance genes in the K, M, N, and P loci may possess gene L9. Therefore, these lines are not satisfactory for determining the pathogenicity genotype of the races nonpathogenic on Bison.

In the Linum-M elampsora system, resistance is invariably dominant and virulence is invariably recessive. A study on the inheritance of pathogenic­ity in the cross of race 22 X race 24 of M. lini was interpreted as indicating that virulence on the differential variety Williston Brown was dominant (21). A later study (30) indicated that this interpretation was erroneous.

Hybridization.�In the principal seed flax producing region of North America, flax rust overwinters only in the telial stage. Consequently, hybri­dization of the fungus is concomitant with the initiation of each year's in­fection. Controlled experiments have demonstrated the role of hybridiza­tion in pathogenic variability (20, 21, 25, 30). In a cross of race 6 X race 22 of M. lini (25),54 pathogenically different races were identified in a population of 67 F2 cultures. The parent races differed in pathogenicity on 12 of the monogenic differentials. Had enough F2 cultures been studied, 212

or 4096 races could have been identified from this cross. Hybridization is an important means of pathogenic variation in the rust and other fungi having a sexual stage as part of their life cycle, e.g., smuts (36), powdery mildews (59), and apple scab (2, '5,6). However, hybridization results only in a re­combination of the genes already existing in the parental races.

Mutation.-New genes for virulence are the results of mutation. Numer­ous attempts to induce mutations to virulence in the rust fungi have been made, but because the results were negative few have been published. In most instances the investigator knew neither the number nor the identity of the resistance genes in the host varieties on which he attempted to screen his mutants, nor the pathogenicity genotypes of the urediospores that he sought to mutate. Since resistance and avirulence usually are dominant, the chance of detecting a mutation to virulence is slight when urediospores are screened on varieties having two or more resistance genes. In order to at­tack such a variety the urediospore would have to be virulent on all the respective resistance genes of the host. Furthermore, a mutation to viru­lence in a urediospore that is homozygous for the dominant avirulence genes can be detected only if the corresponding gene in each dikaryotic nu­cleus mutates.

Anderson & Hart (1) screened urediospores of race 48 of Puccinia graminis f. sp. tritici that had been subjected to several dosages of thermal neutron and X-radiation on four wheat stem rust differentials. No muta­tions in the rust were observed. Rowell, Loegering & Powers (71) irradi­ated urediospores of race 111, P. graminis tritici, and those of an F 1 culture of race 111 X race 36. Although the cultures were phenotypically similar,

Ann

u. R

ev. P

hyto

path

ol. 1

971.

9:27

5-29

6. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tanf

ord

Uni

vers

ity -

Mai

n C

ampu

s -

Lan

e M

edic

al L

ibra

ry o

n 05

/18/

13. F

or p

erso

nal u

se o

nly.

CURRENT STATUS OF THE GENE-FaR-GENE CONCEPT 279

race 111 was of unknown pathogenic genotype, whereas the hybrid was het­erozygous for pathogenicity on the variety Marquis. No mutants were re­covered from screening the X-irradiated spores of race 111 on Marquis. Numerous mutants to virulence were obtained when the irradiated spores of the hybrid were screened on Marquis.

Varieties monogenic for resistance and a urediospore culture heterozy­gous for all genes for pathogenicity should be ideal for mutation studies. To obtain the closest approximation to this ideal, race 22 of M. lini, the race of widest virulence, was crossed with race I, the race of narrowest viru­lence. Selling studies showed that race 22 was homozygous for virulence on 14 and for avirulence on 2 differentials, while race 1 was homozygous for virulence on 1, homozygous for avirulence on 11 and heterozygous on 4 (Table 1). For example, since both parent races were avirulent on Bombay (NN), a urediospore of Fl culture A of race 22 X race 1 would remain avirulent on Bombay even if an AN gene in one nucleus mutated to the recessive aN' The AN gene in the other nucleus would condition avirulence of that spore on Bombay. But the mutation of the dominant AM gene in the race 1 nucleus would make the spore homozygous for the recessive genes aMaM and should enable it to attack Dakota (MM).

To test this hypothesis, Fl urediospores of race 22 X race 1 culture A of M. lini (Table 1) were subjected to X-irradiation at LD50 and LD90 and screened for mutations on the differentials monogenic for resistance (27). No uredia developed on Stewart or Bombay, the differentials to which the urediospores were homozygous for avirulence. Of the 154 uredia that devel­oped on differentials to which the F 1 culture was heterozygous for patho­genicity, 94 produced sufficient inoculum for pathogenicity tests. Of these, 92 differed from the F 1 culture by virulence to the differential on which they were screened. This would be expected of one "hit" on the gene string. The other two cultures were virulent on an additional differential as would be expected of two "hits." Mutations to virulence on Birio (L6) and Cass (MS) were about 20 times as frequent as those on Koto (P) and Leona (P3). Spontaneous mutations were obtained only on Birio and Casso This sug­gests a parallelism between rates of X-ray-induced and spontaneous muta­tions. In regions where they are effective, varieties having resistance genes to which the parasite shows a low rate of mutation to virulence may main­tain their resistance for a longer period of time.

Schwinghamer (75) concluded that in race 1 of M. lini, induced muta­tions to virulence on the flax variety Dakota were due to a loss (deletion) of a chromosomal segment bearing the AM gene. I (28) selfed two of Schwinghamer's (74) race 1 X-ray-induced mutations to virulence on Koto. Each of the 198 selfed cultures was virulent on Koto and 16 were virulent on Abyssinian (F�) and Leona (PS), varieties to which race I was homozy­gous for avirulence (Table I). Therefore, new genomes for virulence were induced in a rust culture by X-irradiation. The mutations were attributed to the deletion of that portion of the chromosome carrying the closely linked

Ann

u. R

ev. P

hyto

path

ol. 1

971.

9:27

5-29

6. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tanf

ord

Uni

vers

ity -

Mai

n C

ampu

s -

Lan

e M

edic

al L

ibra

ry o

n 05

/18/

13. F

or p

erso

nal u

se o

nly.

280 FLOR

genes for avirulence to Koto, Abyssinian, and Leona. The fact that the dele­tion of a dominant gene for avirulence changed the pathogenicity of a spore from avirulent to virulent is an important consideration in studies on the physiology of host-pathogen interaction.

Heterokaryosis.-Parmeter, Snyder, & Reichle (63) have reviewed the role of heterokaryosis in the variability of plant pathogenic fungi. They conclude that, except in the rust and smut fungi where heterokaryosis takes a special form of the dikaryon, the evidence is inadequate to support the hypothesis that heterokaryosis affects pathogenicity or virulence.

Fusions between germ tubes and hyphae in various rust fungi have been observed frequently. Pathogenic and color variants obtained by screening urediospore mixtures of two races of P. graminis tritici have been attrib­uted to nuclear exchange through hyphal fusions or, if more than two new

variants were isolated, to somatic hybridization or parasexualism (7, 16, 82-84). The pathogenicity genotypes of the parental cultures and of the variants derived from them were not established.

I screened four Fl urediospore cultures of race.22 X race 1 of M. lini, designated A, B, C, and D, singly and variously paired on three differentials that were highly resistant to all four cultures (29). Race 22 was homozy­gous for pathogenicity. Therefore, one nucleus in each F 1 culture had the pathogenicity genotype of race 22 (Table 1). Since race 22 was virulent on the four differentials to which race 1 segregated for pathogenicity, a routine race test established the pathogenicity genotype of the race 1 nucleus in each F 1 culture. Race 22 was recovered only when urediospores of F 1 cul­ture B were mixed with those of A, C, or D. This would be expected if the race 22 nucleus of culture B was of one mating type (+) and those of cultures A, C, and D of the other (-) and that the (+) and ( -) race 22 nuclei became associated in one cell by nuclear migration through hyphal fusion. Some cultures were recovered that differed from one or another of the F 1 cultures by increased virulence on the differential on which they were screened. This would be expected of a spontaneous mutation at a sin­gle locus for which the culture was heterozygous.

Parase:ntalism or somatic hybridization.-The experiments in which the paired F 1 cultures of race 22 X race 1 were screened for pathogenic vari­

ants (29) served as a test for somatic hybridization as well as for heteroka­ryosis. Because of the numerous loci at which the Fl cultures were hetero­zygOttS for pathogenicity, variants virulent at several loci would have been expected if a parasexual process or somatic hybridization had occurred.

Ellingboe (16) has noted that some rust races are known to be heterozy­gous for several virulence genes. Since each race possesses a (+) and a ( -) haploid nucleus it should be as' capable of giving rise to variants through the parasexual process as a mixture of urediospores of two or more

Ann

u. R

ev. P

hyto

path

ol. 1

971.

9:27

5-29

6. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tanf

ord

Uni

vers

ity -

Mai

n C

ampu

s -

Lan

e M

edic

al L

ibra

ry o

n 05

/18/

13. F

or p

erso

nal u

se o

nly.

CURRENT 'STATUS OF THE GENE-FaR-GENE CONCEPT 281

races. Yet variants ascribed to parasexualism have been derived only from race mixtures (7,9, 16,82-84).

GENE-FoR-GENE SYSTEMS

Thc gene-for-gene hypothesis was based on correlated genetic studies of both host and parasite (20, 25). For each gene that conditions resistance in the host there is a corresponding gene that conditions pathogenicity in the parasite. Acceptance of the hypothesis enables one to construct hypothetical genotypes of host and pathogen in the absence of direct genetic studies by determining the reaction of a range of host varieties to a range of pathogen races (64). The following examples illustrate some of the data upon which the gene-far-gene relationships in the host-parasite systems listed in Table 2 are based.

Triticum-Pttccinia graminis.-Like the Linum-M. lini system, the Triti­cum-Po graminis gene-far-gene relationship is based on genetic studies of

TABLE 2. HOSl"-PAII.MIl"E SYSTEMS 1'011. WHICH GENE-FOII.·GENE

RELAl"lONSHIPS HAVE BEEN SUGGESl"El>

System and Referenee(s)

Linum-Melampsora lini (20, 25) Triticum-Puccinia graminis tritici (34, 43, 53, 86) Triticum-Puccinia striiformis (87) Triticum-Puccinia recondita (32, 73) Avenae-Puccinia graminis avenae (54) Coffea-Hemileia vastatrix (61) Helianthus-Puccinia helianthi (72) Avenae-Ustilago avenae (38) Triticum-Ustilago tritici (62) Triticum-Tilletia caries (56) Triticum-Tilletia contraversa (39) Hordeum-Erysiphe graminis hordei (58) Triticum-Erysiphe graminis tritici (68) Malus- Venturia inaequalis (2, 6) Avenae-Helminthosporium victoriae (85) Solanum-Phytophthora 'infestans (77) Lycopersicum-Cladosporium fulvum (11) Solanum-Synchytrium endobioticum (41)

both the host and the dikaryotic parasite. Loegering & Powers (53) deter­mined the inheritance of pathogenicity of 108 F 2 cultures of race 1 1 1 X race 36 on 20 wheat varieties. Their results indicated eight independently inherited genes for pathogenicity. On the variety Marquis, to which race 111 was avirulent and race 36 virulent, the F 2 cultures segregated into a ratio ap­proximating 12 ayirl)lel1t: 3 intermediate: 1 virulent. Thi� inqicateq that

Ann

u. R

ev. P

hyto

path

ol. 1

971.

9:27

5-29

6. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tanf

ord

Uni

vers

ity -

Mai

n C

ampu

s -

Lan

e M

edic

al L

ibra

ry o

n 05

/18/

13. F

or p

erso

nal u

se o

nly.

282 FLOR

Marquis possesses two genes for resistance to the parent culture of race 111. Williams, Gough & Rondon (86) determined the pathogenicity of the parental, Fl and F2 cultures of Loegering & Powers on Little Qub, Mar­quis, and on lines homozygous for single genes for resistance derived from Little Club X Marquis. Little Club was susceptible to all cultures. The tests confirmed those previously published (53). Both genes for avirulence are dominant and moderate avirulence is masked by high avirulence.

Green (34) found that virulence on Marquis lines carrying genes srS, Sr9a, Sr9b, or Srll is due to single recessive genes in each case and virulence on tines Sr7 or Srl0 is probably so controlled.

Kao & Knott (43) studied the inheritance of pathogenicity in the cross race 29 X race IlIon Marquis and Prelude lines carrying single genes for resistance to stern rust. Virulence on Sr5, Sr6, Sr8, Sr9a, Sr14 and a gene in Marquis is recessive in a gene-far-gene relationship. However, virulence on SrI is possibly controlled by two complementary dominant genes.

Coffea-Hemileia vastatrix.-Since the perfect stage of the coffee rust fungus is not known, studies on the genetics of its pathogenicity are not

possible. By determining the pathogenicity of 12 races of the parasite to clonal lines and hybrids between clonal lines, Noronha-Wagner & Betten­court (61) were able to postulate the resistance gene or genes in the clonal lines and hybrids. They determined that resistance to H. vastatrix is condi­tioned by four dominant genes. Following Person's (64) method of analy­sis, they inferred the pathogenicity genotypes of the 12 races and predicted the likely occurrence of the four races not yet known.

Triticum-Ustilago tritici.-Oort (62) determined the reaction of eight variety groups of wheat to six physiologic races of U. tritici. He explained this study by assuming that two sets of resistance genes and two sets of incompatibility genes (dwarfing of susceptible plants but no smutted heads) in the host interacted with four sets of complementary genes for virulence in the parasite. Thus, he demonstrated a gene-far-gene relationship in this host-parasite system without studying the genetics either of resistance in the host or pathogenicity in the parasite.

Halisky (36) has reviewed the occurrence and significance of other gene-for-gene relationships in the smut fungi.

H ordeum-Erysiphe graminis and Triticum-Erysiphe graminis.-The ge­netics of host-pathogen interaction in the powdery mildews has been re­viewed by Moseman (59). He (58) and Powers & Sando (68) extended the gene-for-gene hypothesis to the powdery mildews. Since the pathogenic phase is haploid, all cultures are either ;{ (avirulent ) or V (virulent). Moseman (58) crossed a culture of E. graminis f. sp. hordei, virulent on the barley varieties Goldfoil and Kwan, with one avirulent on each variety. Re� ::iistance in each variety was conditioned by a single gene. The F2 cultures

Ann

u. R

ev. P

hyto

path

ol. 1

971.

9:27

5-29

6. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tanf

ord

Uni

vers

ity -

Mai

n C

ampu

s -

Lan

e M

edic

al L

ibra

ry o

n 05

/18/

13. F

or p

erso

nal u

se o

nly.

CURRENT STATUS OF THE GENE-FaR-GENE CONCEPT 283

segregated for pathogenicity into the 1:1:1:1 ratio predicted by the gene-for­gene hypothesis. Powers & Sando (68) in a somewhat similar study found that for each of the two genes conditioning resistance in the wheat variety N ormandie there were corresponding genes for virulence in E. graminis f. sp. tritici.

Malus-Venturia inaequalis.-Boone & Keitt (6) studied the inheritance of pathogenicity of 10 parent lines of V. inaequalis on 15 apple varieties. Their results indicate that 7 different gene pairs condition lesion (suscepti­ble) and fleck (resistant ) reactions. From the reactions of the avirulent alleles they postulated the number and identity of resistance genes carried by seven differential apple varieties. In subsequent studies (2, 5), 19 inde­pendently inherited pathogenicity genes have been identified. Boone (5) presents a comprehensive discussion of the genetics of V. inaequalis in this volume.

Flangas & Dickson (18) and Berry (3) studied the Z ea-Puccinia s01'ghi system. They reported that pathogenicity of P. sorghi was not inherited in a Mendelian manner and postulated "gene pools" or "indeterminate inheri­tance" to explain their results. These are the only reports of nonMendelian inheritance of pathogenicity to specific or major genes for resistance of which I am aware. Berry suggested that a larger number of inbred clones of one line be studied and that failure to obtain consistent infection on Ox­alis spp. with teliospores is a major difficulty. This suggests that lethal genes may be associated with repeated selfing of P. sorghi, and that the popula­tions studied were not random samples.

The rust differentials were lines of corn monogenic for resistance (18). However, Hooker (40) observed that nearly all corn grown in the United States possesses generalized or polygenic resistance. Perhaps more consis­tent results could be obtained by the use of differentials in which monogenic resistance had been added to a variety lacking both specific and generalized resistance. Such differentials would also possess the same cytoplasmic back­ground.

DEVELOPMENT OF RESISTANT VARIETIES

The control of plant diseases by breeding for resistance has been the subject of several recent reviews (8, 12, 13, 40 ). In these, the advantages and limitations of the use of major gene resistance, polygenic resistance, and multiline varieties have been adequately discussed.

Identification of new major resistance genes.-In the North Central States ·flax rust has been satisfactorily controlled by major resistance genes. In fact, polygenic resistance has not been demonstrated. This may be be­cause it has not been actively sought. Since 1940, six genes, then considered to condition resistance to all North American races of M. lini, have become ineffective. Varieties in which these genes condition the resistance have suc-

Ann

u. R

ev. P

hyto

path

ol. 1

971.

9:27

5-29

6. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tanf

ord

Uni

vers

ity -

Mai

n C

ampu

s -

Lan

e M

edic

al L

ibra

ry o

n 05

/18/

13. F

or p

erso

nal u

se o

nly.

284 FLOR

cum bed to new races or to races that had remained undetected until a vari­ety dependent on a particular gene for its resistance became widely grown. Although there are seven genes that condition resistance to all known North American races, additional sources of rust resistance are sought.

New material for testing is provided by the Plant Introduction Service, USDA. Genes for rust resistance can be distinguished by tests for allelism, differences in their resistance phenotype, and the selective pathogenicity of physiologic races of the parasite. Six of the seven genes that condition resis­tance to M. lini in North America arc ineffective against the widely virulent South American race 22. F2 cultures of race 22 crossed with North Ameri­can races have been effective in isolating and identifying unknown resis­tance genes (22,26).

Development of varieties multigenic for resistance.-In the flax-flax rust system, virulence has been recessive and resistance dominant. A new race can attack a heretofore resistant variety only if it is virulent on all of that variety's resistance genes. The probability that a race will become ho­mozygous for two or more new virulence genes is much less than that it will do so for one pair. Therefore, varieties with two or more genes for rust re­sistance should be less apt to succumb to new races than varieties possessing a single resistance gene.

Flax varieties having almost any combination of resistance genes, within the limits of allelism, may be developed by utilizing the selective pathogen­icity of races of M. lini. Since the flax plant grows by the unfolding of a terminal bud, the resistant genes in each plant of a hybrid progeny may be ascertained by successive inoculation at approximately weekly intervals with races that identify the respective genes. For example, all plants in the progeny of a cross of La X M3 that are resistant to race 218, (virulent on La and avirulent on MS) possess the gene MS. Likewise, all plants resistant to race 106 (virulent on MS and avirulent on L6) possess gene L6. Plants re­sistant to both races possess both genes. Lines carrying three genes for resis­tance to all North American races of M. lini were developed by applying the principle that each resistance gene in a hybrid progeny is identified by inoculation with a race avirulent to that gene but virulent on all other re­sistance genes involved in the cross (25, 31).

Studies of F2 populations indicate that the rust-resistance genes in flax occur as multiple alleles in five loci (25). A more efficient test for allelism is a study of testcross populations (55). I (30) studied testcross progenies involving resistance alleles in the L, M, N, and P loci. No dominant cross­over was obtained in tests of more than 30,000 plants of the L heterozy­gates. Likewise, no crossovers were obtained in tests of 7354 P pl X pp and 2306 pp2 X pp plants. Resistance genes lying in the Land P loci are either alleles or very closely linked. Crossing over in progenies of MMl X mm,

MM3 X mm, and MM4 X mm ranged from 0.17 to 0.39%. In a test of 9115 NNl .x n.n plapts, O,IJ% .�rossjn.g over w�s .obtaiped .. ;£"ines .pure for Ml'vf3,

Ann

u. R

ev. P

hyto

path

ol. 1

971.

9:27

5-29

6. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tanf

ord

Uni

vers

ity -

Mai

n C

ampu

s -

Lan

e M

edic

al L

ibra

ry o

n 05

/18/

13. F

or p

erso

nal u

se o

nly.

CURRENT STATUS OF THE GENE-FOR-GENE CONCEPT 285

MM4, and NNl were derived from the dominant gene crossovers. Genes for pathogenicity on lines of flax possessing genes M, M3, M4, N,

and Nl are independently inherited. The MM3, MM\ and NNl resistance genes are so closely linked that each pair would have been regarded as a single gene in screening tests of varieties to serve as flax rust differentials. Had differentials possessing genes MM3, MM4, or NNl been used in patho­genicity tests of F2 cultures (21,25), ratios approximating 15 avirulent: 1 virulent would have been obtained. This would have been interpreted as an instance in which two complementary genes in the parasite conditioned pa­thogenicity on a differential possessing a single gene for resistance-an ex­ception to the gene-for-gene hypothesis.

Induction of rust-resistant mutants by irradiation.-Reports of the suc­cessful induction of disease-resistant mutants was reviewed by Konzak ( 48). Later studies (10, 46) suggest that the earlier reports be reassessed and that tests be conducted under more rigid standards of isolation.

The variety Bison, which has no genes for resistance to North Ameri­can races of M. lini, and Bison backcrosses monogenic for resistance were used in a study on induced mutations (Flor, unpublished). Seed was pro­duced in the greenhouse at Fargo, North Dakota. The parental plants were tested for purity with appropriate races. Outcrossing has never been ob­served in plants grown in the greenhouse. These seeds were irradiated by Drs. Calvin Konzak and Seymour Shapiro at the Brookhaven National Lab­oratory.

In preliminary field tests of several lines monogenic for rust resistance, up to 10% of the plants in some irradiated progenies were resistant to races to which the parent was susceptible. Many of the plants from irradiated seed were obviously damaged. Frequently the anthers did not dehisce nor the filaments elongate. When this occurred the petals, which are usually shed within two to three hours of opening, adhered most of the day and attracted .flying insects. All the "mutants" tested proved to be resistance genes from adjacent plantings.

In another trial, seed of the monogenic differential Cass (MS), that had been subjected to dosages of 35,000 and 40,000 r was sown in a wheat nur­sery. All except two plants in the 734 progenies had the reaction of the par­ent. These two plants were resistant to race 107, virulent on Casso Genetic studies showed them to be heterozygous for resistance gene Nl. Gene Nl conditioned a resistance in a line of flax that was increased in a plot more than 200 m from the wheat nursery. Obviously, excessive care to avoid nat­ural crossing is warranted in attempts to induce mutations for resistance in plants.

Seed of a Bison plant was increased in the greenhouse during the winter of 1956-1957. This seed, subjected to X-ray dosages of 60,000, 80,000 and 100,000 r, was sown in field plots at Brookhaven National Laboratory, Long Jsland, N.Y. These ,Plots wen: remote frOm any known commercial Or eX�

Ann

u. R

ev. P

hyto

path

ol. 1

971.

9:27

5-29

6. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tanf

ord

Uni

vers

ity -

Mai

n C

ampu

s -

Lan

e M

edic

al L

ibra

ry o

n 05

/18/

13. F

or p

erso

nal u

se o

nly.

286 FLOR

perimental flax. Dr. Konzak harvested these plots and sent the seed to me for testing. I determined the reaction to race 1 of M. lini of more than 375,000 X2 plants in the greenhouses at Fargo from 1958 to 1969. Race 1, virulent on Bison, is the race of narrowest virulence and consequently the best race available to detect a mutation to resistance in Bison. Mutations for color of the petals, chlorotic sectors and loss of vigor were observed but none for rust reaction.

Dominant and beneficial mutations in flax induced by X-irradiation are rare. However, nature operates on an almost infinitely greater scale than man, and mutation offers the most plausible explanation for the diversity of rust-resistant germ plasm in flax.

EPIDEMIOLOGY

Browning & Frey (8) and Watson (82) reviewed numerous reports on the relation of virulence to survival of plant pathogenic fungi. Virulence usually is recessive and most reports agree with the hypothesis that physio­logic races having wide host ranges compete poorly with races having a narrower host range on nonresistant varieties (79). This may be an over­simplification as undoubtedly there are numerous other factors involved in aggressiveness or fitness of pathogenic strains to survive.

The flax-flax rust system is well suited for a study of the survival of simple (few genes for virulence) and complex (many genes for virulence) races under field conditions. The pathogen is autoecious and, in the North Central States, each year's infection is initiated through the sexual phase. Browning & Frey (8) have outlined such an experiment.

I (23) identified the virulence (homozygous recessive) genes in each race of M. lini isolated from field collections during 1931-1951. The pre­dominant races carried the least number of virulence genes that permitted survival of an obligate parasite. The percent of races with unnecessary vir­ulence genes declined during the period of the study. A comparison of the 1931-1951 surveys with those of 1964-1968 (Flor, unpublished) supports this conclusion. However, it shows the influence of the resistance genes in varieties that were grown during each period.

When a wheat variety is attacked by new races of stem rust the breeder often adds a new resistance gene to its resistance genotype (82). In order to attack the new variety, a rust strain must retain its virulence to the old variety with added virulence to the new gene. With .flax, the new resistance gene has replaced rather than been added to the rust-resistance genes in the old variety. Rust resistance has been conditioned by a single gene in each of the widely grown commercial varieties that have succumbed during the past 30 years. To attack a new flax variety, only the virulence gene that will overcome the new resistance gene is needed.

An unnecessary resistance gene is one of which there is no record that it has conditioned rust resistance in a commercially grown variety in the

Ann

u. R

ev. P

hyto

path

ol. 1

971.

9:27

5-29

6. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tanf

ord

Uni

vers

ity -

Mai

n C

ampu

s -

Lan

e M

edic

al L

ibra

ry o

n 05

/18/

13. F

or p

erso

nal u

se o

nly.

CURRENT STATUS OF THE GENE-FaR-GENE CONCEPT 287

North Central States. Bison, susceptible to all North American races of M. lil1i, was the predominant variety from 1931 to the early 1940s. Rust col1ec­tions made during this period may be regarded as random samples. In the 1940s Bison was replaced by Koto (P), Dakota (M), and Marine (L). Thereafter, races col1ected in the field necessarily were virulent on one of these varieties. Races attacking Koto, Dakota, and Marine, respectively, were first found in 1940, 1948, and 1962. In some years, no rust was ob­served in commercial fields following the introduction of resistant varieties, and a sizeable proportion of the cultures studied were from experimental plots.

The immediate and residual effect of a resistance gene on virulence of collections is related to the popularity of the variety. Koto was not as widely grown as Dakota. Koto was susceptible to 24% of the rust collec­tions made during 1942-1947 when the Bison acreage was declining. Dur­ing 1948-1951, when Dakota was the most widely grown variety, Koto was susceptible to 51 % and Dakota to 79% of the collections. By 1954 both Koto

and Dakota had been largely replaced by other resistant varieties. Koto was susceptible to 34% and Dakota to 60% of the collections made during the 1964-1968 period. Thus, if a virulence gene becomes established in a rust population, it apparently persists for many years, although it is no longer essential to the survival of the parasite.

Cultures identified as having unnecessary genes for virulence declined from 65% of the 1931-1947 collections to 17% of those made during 1964-1968. By way of contrast, 90% of the 1931-1947 cultures possessed none or one pair of recessive genes whereas 94% of the cultures studied from 1964 to 1968 had two or more such pairs. This agrees with Watson's (82) obser­vations that genes for virulence continually increase as they overcome the resistance genes in popular varieties.

None of the cultures studied during 1964-1968 was virulent on the mon­ogenic differentials possessing genes K, L2, or N. These differentials were susceptible to some of the races identified during 1931-1951 (23). It is not known if gcncs for virulence on these differentials have disappeared from the flax rust population. They may remain as undetected heterozygotes.

Although they are not conclusive, the studies with flax rust tend to sup­port the thesis that aggressiveness is associated with dominant avirulence genes. Flax rust is a destructive disease in Argentina (78) and Egypt (communication from O. I. Selim, Sakha Research Station, Kafr-El­Sheikh, U.A.R.). In both areas most races are virulent on more than 20 differentials. If mutations to virulence are the result of chromosomal dele­tions (28, 74) is it likely that 20 deletions short enough to exclude vital functions should occur? More than one-third of the X-ray-induced muta­tions to virulence failed to produce sufficient spores for a pathogenicity test (27). Further studies of the correlation between a narrow pathogenicity spectrum and aggressiveness or vigor should be made.

Ann

u. R

ev. P

hyto

path

ol. 1

971.

9:27

5-29

6. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tanf

ord

Uni

vers

ity -

Mai

n C

ampu

s -

Lan

e M

edic

al L

ibra

ry o

n 05

/18/

13. F

or p

erso

nal u

se o

nly.

288 FLOR

NATURE OF RESISTANCE

The nature of specific or major gene resistance in the rusts and powdery mildews is one of the most challenging yet elusive problems in plant pathol­ogy. What are the mechanisms whereby the genes in the host and the para­site interact to control the development of the disease? Chemical, morpho­logical, or physiological variations observed between phenotypes that differ only in one gene for resistance are more apt to be related to resistance than variations observed between phenotypes that, in addition to the gene for reaction, differ in many other genes. The flax rust differentials, essentially monogenic for resistance, have been backcrossed to Bison as many as 15 times. Races of the parasite that differ by a single gene for virulence may be identified by a routine test on the differential lines monogenic for resis­tance (24). A Iso, mutant lines of the pathogen that differ from the parental culture by a single virulencc gene were develop cd by irradiation (27, 74). The Bison backcrossed lines and cultures of M. lini that differ by single genes for virulence have been used in studies on the nature of resistance and susceptibility.

Doubly et al (15) compared the globulin antigens of four races of M. lini with those of the rust-susceptible variety Bison and selections of the seventh backcross with Bison of the monogenic differentials Cass, Koto, and Ottawa 770B. Their results indicated that avirulence and virulence were re­lated to resistance and susceptibility through the specific rust antigens. A host variety was susceptible to a race of the pathogen when it contained, as a minor antigenic component, a protein serologically related to one of the proteins characterizing that particular race. Doubly and I have confirmed these results in tests (unpublished) in which the urediospores of the four races were produced on Bison. Thus, it is unlikely that rust spores carried over host antigens.

Studies with nonobligate parasites give some support to the common an­tigen hypothesis (14). Races of Xanthomonas nwlvacearum, the cause of angular leaf spot of cotton, to which the host was susceptible, had more antigens in common with cotton leaves than did nonpathogenic races.

Knott (47) used the agar-gel diffusion plate method to study the pro­teins in stem rust spores and in lines of Marquis wheat carrying single genes for resistance. In contrast to the cases mentioned above, he found no association of an antigen with a particular gene for resistance and could detect no antigens common to both host and parasite. The serology approach needs further study before any definite statement can be made concerning its validity.

Hadwiger & Schwochau (35) propose an induction hypothesis to ex­plain the gene-far-gene mechanism in flax and flax rust. The dominant genes in the pathogen direct the synthesis of compounds that activate the corresponding dominant genes in the host. The activated host genes affect an alteration or disintegration of the cellular organization of the host that results in a hypersensitive type of response. Tests of the induction hypothe-

Ann

u. R

ev. P

hyto

path

ol. 1

971.

9:27

5-29

6. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tanf

ord

Uni

vers

ity -

Mai

n C

ampu

s -

Lan

e M

edic

al L

ibra

ry o

n 05

/18/

13. F

or p

erso

nal u

se o

nly.

CURRENT STATUS OF THE GENE-FaR-GENE CONCEPT 289

sis with the flax-flax rust system are being made by Hadwiger and associ­ates.

Support of the induction hypothesis (35) was provided by Littlefield (51), who showed that flax rust infection is significantly reduced as a result of prior inoculation with an avirulent race of the pathogen. The induced resistance caused a reduction in the number, size, and rate of development of uredia by the normally pathogenic race. That study as well as subsequent investigations on the histology of flax rust resistance (52) suggest that dif­fusible substances from the pathogen may be responsible for triggering the resistance reactions.

CO-EVOLUTION OF HOST-PARASITE SYSTEMS

Parasitism is an antagonistic rather than a mutualistic form of sym­biosis. Unopposed selection for aggressiveness in the parasite would lead to elimination of the host. Unopposed selection for resistance in the host would lead to elimination of the parasite.

Mode (57) applied modern genetic concepts to the problem of microevo­lution in parasitic systems. He proposed' a mathematical model that incorpo­rates specific gene-for-gene relationships to explain the co-evolution of hosts and their parasites. He assumed that during their evolution, flax and the small grains were open pollinated. If host and parasite evolved together, the genic systems controlling their interaction have been established by the selective pressure exerted by one on the other. Mode's model, in addition to specific gene-for-gene interactions in host and parasite, assumes allelism of resistance genes, and independently inherited pathogenicity genes. He dem­onstrated that, under certain conditions, parasitic systems having these characteristics would be in equilibrium, thus assuring the survival of both host and parasite.

Resistance in the host usually is dominant, and the resistance genes of­ten occur as multiple alleles. Virulence in the parasite is usually recessive, and the genes for pathogenicity are independently inherited. Therefore, most parasitic systems possess the essential features included in Mode"s mathematical model of co-evolution.

Theoretical consideration of the origin of gene-for-gene relationships led Person (64, 65) to conclude that relationships such as that postulated for Linum-M elampsora should be the rule rather than an exception in host­parasite systems. This kind of relationship is an automatic result of muta­tion and natural selection for resistance in the host and for virulence in the parasite. Analysis of the data available for the Solanum-Phytophthora sys­tem showed it to be ideal to illustrate the properties inherent in gene-for­gene systems and to illustrate a new method of analysis for these properties.

Person (64) constructed a theoretical model of host-parasite interac­tions with gene-for-gene relationships, at five pairs of loci. With a dichoto­mous classification such as that used by Black et al (4), and with 5 genes for resistance this model differentiates 25 or 32 races. However, if one of

Ann

u. R

ev. P

hyto

path

ol. 1

971.

9:27

5-29

6. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tanf

ord

Uni

vers

ity -

Mai

n C

ampu

s -

Lan

e M

edic

al L

ibra

ry o

n 05

/18/

13. F

or p

erso

nal u

se o

nly.

290 FLOR

the five differentials carries two resistance genes ( e.g. Rv R2, R3, R., and Rl R5, in the respective varieties) only 24 races can be identified. Another feature of Person's theoretical model, which contained host varieties with 5, 4, 3, 2, 1, and 0 resistance genes, is that the number of races virulent on these varieties increase geometrically, 1, 2, 4, 8, 16, 32. Only one race at­tacks the variety with five resistance genes, etc. The variety lacking a gene for resistance is susceptible to all 32 races.

Person (64) analyzed my data on races of M. lini ( 24 ) in accordance with his theoretical model and concluded that many of the flax rust differen­tials, that I had presumed to be monogenic for resistance, possess two or more resistance genes. Johnson ( 42) has shown that evolution in the plant rusts is man-guided. In M. lini this may be illustrated by the pathogenicity of races obtained from different flax growing regions on the flax rust differ­entials carrying the four alleles for resistance in the P locus.

The differential varieties, Koto, Akmolinsk, Abyssinian, and Leona, that, respectively, carry resistance genes P, Pi, p2, and pa were susceptible to all Argentine races of M. lini described by Vallega & Antonelli ( 78) . Had Per­son's study been confined to Argentine races, no P resistance genes would have been detected and the four differentials possessing genes in the p locus could have served as universally susceptible varieties.

All four P differentials are highly resistant to Australian race A (81 ) . Tests with this race could be explained by assuming that the same gene con­ditions resistance in each differential carrying a gene in the P locus.

Koto is highly resistant and Akmolinsk, Abyssinian, and Leona are sus­ceptible to most Australian races other than race A (44) . These races iden­tify only the resistance gene in Koto. The three other differentials having a

P resistance gene would have been regarded as universally susceptible vari­eties.

Koto and Leona have been highly resistant and Akmolinsk and Abyssi­nian susceptible to all Oregon races. In tests with Oregon races, the same resistance gene would have been assigned to Koto and Leona. Akmolinsk and Abyssinian, being susceptible, would have no recognizable resistance gene.

The four varieties having resistance genes in P locus have been differen� tiated by races collected in the North Central States. I have identified hun­dreds of collections of urediospores from this region that attack Koto and hundreds that attack Akmolinsk, but none that attack both differentials. This was interpreted as indicating that, in this region, virulence on Koto and Akmolinsk is conditioned by closely linked genes in the repUlsion phase. Genetic studies (28) have verified this interpretation. Races to which Ak­molinsk is susceptible produce resistant infection type 1 (24) on Abyssinian in contrast to the near immune infection type 0 produced by races to which Akmolinsk is resistant. Leona has shown practical immunity to all uredio­spore collections from the North Central States.

These studies indicate that single genes in the P locus condition resis-

Ann

u. R

ev. P

hyto

path

ol. 1

971.

9:27

5-29

6. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tanf

ord

Uni

vers

ity -

Mai

n C

ampu

s -

Lan

e M

edic

al L

ibra

ry o

n 05

/18/

13. F

or p

erso

nal u

se o

nly.

CURRENT STATUS OF THE GENE-FOR-GENE CONCEPT 291

tancc in Koto, Akmolinsk, and Leona. Abyssinian could, although not neces­

sarily, possess two genes, the Akmolinsk gene pi plus an additional closely linked gene.

Person (64 ) recognized the possibility of man-directed evolution and ex­cluded some races from his analysis. His sample of 179 races in a system of 16 resistance genes is inadequate. Furthermore, 131 of the 179 races that comprised his- population were hybrids of the narrowly virulent North American races 6 or 24 with the widely virulent South American race 22. Many of the deviations from Person's model can be accounted for by simple Mendelian segregation in these hybrid populations.

The investigator's 1ependence on the avirulence gene complement of a pathogen to resolve and identify the resistance gene complement of a host may be further illustrated from studies of resistance to stem rust in the wheat variety Marquis. Puttick (70) reported that resistance in Marquis to race 19 is conditioned by a single gene. Williams et al ( 86 ) found that Mar­quis possesses two genes for resistance to one culture of race 111 and three genes for resistance to another. Sheen ( 76) identified at least four resis­tance genes in Marquis substitution lines of the wheat variety Chinese Spring.

Fincham & Day (17) point out that there can be no conclusive proof of the correctness of Person's hypothesis except through conventional genetic analysis of host and parasite. The primary centers of origin of many culti­

vated plants is well established (50). It was in these gene centers that both host and parasite evolved and developed a biological balance. The primary gene center has been and probably will continue to be the plant breeder's principal source of both vertical and horizontal or polygenic resistance. Re­sistance may be obtained from the cultivated host plants or their wild pro­genitors. Many primary gene centers have been extensively explored for resistant host germ plasm, but the pathogenic range of the accompanying parasites is unexplored (50). The Solanum-Po infestans system is an excep­tion in that both host resistance and pathogenic diversity have been studied

in their centers of primary origin (33). This system conforms ideally to Person's (64) analysis.

Experimental data to examine Mode's (57) or Person's (64) hypotheses on the co-evolution of host-parasite systems are difficult to obtain. Culti­vated plants and their parasites evolved over a long period of time and left few records. Mode assumed that the hosts : wheat, oats, barley, flax, etc. were open pollinated during their evolution. Today, these crops are rather

closely self-pollinated. Much evidence supports the assumption that aggres­siveness and wide virulence are negatively correlated (23, 49, 79, 82) but some data contradict it (54, 82). Modern man, by his breeding program and transportation of plants and their pathogens throughout the world, has up­set Nature's balance and has impeded the study of the co-evolution of host and parasite.

Horizontal resistance or multi genic resistance may have had a role in

Ann

u. R

ev. P

hyto

path

ol. 1

971.

9:27

5-29

6. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tanf

ord

Uni

vers

ity -

Mai

n C

ampu

s -

Lan

e M

edic

al L

ibra

ry o

n 05

/18/

13. F

or p

erso

nal u

se o

nly.

292 FLOR

maintammg the biological balance between host and parasite throughout their evolution. Most cultivated plants and their wild progenitors in primary centers of origin possess some degree of horizontal resistance to endemic pathogens (50) . During their evolution, host plants that possessed a new gene for vertical resistance should have had a survival advantage. That this advantage was temporary and of limited value is shown by the diversity of disease-resistant germ plasm that persists in plants obtained from these areas of primary origin. Seedlings of plants possessing horizontal resistance often are relatively susceptible to parasites, especially the plant rusts. This permits the fungus to become established early in the growing season when the amount of inoculum is limited. The host plant becomes more resistant as it grows and as the inoculum increases. Mature plants possessing only hori­zontal resistance rarely have the immunity conferred by many genes condi­tioning vertical resistance. This permits the continued development of the parasite with minimum damage to the host. Thus, a state of equilibrium is established between host and parasite that assures the survival of both. It may explain the survival in the centers of primary origin of the many resistance genes that developed during the evolution of the host, as well as that of plants lacking genes for vertical resistance.

Disease epidemics that are familiar in today's agriculture rarely occur in nature (67) . These epidemics develop as a result of man's interference. In his varietal breeding program he has either ignored horizontal resistance or subjected a host species, such as Pinus strobus, which evolved in North America, to a parasite, Cronartium ribicola, which evolved in Central Asia.

CONCLUDING REMARKS

The gene-for-gene hypothesis is applicable to most host-parasite systems in which resistance is conditioned by major (vertical resistance ) genes, and virulence increases in a step-wise fashion. A recent interesting extension of the gene-for-gene hypothesis is the study of Hatchett & Gallun (37) . They found that the ability of specific races of Hessian ,fly to survive on certain wheat varieties is controlled by genetic systems in the host and insect that are complementary. There is a gene-for-gene relationship between each gene for resistance in a wheat variety and a gene for survival in a race of the insect. This is the first report of gene-for-gene relationships between plants and insects. Similar relationships may exist between other insects, nematodes, bacteria, and viruses and their plant hosts.

The gene-for-gene concept has been useful as a tool to identify the roles of hybridization, mutation, heterokaryosis, and somatic hybridization in pathogenic variation of parasitic fungi. Also, it has served to survey and identify the resistance germ plasm in the host, to study induced mutations to resistance, and to develop multiline and multigenic varieties. Most data indi­cate that races of wide virulence compete poorly with races of narrow viru­lence on varieties susceptible to both. However, races having a wide viru-

Ann

u. R

ev. P

hyto

path

ol. 1

971.

9:27

5-29

6. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tanf

ord

Uni

vers

ity -

Mai

n C

ampu

s -

Lan

e M

edic

al L

ibra

ry o

n 05

/18/

13. F

or p

erso

nal u

se o

nly.

CURRENT STATUS OF THE GENE-FOR-GENE CONCEPT 293

lence spectrum dominate in the race population of flax rust in Argentina (78 ) , oat stem rust in Canada (54) , and wheat stem rust in Australia (82). Aggressiveness may not invariably be associated with restricted virulence, and the possibility that a super-race may develop on multiline and multi­genic varieties warrants further study.

Hypotheses based on the gene-for-gene concept have been proposed (57, 64) to explain the co-evolution of host plants and their parasites . In these hypotheses no consideration is given to the possible role of horizontal re­

sistance in co-evolution. Actually, there have been few recent studies in which resistance has been attributed to minor or polygenes that condition horizontal resistance (67). Some instances of resistance formerly attributed

to polygenes have subsequently been found to be conditioned by major genes. Also, the incorporation of several genes for vertical resistance into a plant may confer to it a degree of horizontal resistance. While there are major genes that condition only vertical resistance and polygenes that condition

only horizontal resistance, there is an overlapping area that needs resolving.

ACKNOWLEDGEMENT

The writer expresses his appreciation to C. Person for constructive com­ments and references to the literature.

LITERATURE CITED

1. Anderson, R. V., Hart, H. 1956. Effects of ionizing radiation on the host-parasite relationship of stem rust of wheat. Phytopathology 46:6 (Abstr.)

2. Bagga, H. S., Boone, D. M. 1968. Genes in Venturia inaequalis con­trolling pathogenicity to c rabap­pIes. Phytopathology 58: 1 1 76-82

3. Berry, R. W. 1963. Studies on the genetics of the host-parasite rela­tionships in corn leaf rust. PhD thesis. Univ. Wisconsin, Madison. 86 pp.

4. Black, W.. Mastenbroek, C., Mills, W. R., Peterson, L. C. 1953. A proposal for an international no­menclature of the races of Phy­tophthora infestans and of genes controlling immunity in Solanum demissum derivatives. Euphytica 2 : 173-79

5. Boone, D. M. 1971. Genetics of Ven­turia inaequalis. Ann. Rev. Phy­topathol. 9:297-3 18

6. Boone, D. M., Keitt, G. W. 1957. Ven­turia inaequalis (Cke.) Wint. XII. Genes controlling pathogenicity of wild-type lines. Phytopathology 47:403-09

7. B ridgmon, G. H., Wilcoxson, R. D. 1959. New races from mixtures of urediospores of varieties of Puccinia graminis. PhytoPathology 49:428-29

8. Browning, J. A., Frey, K. J. 1969. Multiline cultivars as a means of disease control. Ann. Rev. Phyto­pathol. 7:355-82

9. Bugbee, W. M., Line, R. F., Kern­kamp, M. F. 1968. Pathogenicity of progenies from selfing race 1 5B and somatic and sexual crosses of races 1 5B and 56 of Puccinia graminis f. sp. tritici. Phytopathology 58: 129 1-93

10. Caldecott, R. S., Stevens, H., Roberts, B. J; 1959. Stem rust resistant variants in irradiated populations -mutations or field hybrids ? Agron. J. 5 1 :401-03

1 1. Day, P. R. 1956. Race names of Cladosporium fulvum. Tomato Genet. Coop. Rep. 6, 1 3

1 2. Day, P . R . 1966. Recent developments in the genetics of the host-parasite system. Ann. Rev. Phytopathol. 4:245-68

1 3 . Day, P. R. 1968. Plant disease resis-

Ann

u. R

ev. P

hyto

path

ol. 1

971.

9:27

5-29

6. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tanf

ord

Uni

vers

ity -

Mai

n C

ampu

s -

Lan

e M

edic

al L

ibra

ry o

n 05

/18/

13. F

or p

erso

nal u

se o

nly.

294 FLOR

tance. Sci. Progr. London. 56: 35 7-70

14. DeVay, J. E., Schnathorst, W. C., Foda, M. S. 1967. Common anti­gens and host-parasite interac­tions. In The Dynamic Role of M oleclliar C onstitllents in Plant­Parasite Interaction, ed. C. J. Mirocha, I. Uritani, 3 1 8-28. St. Paul, Minn. : Am. Phytopathol. Soc. 372 pp.

15 . Doubly, J. A., Flor, H. H., Clagett, C. O. 1960. Relation of antigens of M elampsora lini and Linllm usitatissimum to resistance and susceptibility. Science 1 3 1 :229

1 6. Ellingboe, A. H. 1961. Somatic recom­bination in Puccinia graminis var. tritici. Phytopathology 5 1 : 1 3-15

17. Fincham, J. R. S., Day, P. R. 1965. Fungal Genetics, Blackwell, Ox­ford ; Davis, Philadelphia. 326 pp.

1 8 . Flangas, A. L., Dickson, J. G. 1 9 6 1 . The genetic control of pathogen­icity, serotypes, and variability in Puccinia sorghi. Am. J. Bot. 48: 275-85

19. Flor, H. H. 1935. Physiologic special­ization of M elampsora lini on Linum usitatissimum. J. Agr. Res. 5 1 :819-37

20. Flor, H. H. 1942. Inheritance of pathogenicity in a cross between physiologic races 22 and 24 of M elampsora lini. Phytopathology 32:5 (Abstr.)

2 1 . Flor, H. H. 1946. Genetics of patho­genicity in Melampsora lini. J. Agr. Res. 73:337-57

22. Flor, H. H. 1947. Inheritance of re­action to rust in flax. J. Agr. Res. 74:24 1-62

23. Flor, H. H. 1953. Epidemiology of flax rust in the North Central States. Phytopathology 43:624-28

24. Flor, H. H. 1954. Identification of races of flax rust by lines with single rust conditioning genes. U. S. Dept. Agr. Tech. Bull. 1087: 25

25. Flor, H. H. 1955. Host-parasite inter­action in flax rust-its genetics and other implications. Phytopath­ology 45: 680-85

26. Flor, H. H. 1956. The complementary genic systems in flax and flax rust. Advan. Genet. 8 :29-54

27. Flor, H. H. 1958. Mutation to wider virulence in Melampsora lini. Phy­topathology 48:297-301

28. Flor, H. H. 1960. The inheritance of X-ray induced mutations to vir-

ulence in a urediospore culture of race 1 of Melampsora lini. Phy­topathology 50: 603-05

29. Flor, H. H. 1964. Genetics of somatic variation for pathogenicity in M elampsora li"i. Phytopathology 54:823-26

30. Flor, H. H. 1965. Inheritance of smooth spore waH and pathogen­icity in Melampsora lini. Phyto­pathology 5 5 : 724-27

3 1 . Flor, H. H., Comstock, V. E. 1 9 7 1 . Development of flax cultivars with multiple rust-conditioning genes. Crop. Sci. 1 1 : 64-66

32. Freitas, A.P.C.e. 1969. Monogenic lines of wheat that differentiate Puccinia recondita Rob. ex Desm. Anales de la Estacion Experimen­tal de Aula Die. 9 (2-4) :235-44

33. Gallegly, M. E. 1968. Genetics of pathogenicity of Phytophthora infestans. Ann. Rev. Phytopathol. 6:3 75-96

34. Green, G. J. 1966. Selfing studies with races 10 and 1 1 of wheat stem rust. Can. J. Bot. 4 4 : 1 255-60

35. Hadwiger, L. A., Schwochau, M. E. 1969. Host resistance responses­

an induction hypothesis. Phyto­pathology 59 :223-27

36. Halisky, P. M. 1965. Physiologic specialization and genetics of the smut fungi, III. Botan. Rev. 31 : 1 14-50

37. Hatchett, J. H., GaUun, R. L. 1970. Genetics of the ability of the Hessian Fly. Mayetiola destructor, to survive on wheats having differ­ent genes for resistance. Annals Ent. Soc. America 63:1 400-07

38. Holton, C. 5., Halisky, P. M. 1960.

Dominance of avirulence and monogenic control of virulence in race hybrids of Ustilago avenae. Phytopathology 5 0:766-70

39. Holton, C. S., Hoffmann, J. A. 1968. Variation in the smut fungi. Ann. Rev. Phytopathol. 6:213-42

40. Hooker, A. L. 1967. The genetics and expression of resistance in plants to rusts of the genus Puccinia. Ann. Rev. Phytopathol. 5 : 1 63-82

4 1 . Howard, H. W. 1968. The relation between resistance genes in pota­toes and pathotypes of potato­root ee1worm (Heterodera ros­tochiensis), wart disease (Synchy­trium endobioticum) . and potato virus X. (Abstr.) Int. Congr. Plant Pathol., 1st, London, 92. 225 pp.

Ann

u. R

ev. P

hyto

path

ol. 1

971.

9:27

5-29

6. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tanf

ord

Uni

vers

ity -

Mai

n C

ampu

s -

Lan

e M

edic

al L

ibra

ry o

n 05

/18/

13. F

or p

erso

nal u

se o

nly.

CURRENT STATUS OF THE GENE-FaR-GENE CONCEPT 295

42. Johnson, T. 1961. Man-guided evolu­tion in plant rusts. Science 133: 357-72

43. Kao, K. N., Knott, D. R. 1969. The inheritance of pathogenicity in races 1 1 1 and 29 of wheat stem rust. Can. J. Genet. Cytol. 1 1 :266-74

44. Kerr, H. B. 1959. Physiological specialization of M elampsora lini (Pers. ) Lev. in Australia. Proc. Linn. Soc. N. S. W. 84:36-63

45. Kerr, H. B. 1960. The inheritance of resistance of Linum usitatissimum L. to the Australian Melampsora lini (Pers.) Lev. race complex. Proc. Linn. Soc. N. S. W. 85 :273-321

46. Knott, D. R. 1961. Attempts to pro­duce rust resistant mutants in wheat by irradiation. Can. J. Genet. Cytol. 3:20-22

47. Knott, D. R. 1966. The inheritance of stem rust resistance in wheat. Hereditas SHPPI. 2 : 1 5 6-66

48. Konzak. C. F. 1956. Induction of mutations for disease resistance in cereals. Brookhaven Symp. in BioI. No. 9 . Genetics in Plant Breeding. 236 pp.

49 . Leonard. K. J. 1969. Genetic equilib­ria in host-pathogen systems. Phy­topathology 59 : 1858-63

50. Leppik. E. E. 19 70. Gene centers of plants as sources of disease resis­tance. Ann. Rev. Phytopathol. 8 : 323-44

5 1 . Littlefield. L. J. 1969 . Flax rust re­sistance induced by prior inocula­tion with an avirulent race of M elampsora lini. Phytopathology 59: 1 323-28

52. Littlefield, L. J .• Aronson. S. J. 1969. Histological studies of M elampsora lini resistance in flax. Can. !. Bot. 47:1713-17

53. Loegering, W. Q., Powers, H. R., Jr. 1962. Inheritance of pathogenicity in a cross of physiologic races 1 1 1 and 36 of Puccinia graminis f. sp. tritid. Phytopathology 52: 547-54

54. Martens, J. W., McKenzie, R. H. 1., Green. G. W. 1970. Gene-for-gene relationships in the Avena :Pucci­nia graminis host-parasite system in Canada. Can. I. Bot. 48:969-75

55. Mayo. G. M. E . 1956. Linkage in LiHum flsitatissimum and in M e­lampsora lin; between genes con­.t.r,9.llin,l: host-patho§:en r�acti.ol),�.

Aust. I. Bioi. Sci. 9 : 1 8-36 56. Metzger, R. J., Trione, E. J. 1962.

Application of the gene-for-gene relationship hypothesis to. the Triticum-Tilletia system. Phyto­pathology 52:363 (Abstr.)

57. Mo.de, C. J. 1 9 58. A mathematical model for the co-evolution of obligate parasites and their hosts. Evolution 1 2 : 1 5 8-65

58. Moseman. J. G. 1957. Host-parasite interactions between culture 12A1 of the powdery mildew fungus and the MIt and Mig genes in barley. Phytopathology 47:453 (Abstr.)

59. Moseman, J. G. 1966. Genetics of powdery mildews. Ann. Rev. Phy­tapa tho I. 4:269-90

60. Myers, W. M. 1937. The nature and interaction of genes conditioning reaction to rust in flax. I. Agr. Res. 55:63 1-66

6 1 . No.ro.nha-Wagner, M.. Bettencourt, A. J. 1 9 67. Genetic study of the resistance of C offea SPp. to leaf rust. 1. Identification and behav­ior of four factors conditioning disease reaction in C offea arabica to twelve physiologic races of Hemileia vastatrix. Can. I. Bot. 45:2021-31

62. Oort, A. J. P. 1963. A gene-for-gene relationship in the TriticHm-U sti­lago system and some remarks o.n host-parasite combinations in gen­eral. Neth. I. Plant Pathol. 69: 104-09

63. Parmeter. J. R., Snyder, W. C., Reichle, R. E. 1963. Heterokaryo­sis and variability in plant-path­ogenic fungi. Ann. Rev. Phyto­pathol. 1:51-76

64. Person, C. 1959. Gene-for-gene rela­tionships in host:parasite systems. Can. !. Bot. 37: 1 10 1-30

65. Person, C. 1967. Genetic aspects of parasitism. Can. I. Bot. 45:1 193-1 204

66. Person, C., Samborski, D. J.. Roh­ringer, R. 19 62. The gene-for-gene concept. Nature 194:56 1-62

67. Person. c., Sidhu. G. 197 1 . Genetics of host-parasite interrelationships. (In press)

68. Powers. H. R.. Jr .• Sando. W. J. 1957. Genetics of host-parasite relationship in powdery mildew of wheat. Phytopathology 47:453 (Abstr.)

69. Prasada, R. 1948. Studies in linseed .rus�, M e�am'psor/l' li"i S P.ers,) J...ev,.

Ann

u. R

ev. P

hyto

path

ol. 1

971.

9:27

5-29

6. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tanf

ord

Uni

vers

ity -

Mai

n C

ampu

s -

Lan

e M

edic

al L

ibra

ry o

n 05

/18/

13. F

or p

erso

nal u

se o

nly.

296 FLOR

in India. Indian Phytopathol. 1 : 1-1 8

70. Puttick, G. F . 1921. The reaction of the F. generation of a cross be­tween a common and a dUTum wheat to two biologic forms of Puccinia graminis. Phytopathology 1 1 : 208-12

7 1 . Rowell, J. B., Loegering, W. Q., Pow­ers, H. R., Jr. 1963. Genetic model for physiologic studies of mechanisms governing develop­ment of infection type in wheat stem rust. Phytopathology 53:932-3 7

72. Sackston, W . E . 1 9 62. Studies o n sun­flower rust. III. Occurrence, dis­tribution and significance of races of Puccinia helianthi Schw. Can. 1. Bot. 40: 1449-58

73. Samborski, D. J., Dyck, P. L. 1968. Inheritance of virulence in wheat leaf rust on the standard differ­ential wheat varieties. Can. I. Genet. Cytol. 1 0:24-32

74. Schwinghamer, E. A. 1957. Radiation­induced mutations in race 1 of the flax rust fungus. Phytopathology 47:31 CAbstr.)

75. Schwinghamer, E. A. 1959. The re­lation between radiation dose and the frequency of mutations for pathogenicity in M elampsora lini. Phytopathology 49:260--69

76. Sheen, S. J. 1962. Studies on genetic factors controlling stem rust re­sistance in chromosome substitu­tim' lines of two wheat varieties. PhD thesis, Univ. Minnesota, St. Paul, 1 54 pp.

77. Toxopeus, H. J. 1956. Reflections on the origin of new physiologic races in Phytophthora infestans and the breeding of resistance in potatoes. Euphytica 5:221-37

78. Vallega, J.. Antonelli, E . F. 1960. Variaciones en la poblac i6n para-

sita de la roya del lino en la Argentina. Rev. Inv. Agric. 1 4: 403-20

79. van der Plank, J. E. 1968. Disease resistance in plants. New York: Academic 206 pp.

80. Walker, J. C. 1959. Progress and problems in controlling plant dis­eases by host resistance. In Plant Pathology, Problems and Progress, 1908-1958, 32-41. Univ. of Wis­consin Press, Madison, Wis. 588 pp.

81 . Waterhouse, W. L., Watson, 1. A. 1941. A note on determinations of physiologic specialization in flax rust. I. Proc. Roy. Soc. N.S.W. 7 5 : 1 1 5- 1 7

82. Watson, 1. A . 1970. Changes i n viru­lence and population shifts in plant pathogens. Ann. Rev. Phy­topathol. 8:209-30

83. Watson, 1. A., Luig, N. H. 1958. Somatic hybridization in Puccinia graminis var. tritici. Proc. Linn. Soc. N.s.W. 83:190-95

84. Watson, 1. A., Luig, N. H. 1962. Asexual intercrosses between so­matic recombinations of Puccinia graminis. Proc. Linn. Soc. N.S.W. 87:99-104

85. Wheeler, H. 1969. Genetics of path­ogenesis p. 9-1 3 . In Proceedings of a Symposium in Crop Protec­tion. New York State Agr. Exp. Sta., Geneva

86. Williams, N. D., Gough, F. J., Ron­don , M. R. 1966. Interaction of pathogenicity genes in Puccinia graminis f. sp. tritid and reac­tion genes in Triticum aestivum ssp. vulgare 'Marquis' and 'Re­liance'. Crop. Sci. 6:245-48

87. Zadoks, J. C. 196 1 . Yellow rust on wheat : studies in epidemiology and physiologic specialization. T. Pl.-Ziekten 67: 69-256

Ann

u. R

ev. P

hyto

path

ol. 1

971.

9:27

5-29

6. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tanf

ord

Uni

vers

ity -

Mai

n C

ampu

s -

Lan

e M

edic

al L

ibra

ry o

n 05

/18/

13. F

or p

erso

nal u

se o

nly.