| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Previous Article | Next Article ![]()
Infection and Immunity, May 2005, p. 2974-2985, Vol. 73, No. 5
0019-9567/05/$08.00+0 doi:10.1128/IAI.73.5.2974-2985.2005
Copyright © 2005, American Society for Microbiology. All Rights Reserved.
Departement de Sciences Biologiques, Université du Québec à Montréal, 1200 Rue Saint Alexandre, S-2055, Montréal, Québec H3B 3H5, Canada,1 Centre for Host-Parasite Interactions,2 Institute of Parasitology, McGill University, 21111 Lakeshore Road, Ste.-Anne-de-Bellevue, H9X3V9, Quebec, Canada,3 Department of Biotechnology and Environmental Biology, RMIT University, P.O. Box 71, 3083, Bundoora, Victoria, Australia,4 Department of Biochemistry & Molecular Biology, P.O. Box 13D, Monash University, Victoria 3800, Australia,5 Eijkman Institute for Molecular Biology, Jakarta Pusat 10430, Indonesia6
Received 11 October 2004/ Returned for modification 22 November 2004/ Accepted 21 January 2005
|
|
|---|
|
|
|---|
One strategy to circumvent problems associated with allelic polymorphism in malaria antigens is to focus on conserved antigens/epitopes that can be potentially cross-protective (19, 44). For example, the candidate vaccine antigen MSP4/5 is highly conserved in 14 strains (isolates) of Plasmodium yoelii (95 to 100% sequence identity), and vaccination with MSP4/5 proteins from two strains (Plasmodium yoelii killicki and Plasmodium yoelii nigeriensis) cross-protected mice against challenge with P. y. yoelii YM (20). A lesser degree of cross-protection against P. y. yoelii was also afforded by vaccination with MSP4/5 from Plasmodium berghei ANKA (81% sequence identity to P. yoelii), but no protection was observed with MSP4/5 from Plasmodium chabaudi adami DS (55% identity) (20). Such results provide evidence that conserved antigens can elicit good cross-strain protection and even partial cross-species protection. It is notable that several antigens from P. falciparum have been shown to cross-protect mice against murine malaria, further demonstrating that cross-species protection is feasible, presumably due to conservation of protective epitopes between species (6, 9, 10, 34, 35, 37, 53, 54): vaccine-induced partial cross-protection among P. falciparum strains is also described (38).
A key challenge, therefore, is the identification of conserved antigens/epitopes that could be employed for use as cross-strain vaccines. It is possible that among the 5,300 genes in the malaria genome, novel conserved antigens exist that have not yet been evaluated for protective efficacy (15, 18). In the present study, we have used expression library immunization (ELI) as an approach to identify new combinations of cross-protective antigens within the malaria genome, using P. c. adami in mice as a model test system. ELI has been shown to be protective against Mycoplasma (2), Leishmania (43), and Plasmodium, where we showed that mice were significantly protected against virulent P. c. adami DS infection by vaccination with a homologous genomic expression library (46, 58). We hypothesized that our library contains multiple antigen sequences, many of which could be novel hypothetical sequences known to exist in the genome (18), and that this multivalent antigen cocktail may elicit good cross-protection due to sequence conservation between strains. Here, we report the segregation of this protective library down to subpools of plasmids encoding several conserved sequences and demonstrate that these sequences induce strain-transcending immunity in a P. c. adami heterologous challenge model.
|
|
|---|
![]() View larger version (29K): [in a new window] |
FIG. 1. Segregation of the P. c. adami DS VR1020 genomic expression library and evaluation of vaccine efficacy. The original library construction and evaluation of the ability of the library to induce survival after homologous challenge with virulent P. c. adami DS are described down to the level of the 3KA subpools (46, 58). Subpool 3KA-4 induced the highest protection (63% survival). Following DNA sequencing of the 733 plasmids in the 3KA-4 subpool, the library was subdivided based on the size of the predicted ORF in-frame with the TPA leader sequence. The V9+ pool encodes ORFs of >9 aa: the V50+ pool encodes ORFs of >50 aa. The V9+ and V50+ pools were evaluated for efficacy against heterologous P. c. adami DK challenge, assessed as reduction in cumulative parasitemia.
|
Sequencing of plasmid inserts and BLAST alignment analysis.
The sequence of all 733 plasmid inserts in the 3KA-4 pool was determined across the TPA leader sequence and cloning site in the VR1020 vector. Each plasmid was grown overnight from individual 15% glycerol stocks in 10 ml of LB broth supplemented with kanamycin (50 µg/ml). Plasmid DNA was extracted according to the QIAGEN mini-prep procedure and sequenced. The results were translated with the Expasy DNA translation tool (http://us.expasy.org/), and the size of the ORF in frame with the TPA sequence was predicted. Initial segregation of the plasmids into the V50+ pool was performed in 2002 before publication of the malaria genome sequence. Subsequently, BLASTP searches, using the National Center for Biotechnology Information (NCBI) database versus the P. falciparum and P. yoelii genomes (http://www.ncbi.nlm.nih.gov/BLAST/Genome/plasmodium.html) and the PlasmoDB database-NCBI BLASTP 2.1.2 program versus Plasmodium ORFs of
50 bp (amino acids [aa]) (http://plasmodb.org/plasmodb/servlet/sv?page=blast), were used to compare the amino acid sequences encoded in the ORF genes to those of other Plasmodium proteins (expect value cutoff of 106) (4). BLASTX and BLASTN searches using the P. chabaudi BLAST Server at The Wellcome Trust Sanger Institute (http://www.sanger.ac.uk/cgi-bin/blast/submitblast/p_chabaudi) were used to identify sequences potentially unique to the P. chabaudi genome. Thirty-eight of the ORFs of >50 aa showed significant similarity with sequences in one or more Plasmodium species. The 29 ORFs in the V50+ pool found to have homology with P. falciparum 3D7 sequences were further analyzed to determine their likely expression profiles using the Winzeler database (32). The maximum hour of expression of each ORF was predicted according to P. falciparum 3D7 expression data published in the DeRisi Lab Malaria transcriptome database (5), where the maximum hour of blood-stage expression corresponds to a specific stage of parasite development (http://malaria.ucsf.edu/). The number of predicted transmembrane domains, as well as the presence of a signal sequence, was described for each ORF that was most similar to a P. falciparum or P. yoelii orthologue using PlasmoDB (http://plasmodb.org/). BLASTX analysis was performed using the PlasmoDB database with the NCBI BLASTX 2.1.2 program and compared with all Plasmodium ORFs of >50 bp (aa) (http://plasmodb.org/plasmodb/servlet/sv?page=BLAST) on the 18 inserts whose ORFs did not show significant similarity to Plasmodium proteins. Sixteen of these inserts showed an ORF with a significant hit to Plasmodium sequences in a nonexpressed frame, i.e., they carry genomic DNA that was inserted out of frame into the vector due to the random cloning strategy originally used to generate the library (58). BLASTN analysis using NCBI against all organisms (http://www.ncbi.nlm.nih.gov/BLAST/) showed that the DNA encoding the remaining two ORFs of >50 aa was homologous to rat or mouse genomic DNA deposited in GenBank in 2003: these ORFs are nonsense sequences, since they show no homology with any sequence in BLASTP searches. Thus, 38/56 ORFs of >50 aa represent authentic Plasmodium ORFs.
Vaccination and infection of mice. For intraepidermal (i.e.d.) vaccination, the gold particles coated with DNA were delivered in mice using a Helios gene gun (Bio-Rad Laboratories) with a pulse of helium gas at 400 lb/in2. Mice received three doses of 1 µg DNA per dose separated by two-week intervals (46). For intramuscular (i.m.) vaccination, DNA in phosphate-buffered saline (PBS) was delivered into the tibialis anterior muscle (three doses of 100 µg total DNA per dose, separated by 2-week intervals). P. c. adami DK parasites (isolate 556KA) were kindly provided by D. Walliker (University of Edinburgh). Inbred female BALB/c mice, ages 6 to 8 weeks old, were inoculated in the peritoneum with 5 x 104 infected red blood cells (IRBC) obtained from a syngeneic infected mouse 2 to 3 weeks following the last immunization. The i.m. experiment was conducted at the same time as the first i.e.d. experiment (Fig. 2). Parasitemia was determined as described previously (46).
![]() View larger version (17K): [in a new window] |
FIG. 2. Kinetics of infection and cumulative parasitemia of mice immunized with the VR1020/30K pool or the control VR1020 empty vector. Panels A-B and C-D represent the results of two independent experiments. Mice (n = 6) were vaccinated by the i.e.d. route (A-D) or the i.m. route (E, F). Two weeks after the third DNA dose, mice were challenged by intraperitoneal infection with 5 x 104 P. c. adami DK IRBC. Daily parasitemia is shown as % parasitemia (mean ± standard error of the mean [SEM]). Cumulative parasitemia values represent the sum of daily parasitemia throughout the period of patent infection. Groups were compared using a Student t test: *, P < 0.001; **, P < 0.01.
|
.
Splenocyte cell cultures were performed as described previously, using cells collected 15 days after the final DNA vaccination (46). Cell viability was greater than 98%, as determined by trypan blue exclusion (Invitrogen). Splenocytes were plated at a final concentration of 5 x 106 cells/ml in 24-well plates and were stimulated for 72 h with 2 x 106/ml noninfected erythrocytes or IRBC obtained from a syngeneic mouse at the moment of peak parasitemia (40 to 45% for P. c. adami DS and 15 to 22% for P. c. adami DK infections, respectively). Cell culture supernatants were harvested and tested for the presence of gamma interferon (IFN-
) by enzyme-linked immunosorbent assay (ELISA) as described previously (46). Opsonization and phagocytosis assays. Phagocytosis assays using peritoneal exudates cells were performed as described previously (41, 46). Adherent cells were incubated for 2 h with P. c. adami DS or DK IRBC (parasitemias between 15 and 30%) previously incubated for 1 h with inactivated immune sera or sera from DNA-vaccinated mice (n = 4; 1:20 serum dilution) collected before parasite challenge. The percentage of macrophages ingesting IRBC (300 macrophages per individual sample) was then quantified by light microscopy.
|
|
|---|
Vaccination of mice with a subpool of plasmids encoding ORFs of nine or more amino acids induces cross-protective immunity. Preliminary experiments using the P. c. adami DS model in BALB/c mice led to the identification of a subpool of 733 plasmids (pool 3KA-4) within the 30K library that elicited significant protection in two independent experiments (Fig. 1) (58; P. Smooker and T. W. Spithill, unpublished data). DNA sequences of the 733 inserts in these plasmids were obtained by sequencing across the cloning site in VR1020, and the sequences were analyzed for the presence of an ORF in frame with the TPA signal peptide and then catalogued according to the length of the encoded peptide. A pool of 303 clones containing plasmids encoding ORFs of >9 aa (large enough to generate major histocompatibility complex (MHC) binding peptides [55]) were selected to generate a new vaccine (V9+ pool), which was assessed against a control pool consisting of 63 plasmids containing out-of-frame (OOF) sequences from the P. c. adami DS library: this OOF control pool delivers malaria DNA but does not encode any ORF (unpublished data). Note that there is variation between experiments in the cumulative parasitemias observed in control mice vaccinated with either VR1020 or OOF DNA, and we have not observed reproducible differences in parasitemia in mice given these control vaccines. Intraepidermal delivery of the V9+ pool induced significant cross-protection (52% reduction in mean cumulative parasitemia, P < 0.001; 60% reduction in peak parasitemia, P < 0.05) against P. c. adami DK infection in BALB/c mice (Fig. 3 A and B). A second experiment confirmed the protective effect of the V9+ pool (45% reduction in cumulative parasitemia, P < 0.01; 45% reduction in peak parasitemia, P < 0.01 [Fig. 3C and D]). This level of protection is comparable to that induced by the original 30K vaccine pool (Fig. 2A to D).
![]() View larger version (19K): [in a new window] |
FIG. 3. Kinetics of infection and cumulative parasitemia of mice immunized with the VR1020/V9+ pool or the control VR1020/OOF pool by the i.e.d. route. Panels A-B and C-D represent the results of two independent experiments. Two weeks after the third DNA dose, the mice (n = 6) were challenged by intraperitoneal infection with 5 x 104 P. c. adami DK IRBC. Daily parasitemia is shown as % parasitemia (mean ± standard error of the mean [SEM]). Cumulative parasitemia represents the sum of daily parasitemia values throughout the period of patent infection. Groups were compared using a Student t test: *, P < 0.001; **, P < 0.01.
|
![]() View larger version (17K): [in a new window] |
FIG. 4. Kinetics of infection and cumulative parasitemia of mice immunized with the VR1020/V50+ pool or the control VR1020 empty vector by the i.e.d. route. Panels A-B and C-D represent the results of two independent experiments. Two weeks after the third DNA dose, the mice (n = 6) were challenged by intraperitoneal infection with 5 x 104 P. c. adami DK IRBC. Daily parasitemia is shown as % parasitemia (mean ± standard error of the mean [SEM]). Cumulative parasitemia represents the sum of daily parasitemia values throughout the period of patent infection. Groups were compared using a Student t test: *, P < 0.05.
|
-domain-related protein, putative PLP-transport ATPase, karyopherin beta, beta dynein heavy chain, ubiquitin-activating enzyme e11, putative eukaryote translation eIF-1a, PfEMP3, clustered asparagine-rich protein, subtilisin-like protease 2, transportin/putative bromodomain protein, putative tubulin-tyrosine ligase family, CCAAT box-binding protein subunit B, putative bir 1 protein) (Fig. 5B); 20 ORFs (52%) showed similarity with various Plasmodium hypothetical proteins, and 2 ORFs showed homology to unknown sequences from P. chabaudi (Fig. 5B). Overall, by BLASTP analysis, 29/38 ORFs (76%) were orthologues of P. falciparum proteins and a further 3 ORFs were orthologous to P. yoelii sequences that had known P. falciparum orthologues reported in PlasmoDB (Table 1). In addition, 32/38 ORFs (84%) shared significant similarity to P. yoelii proteins (Table 1). Interestingly, 22/38 ORFs (57%) aligned significantly with predicted proteins from 3 or more Plasmodium species, suggesting that these sequences encode polypeptides that are conserved during speciation of Plasmodium (Fig. 5A). Seven ORFs contain a predicted signal sequence, and 27 ORFs exhibit one or more predicted transmembrane anchor domains (Table 1). The stage of expression, level of expression, and peak time of expression during blood-stage infection of each of the ORFs homologous with sequences in P. falciparum were analyzed using the data from Le Roch et al. (32) and Bozdech et al. (5) (Table 1; Fig. 5C) (note that five of these ORFs were not found in the Bozdech et al. database). Of the 31 ORFs for which data exist in the Winzeler database, 30 ORFs are expressed in blood stages and 29 ORFs are expressed in each of the blood, gametocyte, and sporozoite stages. The peak times of ORF expression were distributed throughout the blood stage, and many of these ORFs are expressed at a relatively high level in the blood stage of P. falciparum (Table 1; Fig. 5C). These data show that a high proportion of the cross-protective V50+ vaccine sequences orthologous to P. falciparum are potential blood-stage antigens. |
View this table: [in a new window] |
TABLE 1. Sequence analysis of ORFs from P. c. adami DS DNA library
|
![]() View larger version (16K): [in a new window] |
FIG. 5. Sequence homology and blood-stage expression levels of the 38 ORFs orthologous to Plasmodium sequences in the V50+ vaccine. A: homology with other Plasmodium species; B: the annotation of these sequences; C: stacked bar chart showing the blood-stage expression patterns (in P. falciparum) of 31 sequences demonstrating significant orthology with P. falciparum defined by the maximal hour of expression (see reference 5) and by the overall level of blood-stage expression, using 150 units as the cutoff value to differentiate between high and low expression (32; Le Roch and Winzeler, personal communication). No sequence showed maximal expression in the merozoite stage. Twenty-eight sequences showed direct orthology to P. falciparum by BLASTP analysis (no data were available for gene X95276). Three additional sequences were orthologous to a P. yoelii sequence that was defined as orthologous to P. falciparum in PlasmoDB (see Table 1).
|
production by antigen-stimulated splenic cells.
Humoral and cellular immune responses both play a role in the control and resolution of murine malaria infections (31, 33, 42, 60, 61, 62). We have previously shown a role for both opsonizing antibodies and cellular responses in immunity induced by the 30K vaccine against P. c. adami DS challenge (46). In order to assess a role for antibody in vaccine-mediated cross-protection, the opsonizing capacity of the sera (n = 4) from mice vaccinated with the 30K library, the pool of V9+ plasmids, or the V50+ plasmids was assessed using homologous P. c. adami DS and heterologous P. c. adami DK IRBC. All sera significantly opsonized both P. c. adami DS and DK IRBC in a manner comparable to immune sera from mice that had recovered from infection with either the virulent DS or avirulent DK strains (Fig. 6), suggesting that the DS and DK strains share antigenic cross-reactivity at the surface of the infected erythrocyte.
![]() View larger version (26K): [in a new window] |
FIG. 6. Phagocytosis of P. c. adami IRBC preincubated with sera from mice immunized with the VR1020/30K, VR1020/V50+, or VR1020/V9+ vaccines and from mice that had received the control VR1020 empty vector. A. Uptake of P. c. adami DS IRBC. B. Uptake of P. c. adami DK IRBC. C. In a separate experiment, sera from mice vaccinated with the V9+ plasmids were assessed using P. c. adami DS and DK IRBC. DS immune and DK immune: sera from immune mice that had resolved a primary infection with P. c. adami DS and DK parasites were included as positive controls. Each analysis used sera from four individual mice. In each panel, the opsonization observed with the control VR1020 serum is compared individually with each vaccine serum (30K, V50+, or V9+) using a nonparametric Mann-Whitney t test. *, P < 0.03.
|
responses by splenic cells from vaccinated mice in response to native antigens from both P. c. adami DS and DK IRBC were induced by both the V9+ and the V50+ vaccines (Fig. 7), consistent with earlier observations with the 30K vaccine in the homologous P. c. adami DS model (46). We did not detect responses to noninfected erythrocytes in any of the groups evaluated (data not shown). These results show that both the V9+ and V50+ vaccine pools of recombinant plasmids primed cellular immune responses to native blood-stage antigens expressed by both homologous and heterologous IRBC.
![]() View larger version (16K): [in a new window] |
FIG. 7. IFN- secretion by splenocytes from individual BALB/c mice vaccinated i.e.d. with the gene gun. Mice were immunized with the VR1020/V9+ and VR1020/V50+ vaccines. Control values (CTRL) represent splenocyte cultures from mice that received the VR1020 empty vector. Splenocytes were kept either unstimulated (UNST.) or stimulated with 2 x106 IRBC (P. c. adami DK and/or DS) for 17 h. The supernatants from the splenocyte cultures were harvested, and IFN- was measured by ELISA. The responses in splenic cells from mice given empty VR1020 plasmid DNA stimulated with IRBC (CTRL + IRBC) were compared with the responses with splenic cells from mice given vaccine plasmid DNA also stimulated with IRBC (V9+ or V50+ IRBC), using a nonparametric Mann-Whitney t test. Panel A: *, P = 0.0159. Panel B: *, P = 0.0079. Panel C: *, P = 0.0002. Panel D: *, P = 0.0286).
|
|
|
|---|
Within the V9+ plasmid pool, we selected a subset of plasmids encoding ORFs of >50 aa (V50+ vaccine), hypothesizing that these large ORFs would contain a variety of T-cell or B-cell epitopes which may elicit cross-protection. The V50+ vaccine induced significant protection in BALB/c mice, although this was moderate when compared to the high protection induced by the 30K vaccine or by the V9+ vaccine. These differences suggest that shorter ORFs of <49 aa, excluded from the V50+ pool, may synergize with the larger ORFs and contribute to cross-protection against blood-stage infection. We are currently evaluating the immunogenic and synergistic properties associated with ORFs of smaller size.
Sequence analysis of the ORFs of >50 aa showed that 38 ORFs corresponded to authentic Plasmodium polypeptide sequences: 22 ORFs aligned with sequences from 3 or more Plasmodium species, and 29 ORF genes encoded sequences orthologous to P. falciparum sequences, many of which are expressed at high levels in the blood stage, suggesting that the multivalent cross-protective V50+ (and V9+) vaccines contain conserved blood-stage antigens. The V50+ pool contains 20 ORFs orthologous to hypothetical proteins, and 16 ORFs gave significant alignment scores with known Plasmodium proteins, such as the erythrocyte membrane-associated antigen PfEMP3, an erythrocyte membrane protein, the subtilisin-like protease 2, the ubiquitin-activating enzyme e11, and others. Of the seven ORFs exhibiting predicted signal sequences, four are hypothetical proteins with predicted transmembrane domains and one is an erythrocyte membrane-associated antigen: this may suggest that the V9+ and V50+ pools encode several putative membrane-associated antigens, which may contribute to the cross-protection observed.
Analysis of the immune responses induced by the V9+ and V50+ vaccine pools showed that mice primed with the P. c. adami DS vaccine sequences exhibited opsonizing antibodies and IFN-
-producing splenic cells that cross-reacted with native P. c. adami DK antigens expressed in IRBC. Comparables results were obtained in our earlier study that assessed the 30K library in three different targeting vectors: only the vaccine pools inducing both the humoral and cellular arms of immunity (opsonizing antibodies and IFN-
, respectively) were capable of protecting against lethal P. c. adami DS infection (46). Opsonizing antibodies in our model system promote the uptake of IRBC by macrophages via antibody binding to CD16/Fc
II and CD32/Fc
III receptors on macrophages (46). Since IFN-
has been shown to enhance expression of FcR on macrophages, it is possible that production of IFN-
enhances the protection conferred by opsonizing antibodies in the control of blood-stage malaria (21, 22, 47, 64). Antibody-dependent cellular inhibition is involved in protection against blood-stage malaria in humans, and vaccine-induced antibody-dependent cellular inhibition may be a useful strategy for a malaria blood-stage vaccine (3, 23, 36, 56, 57). It has been suggested that antibody binding to FcR on macrophages triggers the release of mediators that inhibit the replication of malaria parasites within opsonized erythrocytes (3, 36). Although FcR do not seem to play a role in the protection conferred by antibodies in the P. yoelii 17XL infection model, data using the P. berghei XAT nonlethal model suggest an important role for FcR-mediated phagocytosis in the protection transferred passively by antibodies (51, 62, 65).
The results suggest that the cross-protection observed here with the multivalent vaccines results from responses to one or more sequences within the P. c. adami genome that are conserved between the DS and DK strains. Since vaccination with known highly protective antigens, such as AMA1 and MSP1, does not induce cross-protection in mouse malaria model systems (13, 48, 52), one interpretation of the data is that cross-protection results from the synergistic or additive activity of responses to the multiple conserved antigens present in the DS vaccine pools. The sequence analysis of the V50+ pool directly shows that 29/38 Plasmodium orthologues in the pool exhibit a high degree of conservation with other Plasmodium species (e.g., E value in BLASTP of <3 x E-14; sequence identity of 53 to 100% with P. yoelii; Table 1). Twenty of the protective sequences >50 aa in length encode previously untested hypothetical proteins, and 14 of these proteins showed conservation across several malaria species, suggesting that conserved hypothetical proteins may be contributing to the observed cross-protection. This is analogous to results of Melby et al. (39), who identified pools of protective hypothetical proteins within a Leishmania ELI vaccine, and Haddad et al. (24), who identified hypothetical proteins in a P. yoelii liver-stage exon DNA vaccine pool. Our next objective is to further segregate the sequences to identify the minimal number of plasmids capable of inducing cross-protection. This may require a subset of sequences that induce both IFN-
secretion by splenic cells and opsonizing antibodies and may include small epitopes of <49 aa as well as conserved hypothetical sequences not previously considered as vaccine candidates. A possible role for such small epitopes is supported by observations that DNA vaccines encoding short peptides of size 17 to 51 aa have previously been shown to be effective vaccines in other systems (11, 66). We suggest that by selecting conserved sequences with significant similarity to several Plasmodium species, we may be able to identify immunogenic antigens and epitopes that will protect against heterologous P. c. adami strains and perhaps even subspecies. This is consistent with the observation that MSP4/5 sequences from strains of P. yoelii, recently shown to be conserved within P. yoelii, induced cross-protection against heterologous P. yoelii challenge: however, less cross-protection was observed with MSP4/5 from P. berghei against P. yoelii challenge (20). Cross-protection following exposure to P. falciparum strains is also described (7, 17, 26). We suggest that a multiepitope or polyprotein-based cross-protective blood-stage vaccine could feasibly be engineered for delivery in a DNA prime-virus boost approach as described for liver stage vaccines (38, 45).
|
|
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»